5 Changes between 1.0.0 and 1.1.0 [xx XXX xxxx]
7 *) Add an "external" session cache for debugging purposes to s_server. This
8 should help trace issues which normally are only apparent in deployed
12 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
13 a few changes are required:
15 Add SSL_OP_NO_TLSv1_1 flag.
17 Update version checking logic to handle version 1.1.
18 Add explicit IV handling (ported from DTLS code).
19 Add command line options to s_client/s_server.
22 *) Experiemental password based recipient info support for CMS library:
26 *) Split password based encryption into PBES2 and PBKDF2 functions. This
27 neatly separates the code into cipher and PBE sections and is required
28 for some algorithms that split PBES2 into separate pieces (such as
32 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
33 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
34 BIO_set_cipher() and some obscure PEM functions were changed so they
35 can now return an error. The RAND changes required a change to the
36 RAND_METHOD structure.
39 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
40 a gcc attribute to warn if the result of a function is ignored. This
41 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
42 whose return value is often ignored.
45 Changes between 0.9.8m (?) and 1.0.0 [xx XXX xxxx]
47 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
48 output hashes compatible with older versions of OpenSSL.
49 [Willy Weisz <weisz@vcpc.univie.ac.at>]
51 *) Fix compression algorithm handling: if resuming a session use the
52 compression algorithm of the resumed session instead of determining
53 it from client hello again. Don't allow server to change algorithm.
56 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
57 change when encrypting or decrypting.
60 *) Add load_crls() function to apps tidying load_certs() too. Add option
61 to verify utility to allow additional CRLs to be included.
64 *) Update OCSP request code to permit adding custom headers to the request:
65 some responders need this.
68 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
70 [Julia Lawall <julia@diku.dk>]
72 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
73 needlessly dereferenced structures, used obsolete functions and
74 didn't handle all updated verify codes correctly.
77 *) Disable MD2 in the default configuration.
80 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
81 indicate the initial BIO being pushed or popped. This makes it possible
82 to determine whether the BIO is the one explicitly called or as a result
83 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
84 it handles reference counts correctly and doesn't zero out the I/O bio
85 when it is not being explicitly popped. WARNING: applications which
86 included workarounds for the old buggy behaviour will need to be modified
87 or they could free up already freed BIOs.
90 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
91 renaming to all platforms (within the 0.9.8 branch, this was
92 done conditionally on Netware platforms to avoid a name clash).
93 [Guenter <lists@gknw.net>]
95 *) Add ECDHE and PSK support to DTLS.
96 [Michael Tuexen <tuexen@fh-muenster.de>]
98 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
102 *) Add "missing" function EVP_MD_flags() (without this the only way to
103 retrieve a digest flags is by accessing the structure directly. Update
104 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
105 or cipher is registered as in the "from" argument. Print out all
106 registered digests in the dgst usage message instead of manually
107 attempting to work them out.
110 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
111 this allows the use of compression and extensions. Change default cipher
112 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
113 by default unless an application cipher string requests it.
116 *) Alter match criteria in PKCS12_parse(). It used to try to use local
117 key ids to find matching certificates and keys but some PKCS#12 files
118 don't follow the (somewhat unwritten) rules and this strategy fails.
119 Now just gather all certificates together and the first private key
120 then look for the first certificate that matches the key.
123 *) Support use of registered digest and cipher names for dgst and cipher
124 commands instead of having to add each one as a special case. So now
131 openssl dgst -sha256 foo
133 and this works for ENGINE based algorithms too.
137 *) Update Gost ENGINE to support parameter files.
138 [Victor B. Wagner <vitus@cryptocom.ru>]
140 *) Support GeneralizedTime in ca utility.
141 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
143 *) Enhance the hash format used for certificate directory links. The new
144 form uses the canonical encoding (meaning equivalent names will work
145 even if they aren't identical) and uses SHA1 instead of MD5. This form
146 is incompatible with the older format and as a result c_rehash should
147 be used to rebuild symbolic links.
150 *) Make PKCS#8 the default write format for private keys, replacing the
151 traditional format. This form is standardised, more secure and doesn't
152 include an implicit MD5 dependency.
155 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
156 committed to OpenSSL should pass this lot as a minimum.
159 *) Add session ticket override functionality for use by EAP-FAST.
160 [Jouni Malinen <j@w1.fi>]
162 *) Modify HMAC functions to return a value. Since these can be implemented
163 in an ENGINE errors can occur.
166 *) Type-checked OBJ_bsearch_ex.
169 *) Type-checked OBJ_bsearch. Also some constification necessitated
170 by type-checking. Still to come: TXT_DB, bsearch(?),
171 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
175 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
176 seconds to a tm structure directly, instead of going through OS
177 specific date routines. This avoids any issues with OS routines such
178 as the year 2038 bug. New *_adj() functions for ASN1 time structures
179 and X509_time_adj_ex() to cover the extended range. The existing
180 X509_time_adj() is still usable and will no longer have any date issues.
183 *) Delta CRL support. New use deltas option which will attempt to locate
184 and search any appropriate delta CRLs available.
186 This work was sponsored by Google.
189 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
190 code and add additional score elements. Validate alternate CRL paths
191 as part of the CRL checking and indicate a new error "CRL path validation
192 error" in this case. Applications wanting additional details can use
193 the verify callback and check the new "parent" field. If this is not
194 NULL CRL path validation is taking place. Existing applications wont
195 see this because it requires extended CRL support which is off by
198 This work was sponsored by Google.
201 *) Support for freshest CRL extension.
203 This work was sponsored by Google.
206 *) Initial indirect CRL support. Currently only supported in the CRLs
207 passed directly and not via lookup. Process certificate issuer
208 CRL entry extension and lookup CRL entries by bother issuer name
209 and serial number. Check and process CRL issuer entry in IDP extension.
211 This work was sponsored by Google.
214 *) Add support for distinct certificate and CRL paths. The CRL issuer
215 certificate is validated separately in this case. Only enabled if
216 an extended CRL support flag is set: this flag will enable additional
217 CRL functionality in future.
219 This work was sponsored by Google.
222 *) Add support for policy mappings extension.
224 This work was sponsored by Google.
227 *) Fixes to pathlength constraint, self issued certificate handling,
228 policy processing to align with RFC3280 and PKITS tests.
230 This work was sponsored by Google.
233 *) Support for name constraints certificate extension. DN, email, DNS
234 and URI types are currently supported.
236 This work was sponsored by Google.
239 *) To cater for systems that provide a pointer-based thread ID rather
240 than numeric, deprecate the current numeric thread ID mechanism and
241 replace it with a structure and associated callback type. This
242 mechanism allows a numeric "hash" to be extracted from a thread ID in
243 either case, and on platforms where pointers are larger than 'long',
244 mixing is done to help ensure the numeric 'hash' is usable even if it
245 can't be guaranteed unique. The default mechanism is to use "&errno"
246 as a pointer-based thread ID to distinguish between threads.
248 Applications that want to provide their own thread IDs should now use
249 CRYPTO_THREADID_set_callback() to register a callback that will call
250 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
252 Note that ERR_remove_state() is now deprecated, because it is tied
253 to the assumption that thread IDs are numeric. ERR_remove_state(0)
254 to free the current thread's error state should be replaced by
255 ERR_remove_thread_state(NULL).
257 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
258 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
259 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
260 application was previously providing a numeric thread callback that
261 was inappropriate for distinguishing threads, then uniqueness might
262 have been obtained with &errno that happened immediately in the
263 intermediate development versions of OpenSSL; this is no longer the
264 case, the numeric thread callback will now override the automatic use
266 [Geoff Thorpe, with help from Bodo Moeller]
268 *) Initial support for different CRL issuing certificates. This covers a
269 simple case where the self issued certificates in the chain exist and
270 the real CRL issuer is higher in the existing chain.
272 This work was sponsored by Google.
275 *) Removed effectively defunct crypto/store from the build.
278 *) Revamp of STACK to provide stronger type-checking. Still to come:
279 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
280 ASN1_STRING, CONF_VALUE.
283 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
284 RAM on SSL connections. This option can save about 34k per idle SSL.
287 *) Revamp of LHASH to provide stronger type-checking. Still to come:
288 STACK, TXT_DB, bsearch, qsort.
291 *) Initial support for Cryptographic Message Syntax (aka CMS) based
292 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
293 support for data, signedData, compressedData, digestedData and
294 encryptedData, envelopedData types included. Scripts to check against
295 RFC4134 examples draft and interop and consistency checks of many
296 content types and variants.
299 *) Add options to enc utility to support use of zlib compression BIO.
302 *) Extend mk1mf to support importing of options and assembly language
303 files from Configure script, currently only included in VC-WIN32.
304 The assembly language rules can now optionally generate the source
305 files from the associated perl scripts.
308 *) Implement remaining functionality needed to support GOST ciphersuites.
309 Interop testing has been performed using CryptoPro implementations.
310 [Victor B. Wagner <vitus@cryptocom.ru>]
312 *) s390x assembler pack.
315 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
319 *) Implement Opaque PRF Input TLS extension as specified in
320 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
321 official specification yet and no extension type assignment by
322 IANA exists, this extension (for now) will have to be explicitly
323 enabled when building OpenSSL by providing the extension number
324 to use. For example, specify an option
326 -DTLSEXT_TYPE_opaque_prf_input=0x9527
328 to the "config" or "Configure" script to enable the extension,
329 assuming extension number 0x9527 (which is a completely arbitrary
330 and unofficial assignment based on the MD5 hash of the Internet
331 Draft). Note that by doing so, you potentially lose
332 interoperability with other TLS implementations since these might
333 be using the same extension number for other purposes.
335 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
336 opaque PRF input value to use in the handshake. This will create
337 an interal copy of the length-'len' string at 'src', and will
338 return non-zero for success.
340 To get more control and flexibility, provide a callback function
343 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
344 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
348 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
351 Callback function 'cb' will be called in handshakes, and is
352 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
353 Argument 'arg' is for application purposes (the value as given to
354 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
355 be provided to the callback function). The callback function
356 has to return non-zero to report success: usually 1 to use opaque
357 PRF input just if possible, or 2 to enforce use of the opaque PRF
358 input. In the latter case, the library will abort the handshake
359 if opaque PRF input is not successfully negotiated.
361 Arguments 'peerinput' and 'len' given to the callback function
362 will always be NULL and 0 in the case of a client. A server will
363 see the client's opaque PRF input through these variables if
364 available (NULL and 0 otherwise). Note that if the server
365 provides an opaque PRF input, the length must be the same as the
366 length of the client's opaque PRF input.
368 Note that the callback function will only be called when creating
369 a new session (session resumption can resume whatever was
370 previously negotiated), and will not be called in SSL 2.0
371 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
372 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
373 for applications that need to enforce opaque PRF input.
377 *) Update ssl code to support digests other than SHA1+MD5 for handshake
380 [Victor B. Wagner <vitus@cryptocom.ru>]
382 *) Add RFC4507 support to OpenSSL. This includes the corrections in
383 RFC4507bis. The encrypted ticket format is an encrypted encoded
384 SSL_SESSION structure, that way new session features are automatically
387 If a client application caches session in an SSL_SESSION structure
388 support is transparent because tickets are now stored in the encoded
391 The SSL_CTX structure automatically generates keys for ticket
392 protection in servers so again support should be possible
393 with no application modification.
395 If a client or server wishes to disable RFC4507 support then the option
396 SSL_OP_NO_TICKET can be set.
398 Add a TLS extension debugging callback to allow the contents of any client
399 or server extensions to be examined.
401 This work was sponsored by Google.
404 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
405 OpenSSL should now compile cleanly on gcc 4.2
406 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
408 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
409 support including streaming MAC support: this is required for GOST
411 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
413 *) Add option -stream to use PKCS#7 streaming in smime utility. New
414 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
415 to output in BER and PEM format.
418 *) Experimental support for use of HMAC via EVP_PKEY interface. This
419 allows HMAC to be handled via the EVP_DigestSign*() interface. The
420 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
421 ENGINE support for HMAC keys which are unextractable. New -mac and
422 -macopt options to dgst utility.
425 *) New option -sigopt to dgst utility. Update dgst to use
426 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
427 alternative signing paramaters such as X9.31 or PSS in the dgst
431 *) Change ssl_cipher_apply_rule(), the internal function that does
432 the work each time a ciphersuite string requests enabling
433 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
434 removing ("!foo+bar") a class of ciphersuites: Now it maintains
435 the order of disabled ciphersuites such that those ciphersuites
436 that most recently went from enabled to disabled not only stay
437 in order with respect to each other, but also have higher priority
438 than other disabled ciphersuites the next time ciphersuites are
441 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
442 the same ciphersuites as with "HIGH" alone, but in a specific
443 order where the PSK ciphersuites come first (since they are the
444 most recently disabled ciphersuites when "HIGH" is parsed).
446 Also, change ssl_create_cipher_list() (using this new
447 funcionality) such that between otherwise identical
448 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
452 *) Change ssl_create_cipher_list() so that it automatically
453 arranges the ciphersuites in reasonable order before starting
454 to process the rule string. Thus, the definition for "DEFAULT"
455 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
456 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
457 This makes it much easier to arrive at a reasonable default order
458 in applications for which anonymous ciphers are OK (meaning
459 that you can't actually use DEFAULT).
460 [Bodo Moeller; suggested by Victor Duchovni]
462 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
463 processing) into multiple integers instead of setting
464 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
465 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
466 (These masks as well as the individual bit definitions are hidden
467 away into the non-exported interface ssl/ssl_locl.h, so this
468 change to the definition of the SSL_CIPHER structure shouldn't
469 affect applications.) This give us more bits for each of these
470 categories, so there is no longer a need to coagulate AES128 and
471 AES256 into a single algorithm bit, and to coagulate Camellia128
472 and Camellia256 into a single algorithm bit, which has led to all
475 Thus, among other things, the kludge introduced in 0.9.7m and
476 0.9.8e for masking out AES256 independently of AES128 or masking
477 out Camellia256 independently of AES256 is not needed here in 0.9.9.
479 With the change, we also introduce new ciphersuite aliases that
480 so far were missing: "AES128", "AES256", "CAMELLIA128", and
484 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
485 Use the leftmost N bytes of the signature input if the input is
486 larger than the prime q (with N being the size in bytes of q).
489 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
490 it yet and it is largely untested.
493 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
496 *) Initial incomplete changes to avoid need for function casts in OpenSSL
497 some compilers (gcc 4.2 and later) reject their use. Safestack is
498 reimplemented. Update ASN1 to avoid use of legacy functions.
501 *) Win32/64 targets are linked with Winsock2.
504 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
505 to external functions. This can be used to increase CRL handling
506 efficiency especially when CRLs are very large by (for example) storing
507 the CRL revoked certificates in a database.
510 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
511 new CRLs added to a directory can be used. New command line option
512 -verify_return_error to s_client and s_server. This causes real errors
513 to be returned by the verify callback instead of carrying on no matter
514 what. This reflects the way a "real world" verify callback would behave.
517 *) GOST engine, supporting several GOST algorithms and public key formats.
518 Kindly donated by Cryptocom.
521 *) Partial support for Issuing Distribution Point CRL extension. CRLs
522 partitioned by DP are handled but no indirect CRL or reason partitioning
523 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
524 selected via a scoring technique which handles IDP and AKID in CRLs.
527 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
528 will ultimately be used for all verify operations: this will remove the
529 X509_STORE dependency on certificate verification and allow alternative
530 lookup methods. X509_STORE based implementations of these two callbacks.
533 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
534 Modify get_crl() to find a valid (unexpired) CRL if possible.
537 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
538 this would be called X509_CRL_cmp() but that name is already used by
539 a function that just compares CRL issuer names. Cache several CRL
540 extensions in X509_CRL structure and cache CRLDP in X509.
543 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
544 this maps equivalent X509_NAME structures into a consistent structure.
545 Name comparison can then be performed rapidly using memcmp().
548 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
552 *) Allow digests to supply their own micalg string for S/MIME type using
553 the ctrl EVP_MD_CTRL_MICALG.
556 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
557 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
558 ctrl. It can then customise the structure before and/or after signing
562 *) New function OBJ_add_sigid() to allow application defined signature OIDs
563 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
564 to free up any added signature OIDs.
567 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
568 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
569 digest and cipher tables. New options added to openssl utility:
570 list-message-digest-algorithms and list-cipher-algorithms.
573 *) Change the array representation of binary polynomials: the list
574 of degrees of non-zero coefficients is now terminated with -1.
575 Previously it was terminated with 0, which was also part of the
576 value; thus, the array representation was not applicable to
577 polynomials where t^0 has coefficient zero. This change makes
578 the array representation useful in a more general context.
581 *) Various modifications and fixes to SSL/TLS cipher string
582 handling. For ECC, the code now distinguishes between fixed ECDH
583 with RSA certificates on the one hand and with ECDSA certificates
584 on the other hand, since these are separate ciphersuites. The
585 unused code for Fortezza ciphersuites has been removed.
587 For consistency with EDH, ephemeral ECDH is now called "EECDH"
588 (not "ECDHE"). For consistency with the code for DH
589 certificates, use of ECDH certificates is now considered ECDH
590 authentication, not RSA or ECDSA authentication (the latter is
591 merely the CA's signing algorithm and not actively used in the
594 The temporary ciphersuite alias "ECCdraft" is no longer
595 available, and ECC ciphersuites are no longer excluded from "ALL"
596 and "DEFAULT". The following aliases now exist for RFC 4492
597 ciphersuites, most of these by analogy with the DH case:
599 kECDHr - ECDH cert, signed with RSA
600 kECDHe - ECDH cert, signed with ECDSA
601 kECDH - ECDH cert (signed with either RSA or ECDSA)
602 kEECDH - ephemeral ECDH
603 ECDH - ECDH cert or ephemeral ECDH
609 AECDH - anonymous ECDH
610 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
614 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
615 Use correct micalg parameters depending on digest(s) in signed message.
618 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
619 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
622 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
623 an engine to register a method. Add ENGINE lookups for methods and
624 functional reference processing.
627 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
628 EVP_{Sign,Verify}* which allow an application to customise the signature
632 *) New -resign option to smime utility. This adds one or more signers
633 to an existing PKCS#7 signedData structure. Also -md option to use an
634 alternative message digest algorithm for signing.
637 *) Tidy up PKCS#7 routines and add new functions to make it easier to
638 create PKCS7 structures containing multiple signers. Update smime
639 application to support multiple signers.
642 *) New -macalg option to pkcs12 utility to allow setting of an alternative
646 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
647 Reorganize PBE internals to lookup from a static table using NIDs,
648 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
649 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
650 PRF which will be automatically used with PBES2.
653 *) Replace the algorithm specific calls to generate keys in "req" with the
657 *) Update PKCS#7 enveloped data routines to use new API. This is now
658 supported by any public key method supporting the encrypt operation. A
659 ctrl is added to allow the public key algorithm to examine or modify
660 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
664 *) Add a ctrl to asn1 method to allow a public key algorithm to express
665 a default digest type to use. In most cases this will be SHA1 but some
666 algorithms (such as GOST) need to specify an alternative digest. The
667 return value indicates how strong the prefernce is 1 means optional and
668 2 is mandatory (that is it is the only supported type). Modify
669 ASN1_item_sign() to accept a NULL digest argument to indicate it should
670 use the default md. Update openssl utilities to use the default digest
671 type for signing if it is not explicitly indicated.
674 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
675 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
676 signing method from the key type. This effectively removes the link
677 between digests and public key types.
680 *) Add an OID cross reference table and utility functions. Its purpose is to
681 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
682 rsaEncryption. This will allow some of the algorithm specific hackery
683 needed to use the correct OID to be removed.
686 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
687 structures for PKCS7_sign(). They are now set up by the relevant public
691 *) Add provisional EC pkey method with support for ECDSA and ECDH.
694 *) Add support for key derivation (agreement) in the API, DH method and
698 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
699 public and private key formats. As a side effect these add additional
700 command line functionality not previously available: DSA signatures can be
701 generated and verified using pkeyutl and DH key support and generation in
706 [Oliver Tappe <zooey@hirschkaefer.de>]
708 *) New make target "install_html_docs" installs HTML renditions of the
710 [Oliver Tappe <zooey@hirschkaefer.de>]
712 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
713 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
714 support key and parameter generation and add initial key generation
715 functionality for RSA.
718 *) Add functions for main EVP_PKEY_method operations. The undocumented
719 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
720 EVP_PKEY_{encrypt,decrypt}_old.
723 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
724 key API, doesn't do much yet.
727 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
728 public key algorithms. New option to openssl utility:
729 "list-public-key-algorithms" to print out info.
732 *) Implement the Supported Elliptic Curves Extension for
733 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
736 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
737 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
740 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
741 utilities such as rsa, dsa, dsaparam etc except they process any key
745 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
746 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
747 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
751 *) Initial support for pluggable public key ASN1.
752 De-spaghettify the public key ASN1 handling. Move public and private
753 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
754 algorithm specific handling to a single module within the relevant
755 algorithm directory. Add functions to allow (near) opaque processing
756 of public and private key structures.
759 *) Implement the Supported Point Formats Extension for
760 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
763 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
764 for the psk identity [hint] and the psk callback functions to the
765 SSL_SESSION, SSL and SSL_CTX structure.
768 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
772 SSL_CTX_use_psk_identity_hint
773 SSL_get_psk_identity_hint
775 SSL_use_psk_identity_hint
777 [Mika Kousa and Pasi Eronen of Nokia Corporation]
779 *) Add RFC 3161 compliant time stamp request creation, response generation
780 and response verification functionality.
781 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
783 *) Add initial support for TLS extensions, specifically for the server_name
784 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
785 have new members for a host name. The SSL data structure has an
786 additional member SSL_CTX *initial_ctx so that new sessions can be
787 stored in that context to allow for session resumption, even after the
788 SSL has been switched to a new SSL_CTX in reaction to a client's
789 server_name extension.
791 New functions (subject to change):
794 SSL_get_servername_type()
797 New CTRL codes and macros (subject to change):
799 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
800 - SSL_CTX_set_tlsext_servername_callback()
801 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
802 - SSL_CTX_set_tlsext_servername_arg()
803 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
805 openssl s_client has a new '-servername ...' option.
807 openssl s_server has new options '-servername_host ...', '-cert2 ...',
808 '-key2 ...', '-servername_fatal' (subject to change). This allows
809 testing the HostName extension for a specific single host name ('-cert'
810 and '-key' remain fallbacks for handshakes without HostName
811 negotiation). If the unrecogninzed_name alert has to be sent, this by
812 default is a warning; it becomes fatal with the '-servername_fatal'
815 [Peter Sylvester, Remy Allais, Christophe Renou]
817 *) Whirlpool hash implementation is added.
820 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
821 bn(64,32). Because of instruction set limitations it doesn't have
822 any negative impact on performance. This was done mostly in order
823 to make it possible to share assembler modules, such as bn_mul_mont
824 implementations, between 32- and 64-bit builds without hassle.
827 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
828 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
832 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
833 dedicated Montgomery multiplication procedure, is introduced.
834 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
835 "64-bit" performance on certain 32-bit targets.
838 *) New option SSL_OP_NO_COMP to disable use of compression selectively
839 in SSL structures. New SSL ctrl to set maximum send fragment size.
840 Save memory by seeting the I/O buffer sizes dynamically instead of
841 using the maximum available value.
844 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
845 in addition to the text details.
848 *) Very, very preliminary EXPERIMENTAL support for printing of general
849 ASN1 structures. This currently produces rather ugly output and doesn't
850 handle several customised structures at all.
853 *) Integrated support for PVK file format and some related formats such
854 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
855 these in the 'rsa' and 'dsa' utilities.
858 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
861 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
862 place for the (very old) "NETSCAPE" format certificates which are now
863 handled using new ASN1 code equivalents.
866 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
867 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
868 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
871 *) Modify CRL distribution points extension code to print out previously
872 unsupported fields. Enhance extension setting code to allow setting of
876 *) Add print and set support for Issuing Distribution Point CRL extension.
879 *) Change 'Configure' script to enable Camellia by default.
882 Changes between 0.9.8l (?) and 0.9.8m (?) [xx XXX xxxx]
884 *) Modify compression code so it frees up structures without using the
885 ex_data callbacks. This works around a problem where some applications
886 call CRYPTO_free_all_ex_data() before application exit (e.g. when
887 restarting) then use compression (e.g. SSL with compression) later.
888 This results in significant per-connection memory leaks and
889 has caused some security issues including CVE-2008-1678 and
893 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
894 connect (but not renegotiate) with servers which do not support RI.
895 Until RI is more widely deployed this option is enabled by default.
898 *) Add "missing" ssl ctrls to clear options and mode.
901 *) If client attempts to renegotiate and doesn't support RI respond with
902 a no_renegotiation alert as required by draft-ietf-tls-renegotiation.
903 Some renegotiating TLS clients will continue a connection gracefully
904 when they receive the alert. Unfortunately OpenSSL mishandled
905 this alert and would hang waiting for a server hello which it will never
906 receive. Now we treat a received no_renegotiation alert as a fatal
907 error. This is because applications requesting a renegotiation might well
908 expect it to succeed and would have no code in place to handle the server
909 denying it so the only safe thing to do is to terminate the connection.
912 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
913 peer supports secure renegotiation and 0 otherwise. Print out peer
914 renegotiation support in s_client/s_server.
917 *) Replace the highly broken and deprecated SPKAC certification method with
918 the updated NID creation version. This should correctly handle UTF8.
921 *) Implement draft-ietf-tls-renegotiation-03. Re-enable
922 renegotiation but require the extension as needed. Unfortunately,
923 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION turns out to be a
924 bad idea. It has been replaced by
925 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
926 SSL_CTX_set_options(). This is really not recommended unless you
927 know what you are doing.
928 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
930 *) Fixes to stateless session resumption handling. Use initial_ctx when
931 issuing and attempting to decrypt tickets in case it has changed during
932 servername handling. Use a non-zero length session ID when attempting
933 stateless session resumption: this makes it possible to determine if
934 a resumption has occurred immediately after receiving server hello
935 (several places in OpenSSL subtly assume this) instead of later in
939 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
940 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
941 fixes for a few places where the return code is not checked
943 [Julia Lawall <julia@diku.dk>]
945 *) Add --strict-warnings option to Configure script to include devteam
946 warnings in other configurations.
949 *) Add support for --libdir option and LIBDIR variable in makefiles. This
950 makes it possible to install openssl libraries in locations which
951 have names other than "lib", for example "/usr/lib64" which some
953 [Steve Henson, based on patch from Jeremy Utley]
955 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
956 X690 8.9.12 and can produce some misleading textual output of OIDs.
957 [Steve Henson, reported by Dan Kaminsky]
959 *) Delete MD2 from algorithm tables. This follows the recommendation in
960 several standards that it is not used in new applications due to
961 several cryptographic weaknesses. For binary compatibility reasons
962 the MD2 API is still compiled in by default.
965 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
969 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
970 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
972 [Guenter <lists@gknw.net>]
974 *) Fix the server certificate chain building code to use X509_verify_cert(),
975 it used to have an ad-hoc builder which was unable to cope with anything
976 other than a simple chain.
977 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
979 *) Don't check self signed certificate signatures in X509_verify_cert()
980 by default (a flag can override this): it just wastes time without
981 adding any security. As a useful side effect self signed root CAs
982 with non-FIPS digests are now usable in FIPS mode.
985 *) In dtls1_process_out_of_seq_message() the check if the current message
986 is already buffered was missing. For every new message was memory
987 allocated, allowing an attacker to perform an denial of service attack
988 with sending out of seq handshake messages until there is no memory
989 left. Additionally every future messege was buffered, even if the
990 sequence number made no sense and would be part of another handshake.
991 So only messages with sequence numbers less than 10 in advance will be
992 buffered. (CVE-2009-1378)
993 [Robin Seggelmann, discovered by Daniel Mentz]
995 *) Records are buffered if they arrive with a future epoch to be
996 processed after finishing the corresponding handshake. There is
997 currently no limitation to this buffer allowing an attacker to perform
998 a DOS attack with sending records with future epochs until there is no
999 memory left. This patch adds the pqueue_size() function to detemine
1000 the size of a buffer and limits the record buffer to 100 entries.
1002 [Robin Seggelmann, discovered by Daniel Mentz]
1004 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
1005 parent structure is freed. (CVE-2009-1379)
1008 *) Handle non-blocking I/O properly in SSL_shutdown() call.
1009 [Darryl Miles <darryl-mailinglists@netbauds.net>]
1012 [Ilya O. <vrghost@gmail.com>]
1014 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
1016 *) Disable renegotiation completely - this fixes a severe security
1017 problem (CVE-2009-3555) at the cost of breaking all
1018 renegotiation. Renegotiation can be re-enabled by setting
1019 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
1020 run-time. This is really not recommended unless you know what
1024 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
1026 *) Don't set val to NULL when freeing up structures, it is freed up by
1027 underlying code. If sizeof(void *) > sizeof(long) this can result in
1028 zeroing past the valid field. (CVE-2009-0789)
1029 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
1031 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
1032 checked correctly. This would allow some invalid signed attributes to
1033 appear to verify correctly. (CVE-2009-0591)
1034 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1036 *) Reject UniversalString and BMPString types with invalid lengths. This
1037 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
1038 a legal length. (CVE-2009-0590)
1041 *) Set S/MIME signing as the default purpose rather than setting it
1042 unconditionally. This allows applications to override it at the store
1046 *) Permit restricted recursion of ASN1 strings. This is needed in practice
1047 to handle some structures.
1050 *) Improve efficiency of mem_gets: don't search whole buffer each time
1052 [Jeremy Shapiro <jnshapir@us.ibm.com>]
1054 *) New -hex option for openssl rand.
1057 *) Print out UTF8String and NumericString when parsing ASN1.
1060 *) Support NumericString type for name components.
1063 *) Allow CC in the environment to override the automatically chosen
1064 compiler. Note that nothing is done to ensure flags work with the
1068 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
1070 *) Properly check EVP_VerifyFinal() and similar return values
1072 [Ben Laurie, Bodo Moeller, Google Security Team]
1074 *) Enable TLS extensions by default.
1077 *) Allow the CHIL engine to be loaded, whether the application is
1078 multithreaded or not. (This does not release the developer from the
1079 obligation to set up the dynamic locking callbacks.)
1080 [Sander Temme <sander@temme.net>]
1082 *) Use correct exit code if there is an error in dgst command.
1083 [Steve Henson; problem pointed out by Roland Dirlewanger]
1085 *) Tweak Configure so that you need to say "experimental-jpake" to enable
1086 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
1089 *) Add experimental JPAKE support, including demo authentication in
1090 s_client and s_server.
1093 *) Set the comparison function in v3_addr_canonize().
1094 [Rob Austein <sra@hactrn.net>]
1096 *) Add support for XMPP STARTTLS in s_client.
1097 [Philip Paeps <philip@freebsd.org>]
1099 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
1100 to ensure that even with this option, only ciphersuites in the
1101 server's preference list will be accepted. (Note that the option
1102 applies only when resuming a session, so the earlier behavior was
1103 just about the algorithm choice for symmetric cryptography.)
1106 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
1108 *) Fix NULL pointer dereference if a DTLS server received
1109 ChangeCipherSpec as first record (CVE-2009-1386).
1112 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
1113 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
1116 *) The fix in 0.9.8c that supposedly got rid of unsafe
1117 double-checked locking was incomplete for RSA blinding,
1118 addressing just one layer of what turns out to have been
1119 doubly unsafe triple-checked locking.
1121 So now fix this for real by retiring the MONT_HELPER macro
1122 in crypto/rsa/rsa_eay.c.
1124 [Bodo Moeller; problem pointed out by Marius Schilder]
1126 *) Various precautionary measures:
1128 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
1130 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
1131 (NB: This would require knowledge of the secret session ticket key
1132 to exploit, in which case you'd be SOL either way.)
1134 - Change bn_nist.c so that it will properly handle input BIGNUMs
1135 outside the expected range.
1137 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
1140 [Neel Mehta, Bodo Moeller]
1142 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
1143 the load fails. Useful for distros.
1144 [Ben Laurie and the FreeBSD team]
1146 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
1149 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
1152 *) Expand ENGINE to support engine supplied SSL client certificate functions.
1154 This work was sponsored by Logica.
1157 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
1158 keystores. Support for SSL/TLS client authentication too.
1159 Not compiled unless enable-capieng specified to Configure.
1161 This work was sponsored by Logica.
1164 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
1165 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
1166 attribute creation routines such as certifcate requests and PKCS#12
1170 Changes between 0.9.8g and 0.9.8h [28 May 2008]
1172 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
1173 handshake which could lead to a cilent crash as found using the
1174 Codenomicon TLS test suite (CVE-2008-1672)
1175 [Steve Henson, Mark Cox]
1177 *) Fix double free in TLS server name extensions which could lead to
1178 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
1181 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
1183 Clear the error queue to ensure that error entries left from
1184 older function calls do not interfere with the correct operation.
1185 [Lutz Jaenicke, Erik de Castro Lopo]
1187 *) Remove root CA certificates of commercial CAs:
1189 The OpenSSL project does not recommend any specific CA and does not
1190 have any policy with respect to including or excluding any CA.
1191 Therefore it does not make any sense to ship an arbitrary selection
1192 of root CA certificates with the OpenSSL software.
1195 *) RSA OAEP patches to fix two separate invalid memory reads.
1196 The first one involves inputs when 'lzero' is greater than
1197 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
1198 before the beginning of from). The second one involves inputs where
1199 the 'db' section contains nothing but zeroes (there is a one-byte
1200 invalid read after the end of 'db').
1201 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1203 *) Partial backport from 0.9.9-dev:
1205 Introduce bn_mul_mont (dedicated Montgomery multiplication
1206 procedure) as a candidate for BIGNUM assembler implementation.
1207 While 0.9.9-dev uses assembler for various architectures, only
1208 x86_64 is available by default here in the 0.9.8 branch, and
1209 32-bit x86 is available through a compile-time setting.
1211 To try the 32-bit x86 assembler implementation, use Configure
1212 option "enable-montasm" (which exists only for this backport).
1214 As "enable-montasm" for 32-bit x86 disclaims code stability
1215 anyway, in this constellation we activate additional code
1216 backported from 0.9.9-dev for further performance improvements,
1217 namely BN_from_montgomery_word. (To enable this otherwise,
1218 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
1220 [Andy Polyakov (backport partially by Bodo Moeller)]
1222 *) Add TLS session ticket callback. This allows an application to set
1223 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
1224 values. This is useful for key rollover for example where several key
1225 sets may exist with different names.
1228 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
1229 This was broken until now in 0.9.8 releases, such that the only way
1230 a registered ENGINE could be used (assuming it initialises
1231 successfully on the host) was to explicitly set it as the default
1232 for the relevant algorithms. This is in contradiction with 0.9.7
1233 behaviour and the documentation. With this fix, when an ENGINE is
1234 registered into a given algorithm's table of implementations, the
1235 'uptodate' flag is reset so that auto-discovery will be used next
1236 time a new context for that algorithm attempts to select an
1238 [Ian Lister (tweaked by Geoff Thorpe)]
1240 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
1241 implemention in the following ways:
1243 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
1246 Lack of BER streaming support means one pass streaming processing is
1247 only supported if data is detached: setting the streaming flag is
1248 ignored for embedded content.
1250 CMS support is disabled by default and must be explicitly enabled
1251 with the enable-cms configuration option.
1254 *) Update the GMP engine glue to do direct copies between BIGNUM and
1255 mpz_t when openssl and GMP use the same limb size. Otherwise the
1256 existing "conversion via a text string export" trick is still used.
1257 [Paul Sheer <paulsheer@gmail.com>]
1259 *) Zlib compression BIO. This is a filter BIO which compressed and
1260 uncompresses any data passed through it.
1263 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
1264 RFC3394 compatible AES key wrapping.
1267 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
1268 sets string data without copying. X509_ALGOR_set0() and
1269 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
1270 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
1271 from an X509_ATTRIBUTE structure optionally checking it occurs only
1272 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
1276 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
1277 to get the expected BN_FLG_CONSTTIME behavior.
1278 [Bodo Moeller (Google)]
1282 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
1283 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
1284 - added some more tests to do_tests.pl
1285 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
1286 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
1287 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
1288 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
1289 - various changes to netware.pl to enable gcc-cross builds on Win32
1291 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
1292 - various changes to fix missing prototype warnings
1293 - fixed x86nasm.pl to create correct asm files for NASM COFF output
1294 - added AES, WHIRLPOOL and CPUID assembler code to build files
1295 - added missing AES assembler make rules to mk1mf.pl
1296 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
1297 [Guenter Knauf <eflash@gmx.net>]
1299 *) Implement certificate status request TLS extension defined in RFC3546.
1300 A client can set the appropriate parameters and receive the encoded
1301 OCSP response via a callback. A server can query the supplied parameters
1302 and set the encoded OCSP response in the callback. Add simplified examples
1303 to s_client and s_server.
1306 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
1308 *) Fix various bugs:
1309 + Binary incompatibility of ssl_ctx_st structure
1310 + DTLS interoperation with non-compliant servers
1311 + Don't call get_session_cb() without proposed session
1312 + Fix ia64 assembler code
1313 [Andy Polyakov, Steve Henson]
1315 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
1317 *) DTLS Handshake overhaul. There were longstanding issues with
1318 OpenSSL DTLS implementation, which were making it impossible for
1319 RFC 4347 compliant client to communicate with OpenSSL server.
1320 Unfortunately just fixing these incompatibilities would "cut off"
1321 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
1322 server keeps tolerating non RFC compliant syntax. The opposite is
1323 not true, 0.9.8f client can not communicate with earlier server.
1324 This update even addresses CVE-2007-4995.
1327 *) Changes to avoid need for function casts in OpenSSL: some compilers
1328 (gcc 4.2 and later) reject their use.
1329 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
1332 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1333 RFC4507bis. The encrypted ticket format is an encrypted encoded
1334 SSL_SESSION structure, that way new session features are automatically
1337 If a client application caches session in an SSL_SESSION structure
1338 support is transparent because tickets are now stored in the encoded
1341 The SSL_CTX structure automatically generates keys for ticket
1342 protection in servers so again support should be possible
1343 with no application modification.
1345 If a client or server wishes to disable RFC4507 support then the option
1346 SSL_OP_NO_TICKET can be set.
1348 Add a TLS extension debugging callback to allow the contents of any client
1349 or server extensions to be examined.
1351 This work was sponsored by Google.
1354 *) Add initial support for TLS extensions, specifically for the server_name
1355 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1356 have new members for a host name. The SSL data structure has an
1357 additional member SSL_CTX *initial_ctx so that new sessions can be
1358 stored in that context to allow for session resumption, even after the
1359 SSL has been switched to a new SSL_CTX in reaction to a client's
1360 server_name extension.
1362 New functions (subject to change):
1364 SSL_get_servername()
1365 SSL_get_servername_type()
1368 New CTRL codes and macros (subject to change):
1370 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1371 - SSL_CTX_set_tlsext_servername_callback()
1372 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1373 - SSL_CTX_set_tlsext_servername_arg()
1374 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1376 openssl s_client has a new '-servername ...' option.
1378 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1379 '-key2 ...', '-servername_fatal' (subject to change). This allows
1380 testing the HostName extension for a specific single host name ('-cert'
1381 and '-key' remain fallbacks for handshakes without HostName
1382 negotiation). If the unrecogninzed_name alert has to be sent, this by
1383 default is a warning; it becomes fatal with the '-servername_fatal'
1386 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
1388 *) Add AES and SSE2 assembly language support to VC++ build.
1391 *) Mitigate attack on final subtraction in Montgomery reduction.
1394 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
1395 (which previously caused an internal error).
1398 *) Squeeze another 10% out of IGE mode when in != out.
1401 *) AES IGE mode speedup.
1402 [Dean Gaudet (Google)]
1404 *) Add the Korean symmetric 128-bit cipher SEED (see
1405 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
1406 add SEED ciphersuites from RFC 4162:
1408 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
1409 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
1410 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
1411 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
1413 To minimize changes between patchlevels in the OpenSSL 0.9.8
1414 series, SEED remains excluded from compilation unless OpenSSL
1415 is configured with 'enable-seed'.
1416 [KISA, Bodo Moeller]
1418 *) Mitigate branch prediction attacks, which can be practical if a
1419 single processor is shared, allowing a spy process to extract
1420 information. For detailed background information, see
1421 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
1422 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
1423 and Necessary Software Countermeasures"). The core of the change
1424 are new versions BN_div_no_branch() and
1425 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
1426 respectively, which are slower, but avoid the security-relevant
1427 conditional branches. These are automatically called by BN_div()
1428 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
1429 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
1430 remove a conditional branch.
1432 BN_FLG_CONSTTIME is the new name for the previous
1433 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
1434 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
1435 in the exponent causes BN_mod_exp_mont() to use the alternative
1436 implementation in BN_mod_exp_mont_consttime().) The old name
1437 remains as a deprecated alias.
1439 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
1440 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
1441 constant-time implementations for more than just exponentiation.
1442 Here too the old name is kept as a deprecated alias.
1444 BN_BLINDING_new() will now use BN_dup() for the modulus so that
1445 the BN_BLINDING structure gets an independent copy of the
1446 modulus. This means that the previous "BIGNUM *m" argument to
1447 BN_BLINDING_new() and to BN_BLINDING_create_param() now
1448 essentially becomes "const BIGNUM *m", although we can't actually
1449 change this in the header file before 0.9.9. It allows
1450 RSA_setup_blinding() to use BN_with_flags() on the modulus to
1451 enable BN_FLG_CONSTTIME.
1453 [Matthew D Wood (Intel Corp)]
1455 *) In the SSL/TLS server implementation, be strict about session ID
1456 context matching (which matters if an application uses a single
1457 external cache for different purposes). Previously,
1458 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
1459 set. This did ensure strict client verification, but meant that,
1460 with applications using a single external cache for quite
1461 different requirements, clients could circumvent ciphersuite
1462 restrictions for a given session ID context by starting a session
1463 in a different context.
1466 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
1467 a ciphersuite string such as "DEFAULT:RSA" cannot enable
1468 authentication-only ciphersuites.
1471 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
1472 not complete and could lead to a possible single byte overflow
1473 (CVE-2007-5135) [Ben Laurie]
1475 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
1477 *) Since AES128 and AES256 (and similarly Camellia128 and
1478 Camellia256) share a single mask bit in the logic of
1479 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
1480 kludge to work properly if AES128 is available and AES256 isn't
1481 (or if Camellia128 is available and Camellia256 isn't).
1484 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
1485 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
1486 When a point or a seed is encoded in a BIT STRING, we need to
1487 prevent the removal of trailing zero bits to get the proper DER
1488 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
1489 of a NamedBitList, for which trailing 0 bits need to be removed.)
1492 *) Have SSL/TLS server implementation tolerate "mismatched" record
1493 protocol version while receiving ClientHello even if the
1494 ClientHello is fragmented. (The server can't insist on the
1495 particular protocol version it has chosen before the ServerHello
1496 message has informed the client about his choice.)
1499 *) Add RFC 3779 support.
1500 [Rob Austein for ARIN, Ben Laurie]
1502 *) Load error codes if they are not already present instead of using a
1503 static variable. This allows them to be cleanly unloaded and reloaded.
1504 Improve header file function name parsing.
1507 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
1508 or CAPABILITY handshake as required by RFCs.
1511 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
1513 *) Introduce limits to prevent malicious keys being able to
1514 cause a denial of service. (CVE-2006-2940)
1515 [Steve Henson, Bodo Moeller]
1517 *) Fix ASN.1 parsing of certain invalid structures that can result
1518 in a denial of service. (CVE-2006-2937) [Steve Henson]
1520 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
1521 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
1523 *) Fix SSL client code which could crash if connecting to a
1524 malicious SSLv2 server. (CVE-2006-4343)
1525 [Tavis Ormandy and Will Drewry, Google Security Team]
1527 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
1528 match only those. Before that, "AES256-SHA" would be interpreted
1529 as a pattern and match "AES128-SHA" too (since AES128-SHA got
1530 the same strength classification in 0.9.7h) as we currently only
1531 have a single AES bit in the ciphersuite description bitmap.
1532 That change, however, also applied to ciphersuite strings such as
1533 "RC4-MD5" that intentionally matched multiple ciphersuites --
1534 namely, SSL 2.0 ciphersuites in addition to the more common ones
1535 from SSL 3.0/TLS 1.0.
1537 So we change the selection algorithm again: Naming an explicit
1538 ciphersuite selects this one ciphersuite, and any other similar
1539 ciphersuite (same bitmap) from *other* protocol versions.
1540 Thus, "RC4-MD5" again will properly select both the SSL 2.0
1541 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
1543 Since SSL 2.0 does not have any ciphersuites for which the
1544 128/256 bit distinction would be relevant, this works for now.
1545 The proper fix will be to use different bits for AES128 and
1546 AES256, which would have avoided the problems from the beginning;
1547 however, bits are scarce, so we can only do this in a new release
1548 (not just a patchlevel) when we can change the SSL_CIPHER
1549 definition to split the single 'unsigned long mask' bitmap into
1550 multiple values to extend the available space.
1554 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
1556 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
1557 (CVE-2006-4339) [Ben Laurie and Google Security Team]
1559 *) Add AES IGE and biIGE modes.
1562 *) Change the Unix randomness entropy gathering to use poll() when
1563 possible instead of select(), since the latter has some
1564 undesirable limitations.
1565 [Darryl Miles via Richard Levitte and Bodo Moeller]
1567 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
1568 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
1569 cannot be implicitly activated as part of, e.g., the "AES" alias.
1570 However, please upgrade to OpenSSL 0.9.9[-dev] for
1571 non-experimental use of the ECC ciphersuites to get TLS extension
1572 support, which is required for curve and point format negotiation
1573 to avoid potential handshake problems.
1576 *) Disable rogue ciphersuites:
1578 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
1579 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
1580 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
1582 The latter two were purportedly from
1583 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
1586 Also deactivate the remaining ciphersuites from
1587 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
1588 unofficial, and the ID has long expired.
1591 *) Fix RSA blinding Heisenbug (problems sometimes occured on
1592 dual-core machines) and other potential thread-safety issues.
1595 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
1596 versions), which is now available for royalty-free use
1597 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
1598 Also, add Camellia TLS ciphersuites from RFC 4132.
1600 To minimize changes between patchlevels in the OpenSSL 0.9.8
1601 series, Camellia remains excluded from compilation unless OpenSSL
1602 is configured with 'enable-camellia'.
1605 *) Disable the padding bug check when compression is in use. The padding
1606 bug check assumes the first packet is of even length, this is not
1607 necessarily true if compresssion is enabled and can result in false
1608 positives causing handshake failure. The actual bug test is ancient
1609 code so it is hoped that implementations will either have fixed it by
1610 now or any which still have the bug do not support compression.
1613 Changes between 0.9.8a and 0.9.8b [04 May 2006]
1615 *) When applying a cipher rule check to see if string match is an explicit
1616 cipher suite and only match that one cipher suite if it is.
1619 *) Link in manifests for VC++ if needed.
1620 [Austin Ziegler <halostatue@gmail.com>]
1622 *) Update support for ECC-based TLS ciphersuites according to
1623 draft-ietf-tls-ecc-12.txt with proposed changes (but without
1624 TLS extensions, which are supported starting with the 0.9.9
1625 branch, not in the OpenSSL 0.9.8 branch).
1628 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
1629 opaque EVP_CIPHER_CTX handling.
1632 *) Fixes and enhancements to zlib compression code. We now only use
1633 "zlib1.dll" and use the default __cdecl calling convention on Win32
1634 to conform with the standards mentioned here:
1635 http://www.zlib.net/DLL_FAQ.txt
1636 Static zlib linking now works on Windows and the new --with-zlib-include
1637 --with-zlib-lib options to Configure can be used to supply the location
1638 of the headers and library. Gracefully handle case where zlib library
1642 *) Several fixes and enhancements to the OID generation code. The old code
1643 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
1644 handle numbers larger than ULONG_MAX, truncated printing and had a
1645 non standard OBJ_obj2txt() behaviour.
1648 *) Add support for building of engines under engine/ as shared libraries
1649 under VC++ build system.
1652 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
1653 Hopefully, we will not see any false combination of paths any more.
1656 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
1658 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1659 (part of SSL_OP_ALL). This option used to disable the
1660 countermeasure against man-in-the-middle protocol-version
1661 rollback in the SSL 2.0 server implementation, which is a bad
1662 idea. (CVE-2005-2969)
1664 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
1665 for Information Security, National Institute of Advanced Industrial
1666 Science and Technology [AIST], Japan)]
1668 *) Add two function to clear and return the verify parameter flags.
1671 *) Keep cipherlists sorted in the source instead of sorting them at
1672 runtime, thus removing the need for a lock.
1675 *) Avoid some small subgroup attacks in Diffie-Hellman.
1676 [Nick Mathewson and Ben Laurie]
1678 *) Add functions for well-known primes.
1681 *) Extended Windows CE support.
1682 [Satoshi Nakamura and Andy Polyakov]
1684 *) Initialize SSL_METHOD structures at compile time instead of during
1685 runtime, thus removing the need for a lock.
1688 *) Make PKCS7_decrypt() work even if no certificate is supplied by
1689 attempting to decrypt each encrypted key in turn. Add support to
1693 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
1695 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
1698 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
1701 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
1702 key into the same file any more.
1705 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
1708 *) Add -utf8 command line and config file option to 'ca'.
1709 [Stefan <stf@udoma.org]
1711 *) Removed the macro des_crypt(), as it seems to conflict with some
1712 libraries. Use DES_crypt().
1715 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
1716 involves renaming the source and generated shared-libs for
1717 both. The engines will accept the corrected or legacy ids
1718 ('ncipher' and '4758_cca' respectively) when binding. NB,
1719 this only applies when building 'shared'.
1720 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
1722 *) Add attribute functions to EVP_PKEY structure. Modify
1723 PKCS12_create() to recognize a CSP name attribute and
1724 use it. Make -CSP option work again in pkcs12 utility.
1727 *) Add new functionality to the bn blinding code:
1728 - automatic re-creation of the BN_BLINDING parameters after
1729 a fixed number of uses (currently 32)
1730 - add new function for parameter creation
1731 - introduce flags to control the update behaviour of the
1732 BN_BLINDING parameters
1733 - hide BN_BLINDING structure
1734 Add a second BN_BLINDING slot to the RSA structure to improve
1735 performance when a single RSA object is shared among several
1739 *) Add support for DTLS.
1740 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
1742 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
1743 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
1746 *) Remove buggy and incompletet DH cert support from
1747 ssl/ssl_rsa.c and ssl/s3_both.c
1750 *) Use SHA-1 instead of MD5 as the default digest algorithm for
1751 the apps/openssl applications.
1754 *) Compile clean with "-Wall -Wmissing-prototypes
1755 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
1756 DEBUG_SAFESTACK must also be set.
1759 *) Change ./Configure so that certain algorithms can be disabled by default.
1760 The new counterpiece to "no-xxx" is "enable-xxx".
1762 The patented RC5 and MDC2 algorithms will now be disabled unless
1763 "enable-rc5" and "enable-mdc2", respectively, are specified.
1765 (IDEA remains enabled despite being patented. This is because IDEA
1766 is frequently required for interoperability, and there is no license
1767 fee for non-commercial use. As before, "no-idea" can be used to
1768 avoid this algorithm.)
1772 *) Add processing of proxy certificates (see RFC 3820). This work was
1773 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
1774 EGEE (Enabling Grids for E-science in Europe).
1777 *) RC4 performance overhaul on modern architectures/implementations, such
1778 as Intel P4, IA-64 and AMD64.
1781 *) New utility extract-section.pl. This can be used specify an alternative
1782 section number in a pod file instead of having to treat each file as
1783 a separate case in Makefile. This can be done by adding two lines to the
1786 =for comment openssl_section:XXX
1788 The blank line is mandatory.
1792 *) New arguments -certform, -keyform and -pass for s_client and s_server
1793 to allow alternative format key and certificate files and passphrase
1797 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
1798 update associated structures and add various utility functions.
1800 Add new policy related verify parameters, include policy checking in
1801 standard verify code. Enhance 'smime' application with extra parameters
1802 to support policy checking and print out.
1805 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
1806 Nehemiah processors. These extensions support AES encryption in hardware
1807 as well as RNG (though RNG support is currently disabled).
1808 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
1810 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
1813 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
1814 [Andy Polyakov and a number of other people]
1816 *) Improved PowerPC platform support. Most notably BIGNUM assembler
1817 implementation contributed by IBM.
1818 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
1820 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
1821 exponent rather than 'unsigned long'. There is a corresponding change to
1822 the new 'rsa_keygen' element of the RSA_METHOD structure.
1823 [Jelte Jansen, Geoff Thorpe]
1825 *) Functionality for creating the initial serial number file is now
1826 moved from CA.pl to the 'ca' utility with a new option -create_serial.
1828 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
1829 number file to 1, which is bound to cause problems. To avoid
1830 the problems while respecting compatibility between different 0.9.7
1831 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
1832 CA.pl for serial number initialization. With the new release 0.9.8,
1833 we can fix the problem directly in the 'ca' utility.)
1836 *) Reduced header interdepencies by declaring more opaque objects in
1837 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
1838 give fewer recursive includes, which could break lazy source code - so
1839 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
1840 developers should define this symbol when building and using openssl to
1841 ensure they track the recommended behaviour, interfaces, [etc], but
1842 backwards-compatible behaviour prevails when this isn't defined.
1845 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
1848 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
1849 This will generate a random key of the appropriate length based on the
1850 cipher context. The EVP_CIPHER can provide its own random key generation
1851 routine to support keys of a specific form. This is used in the des and
1852 3des routines to generate a key of the correct parity. Update S/MIME
1853 code to use new functions and hence generate correct parity DES keys.
1854 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
1855 valid (weak or incorrect parity).
1858 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
1859 as looking them up. This is useful when the verified structure may contain
1860 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
1861 present unless the new PKCS7_NO_CRL flag is asserted.
1864 *) Extend ASN1 oid configuration module. It now additionally accepts the
1867 shortName = some long name, 1.2.3.4
1870 *) Reimplemented the BN_CTX implementation. There is now no more static
1871 limitation on the number of variables it can handle nor the depth of the
1872 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
1873 information can now expand as required, and rather than having a single
1874 static array of bignums, BN_CTX now uses a linked-list of such arrays
1875 allowing it to expand on demand whilst maintaining the usefulness of
1876 BN_CTX's "bundling".
1879 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
1880 to allow all RSA operations to function using a single BN_CTX.
1883 *) Preliminary support for certificate policy evaluation and checking. This
1884 is initially intended to pass the tests outlined in "Conformance Testing
1885 of Relying Party Client Certificate Path Processing Logic" v1.07.
1888 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
1889 remained unused and not that useful. A variety of other little bignum
1890 tweaks and fixes have also been made continuing on from the audit (see
1894 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
1895 associated ASN1, EVP and SSL functions and old ASN1 macros.
1898 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
1899 and this should never fail. So the return value from the use of
1900 BN_set_word() (which can fail due to needless expansion) is now deprecated;
1901 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
1904 *) BN_CTX_get() should return zero-valued bignums, providing the same
1905 initialised value as BN_new().
1906 [Geoff Thorpe, suggested by Ulf Möller]
1908 *) Support for inhibitAnyPolicy certificate extension.
1911 *) An audit of the BIGNUM code is underway, for which debugging code is
1912 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
1913 is considered valid when processing BIGNUMs, and causes execution to
1914 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
1915 further steps are taken to deliberately pollute unused data in BIGNUM
1916 structures to try and expose faulty code further on. For now, openssl will
1917 (in its default mode of operation) continue to tolerate the inconsistent
1918 forms that it has tolerated in the past, but authors and packagers should
1919 consider trying openssl and their own applications when compiled with
1920 these debugging symbols defined. It will help highlight potential bugs in
1921 their own code, and will improve the test coverage for OpenSSL itself. At
1922 some point, these tighter rules will become openssl's default to improve
1923 maintainability, though the assert()s and other overheads will remain only
1924 in debugging configurations. See bn.h for more details.
1925 [Geoff Thorpe, Nils Larsch, Ulf Möller]
1927 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
1928 that can only be obtained through BN_CTX_new() (which implicitly
1929 initialises it). The presence of this function only made it possible
1930 to overwrite an existing structure (and cause memory leaks).
1933 *) Because of the callback-based approach for implementing LHASH as a
1934 template type, lh_insert() adds opaque objects to hash-tables and
1935 lh_doall() or lh_doall_arg() are typically used with a destructor callback
1936 to clean up those corresponding objects before destroying the hash table
1937 (and losing the object pointers). So some over-zealous constifications in
1938 LHASH have been relaxed so that lh_insert() does not take (nor store) the
1939 objects as "const" and the lh_doall[_arg] callback wrappers are not
1940 prototyped to have "const" restrictions on the object pointers they are
1941 given (and so aren't required to cast them away any more).
1944 *) The tmdiff.h API was so ugly and minimal that our own timing utility
1945 (speed) prefers to use its own implementation. The two implementations
1946 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
1947 its object type properly exposed (MS_TM) instead of casting to/from "char
1948 *". This may still change yet if someone realises MS_TM and "ms_time_***"
1949 aren't necessarily the greatest nomenclatures - but this is what was used
1950 internally to the implementation so I've used that for now.
1953 *) Ensure that deprecated functions do not get compiled when
1954 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
1955 the self-tests were still using deprecated key-generation functions so
1956 these have been updated also.
1959 *) Reorganise PKCS#7 code to separate the digest location functionality
1960 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
1961 New function PKCS7_set_digest() to set the digest type for PKCS#7
1962 digestedData type. Add additional code to correctly generate the
1963 digestedData type and add support for this type in PKCS7 initialization
1967 *) New function PKCS7_set0_type_other() this initializes a PKCS7
1968 structure of type "other".
1971 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
1972 sure the loop does correctly stop and breaking ("division by zero")
1973 modulus operations are not performed. The (pre-generated) prime
1974 table crypto/bn/bn_prime.h was already correct, but it could not be
1975 re-generated on some platforms because of the "division by zero"
1976 situation in the script.
1977 [Ralf S. Engelschall]
1979 *) Update support for ECC-based TLS ciphersuites according to
1980 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
1981 SHA-1 now is only used for "small" curves (where the
1982 representation of a field element takes up to 24 bytes); for
1983 larger curves, the field element resulting from ECDH is directly
1984 used as premaster secret.
1985 [Douglas Stebila (Sun Microsystems Laboratories)]
1987 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
1988 curve secp160r1 to the tests.
1989 [Douglas Stebila (Sun Microsystems Laboratories)]
1991 *) Add the possibility to load symbols globally with DSO.
1992 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
1994 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
1995 control of the error stack.
1998 *) Add support for STORE in ENGINE.
2001 *) Add the STORE type. The intention is to provide a common interface
2002 to certificate and key stores, be they simple file-based stores, or
2003 HSM-type store, or LDAP stores, or...
2004 NOTE: The code is currently UNTESTED and isn't really used anywhere.
2007 *) Add a generic structure called OPENSSL_ITEM. This can be used to
2008 pass a list of arguments to any function as well as provide a way
2009 for a function to pass data back to the caller.
2012 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
2013 works like BUF_strdup() but can be used to duplicate a portion of
2014 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
2018 *) Add the function sk_find_ex() which works like sk_find(), but will
2019 return an index to an element even if an exact match couldn't be
2020 found. The index is guaranteed to point at the element where the
2021 searched-for key would be inserted to preserve sorting order.
2024 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
2025 takes an extra flags argument for optional functionality. Currently,
2026 the following flags are defined:
2028 OBJ_BSEARCH_VALUE_ON_NOMATCH
2029 This one gets OBJ_bsearch_ex() to return a pointer to the first
2030 element where the comparing function returns a negative or zero
2033 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
2034 This one gets OBJ_bsearch_ex() to return a pointer to the first
2035 element where the comparing function returns zero. This is useful
2036 if there are more than one element where the comparing function
2040 *) Make it possible to create self-signed certificates with 'openssl ca'
2041 in such a way that the self-signed certificate becomes part of the
2042 CA database and uses the same mechanisms for serial number generation
2043 as all other certificate signing. The new flag '-selfsign' enables
2044 this functionality. Adapt CA.sh and CA.pl.in.
2047 *) Add functionality to check the public key of a certificate request
2048 against a given private. This is useful to check that a certificate
2049 request can be signed by that key (self-signing).
2052 *) Make it possible to have multiple active certificates with the same
2053 subject in the CA index file. This is done only if the keyword
2054 'unique_subject' is set to 'no' in the main CA section (default
2055 if 'CA_default') of the configuration file. The value is saved
2056 with the database itself in a separate index attribute file,
2057 named like the index file with '.attr' appended to the name.
2060 *) Generate muti valued AVAs using '+' notation in config files for
2064 *) Support for nameConstraints certificate extension.
2067 *) Support for policyConstraints certificate extension.
2070 *) Support for policyMappings certificate extension.
2073 *) Make sure the default DSA_METHOD implementation only uses its
2074 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
2075 and change its own handlers to be NULL so as to remove unnecessary
2076 indirection. This lets alternative implementations fallback to the
2077 default implementation more easily.
2080 *) Support for directoryName in GeneralName related extensions
2084 *) Make it possible to link applications using Makefile.shared.
2085 Make that possible even when linking against static libraries!
2088 *) Support for single pass processing for S/MIME signing. This now
2089 means that S/MIME signing can be done from a pipe, in addition
2090 cleartext signing (multipart/signed type) is effectively streaming
2091 and the signed data does not need to be all held in memory.
2093 This is done with a new flag PKCS7_STREAM. When this flag is set
2094 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
2095 is done after the data is output (and digests calculated) in
2096 SMIME_write_PKCS7().
2099 *) Add full support for -rpath/-R, both in shared libraries and
2100 applications, at least on the platforms where it's known how
2104 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
2105 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
2106 will now compute a table of multiples of the generator that
2107 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
2108 faster (notably in the case of a single point multiplication,
2109 scalar * generator).
2110 [Nils Larsch, Bodo Moeller]
2112 *) IPv6 support for certificate extensions. The various extensions
2113 which use the IP:a.b.c.d can now take IPv6 addresses using the
2114 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
2118 *) Added an ENGINE that implements RSA by performing private key
2119 exponentiations with the GMP library. The conversions to and from
2120 GMP's mpz_t format aren't optimised nor are any montgomery forms
2121 cached, and on x86 it appears OpenSSL's own performance has caught up.
2122 However there are likely to be other architectures where GMP could
2123 provide a boost. This ENGINE is not built in by default, but it can be
2124 specified at Configure time and should be accompanied by the necessary
2125 linker additions, eg;
2126 ./config -DOPENSSL_USE_GMP -lgmp
2129 *) "openssl engine" will not display ENGINE/DSO load failure errors when
2130 testing availability of engines with "-t" - the old behaviour is
2131 produced by increasing the feature's verbosity with "-tt".
2134 *) ECDSA routines: under certain error conditions uninitialized BN objects
2135 could be freed. Solution: make sure initialization is performed early
2136 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
2140 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
2141 and DH_METHOD (eg. by ENGINE implementations) to override the normal
2142 software implementations. For DSA and DH, parameter generation can
2143 also be overriden by providing the appropriate method callbacks.
2146 *) Change the "progress" mechanism used in key-generation and
2147 primality testing to functions that take a new BN_GENCB pointer in
2148 place of callback/argument pairs. The new API functions have "_ex"
2149 postfixes and the older functions are reimplemented as wrappers for
2150 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
2151 declarations of the old functions to help (graceful) attempts to
2152 migrate to the new functions. Also, the new key-generation API
2153 functions operate on a caller-supplied key-structure and return
2154 success/failure rather than returning a key or NULL - this is to
2155 help make "keygen" another member function of RSA_METHOD etc.
2157 Example for using the new callback interface:
2159 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
2163 BN_GENCB_set(&my_cb, my_callback, my_arg);
2165 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
2166 /* For the meaning of a, b in calls to my_callback(), see the
2167 * documentation of the function that calls the callback.
2168 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
2169 * my_callback should return 1 if it wants BN_is_prime_ex()
2170 * to continue, or 0 to stop.
2175 *) Change the ZLIB compression method to be stateful, and make it
2176 available to TLS with the number defined in
2177 draft-ietf-tls-compression-04.txt.
2180 *) Add the ASN.1 structures and functions for CertificatePair, which
2181 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
2183 CertificatePair ::= SEQUENCE {
2184 forward [0] Certificate OPTIONAL,
2185 reverse [1] Certificate OPTIONAL,
2186 -- at least one of the pair shall be present -- }
2188 Also implement the PEM functions to read and write certificate
2189 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
2191 This needed to be defined, mostly for the sake of the LDAP
2192 attribute crossCertificatePair, but may prove useful elsewhere as
2196 *) Make it possible to inhibit symlinking of shared libraries in
2197 Makefile.shared, for Cygwin's sake.
2200 *) Extend the BIGNUM API by creating a function
2201 void BN_set_negative(BIGNUM *a, int neg);
2202 and a macro that behave like
2203 int BN_is_negative(const BIGNUM *a);
2205 to avoid the need to access 'a->neg' directly in applications.
2208 *) Implement fast modular reduction for pseudo-Mersenne primes
2209 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
2210 EC_GROUP_new_curve_GFp() will now automatically use this
2212 [Nils Larsch <nla@trustcenter.de>]
2214 *) Add new lock type (CRYPTO_LOCK_BN).
2217 *) Change the ENGINE framework to automatically load engines
2218 dynamically from specific directories unless they could be
2219 found to already be built in or loaded. Move all the
2220 current engines except for the cryptodev one to a new
2222 The engines in engines/ are built as shared libraries if
2223 the "shared" options was given to ./Configure or ./config.
2224 Otherwise, they are inserted in libcrypto.a.
2225 /usr/local/ssl/engines is the default directory for dynamic
2226 engines, but that can be overriden at configure time through
2227 the usual use of --prefix and/or --openssldir, and at run
2228 time with the environment variable OPENSSL_ENGINES.
2229 [Geoff Thorpe and Richard Levitte]
2231 *) Add Makefile.shared, a helper makefile to build shared
2232 libraries. Addapt Makefile.org.
2235 *) Add version info to Win32 DLLs.
2236 [Peter 'Luna' Runestig" <peter@runestig.com>]
2238 *) Add new 'medium level' PKCS#12 API. Certificates and keys
2239 can be added using this API to created arbitrary PKCS#12
2240 files while avoiding the low level API.
2242 New options to PKCS12_create(), key or cert can be NULL and
2243 will then be omitted from the output file. The encryption
2244 algorithm NIDs can be set to -1 for no encryption, the mac
2245 iteration count can be set to 0 to omit the mac.
2247 Enhance pkcs12 utility by making the -nokeys and -nocerts
2248 options work when creating a PKCS#12 file. New option -nomac
2249 to omit the mac, NONE can be set for an encryption algorithm.
2250 New code is modified to use the enhanced PKCS12_create()
2251 instead of the low level API.
2254 *) Extend ASN1 encoder to support indefinite length constructed
2255 encoding. This can output sequences tags and octet strings in
2256 this form. Modify pk7_asn1.c to support indefinite length
2257 encoding. This is experimental and needs additional code to
2258 be useful, such as an ASN1 bio and some enhanced streaming
2261 Extend template encode functionality so that tagging is passed
2262 down to the template encoder.
2265 *) Let 'openssl req' fail if an argument to '-newkey' is not
2266 recognized instead of using RSA as a default.
2269 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
2270 As these are not official, they are not included in "ALL";
2271 the "ECCdraft" ciphersuite group alias can be used to select them.
2272 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
2274 *) Add ECDH engine support.
2275 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
2277 *) Add ECDH in new directory crypto/ecdh/.
2278 [Douglas Stebila (Sun Microsystems Laboratories)]
2280 *) Let BN_rand_range() abort with an error after 100 iterations
2281 without success (which indicates a broken PRNG).
2284 *) Change BN_mod_sqrt() so that it verifies that the input value
2285 is really the square of the return value. (Previously,
2286 BN_mod_sqrt would show GIGO behaviour.)
2289 *) Add named elliptic curves over binary fields from X9.62, SECG,
2290 and WAP/WTLS; add OIDs that were still missing.
2292 [Sheueling Chang Shantz and Douglas Stebila
2293 (Sun Microsystems Laboratories)]
2295 *) Extend the EC library for elliptic curves over binary fields
2296 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
2299 EC_GF2m_simple_method
2303 EC_GROUP_new_curve_GF2m
2304 EC_GROUP_set_curve_GF2m
2305 EC_GROUP_get_curve_GF2m
2306 EC_POINT_set_affine_coordinates_GF2m
2307 EC_POINT_get_affine_coordinates_GF2m
2308 EC_POINT_set_compressed_coordinates_GF2m
2310 Point compression for binary fields is disabled by default for
2311 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
2314 As binary polynomials are represented as BIGNUMs, various members
2315 of the EC_GROUP and EC_POINT data structures can be shared
2316 between the implementations for prime fields and binary fields;
2317 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
2318 are essentially identical to their ..._GFp counterparts.
2319 (For simplicity, the '..._GFp' prefix has been dropped from
2320 various internal method names.)
2322 An internal 'field_div' method (similar to 'field_mul' and
2323 'field_sqr') has been added; this is used only for binary fields.
2325 [Sheueling Chang Shantz and Douglas Stebila
2326 (Sun Microsystems Laboratories)]
2328 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
2329 through methods ('mul', 'precompute_mult').
2331 The generic implementations (now internally called 'ec_wNAF_mul'
2332 and 'ec_wNAF_precomputed_mult') remain the default if these
2333 methods are undefined.
2335 [Sheueling Chang Shantz and Douglas Stebila
2336 (Sun Microsystems Laboratories)]
2338 *) New function EC_GROUP_get_degree, which is defined through
2339 EC_METHOD. For curves over prime fields, this returns the bit
2340 length of the modulus.
2342 [Sheueling Chang Shantz and Douglas Stebila
2343 (Sun Microsystems Laboratories)]
2345 *) New functions EC_GROUP_dup, EC_POINT_dup.
2346 (These simply call ..._new and ..._copy).
2348 [Sheueling Chang Shantz and Douglas Stebila
2349 (Sun Microsystems Laboratories)]
2351 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
2352 Polynomials are represented as BIGNUMs (where the sign bit is not
2353 used) in the following functions [macros]:
2356 BN_GF2m_sub [= BN_GF2m_add]
2357 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
2358 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
2359 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
2361 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
2362 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
2363 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
2364 BN_GF2m_cmp [= BN_ucmp]
2366 (Note that only the 'mod' functions are actually for fields GF(2^m).
2367 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
2369 For some functions, an the irreducible polynomial defining a
2370 field can be given as an 'unsigned int[]' with strictly
2371 decreasing elements giving the indices of those bits that are set;
2372 i.e., p[] represents the polynomial
2373 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
2375 p[0] > p[1] > ... > p[k] = 0.
2376 This applies to the following functions:
2381 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
2382 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
2384 BN_GF2m_mod_sqrt_arr
2385 BN_GF2m_mod_solve_quad_arr
2389 Conversion can be performed by the following functions:
2394 bntest.c has additional tests for binary polynomial arithmetic.
2396 Two implementations for BN_GF2m_mod_div() are available.
2397 The default algorithm simply uses BN_GF2m_mod_inv() and
2398 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
2399 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
2400 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
2402 [Sheueling Chang Shantz and Douglas Stebila
2403 (Sun Microsystems Laboratories)]
2405 *) Add new error code 'ERR_R_DISABLED' that can be used when some
2406 functionality is disabled at compile-time.
2407 [Douglas Stebila <douglas.stebila@sun.com>]
2409 *) Change default behaviour of 'openssl asn1parse' so that more
2410 information is visible when viewing, e.g., a certificate:
2412 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
2413 mode the content of non-printable OCTET STRINGs is output in a
2414 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
2415 avoid the appearance of a printable string.
2416 [Nils Larsch <nla@trustcenter.de>]
2418 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
2420 EC_GROUP_set_asn1_flag()
2421 EC_GROUP_get_asn1_flag()
2422 EC_GROUP_set_point_conversion_form()
2423 EC_GROUP_get_point_conversion_form()
2424 These control ASN1 encoding details:
2425 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
2426 has been set to OPENSSL_EC_NAMED_CURVE.
2427 - Points are encoded in uncompressed form by default; options for
2428 asn1_for are as for point2oct, namely
2429 POINT_CONVERSION_COMPRESSED
2430 POINT_CONVERSION_UNCOMPRESSED
2431 POINT_CONVERSION_HYBRID
2433 Also add 'seed' and 'seed_len' members to EC_GROUP with access
2436 EC_GROUP_get0_seed()
2437 EC_GROUP_get_seed_len()
2438 This is used only for ASN1 purposes (so far).
2439 [Nils Larsch <nla@trustcenter.de>]
2441 *) Add 'field_type' member to EC_METHOD, which holds the NID
2442 of the appropriate field type OID. The new function
2443 EC_METHOD_get_field_type() returns this value.
2444 [Nils Larsch <nla@trustcenter.de>]
2449 EC_POINT_point2hex()
2450 EC_POINT_hex2point()
2451 providing useful interfaces to EC_POINT_point2oct() and
2452 EC_POINT_oct2point().
2453 [Nils Larsch <nla@trustcenter.de>]
2455 *) Change internals of the EC library so that the functions
2456 EC_GROUP_set_generator()
2457 EC_GROUP_get_generator()
2458 EC_GROUP_get_order()
2459 EC_GROUP_get_cofactor()
2460 are implemented directly in crypto/ec/ec_lib.c and not dispatched
2461 to methods, which would lead to unnecessary code duplication when
2462 adding different types of curves.
2463 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
2465 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
2466 arithmetic, and such that modified wNAFs are generated
2467 (which avoid length expansion in many cases).
2470 *) Add a function EC_GROUP_check_discriminant() (defined via
2471 EC_METHOD) that verifies that the curve discriminant is non-zero.
2473 Add a function EC_GROUP_check() that makes some sanity tests
2474 on a EC_GROUP, its generator and order. This includes
2475 EC_GROUP_check_discriminant().
2476 [Nils Larsch <nla@trustcenter.de>]
2478 *) Add ECDSA in new directory crypto/ecdsa/.
2480 Add applications 'openssl ecparam' and 'openssl ecdsa'
2481 (these are based on 'openssl dsaparam' and 'openssl dsa').
2483 ECDSA support is also included in various other files across the
2484 library. Most notably,
2485 - 'openssl req' now has a '-newkey ecdsa:file' option;
2486 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
2487 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
2488 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
2489 them suitable for ECDSA where domain parameters must be
2490 extracted before the specific public key;
2491 - ECDSA engine support has been added.
2492 [Nils Larsch <nla@trustcenter.de>]
2494 *) Include some named elliptic curves, and add OIDs from X9.62,
2495 SECG, and WAP/WTLS. Each curve can be obtained from the new
2497 EC_GROUP_new_by_curve_name(),
2498 and the list of available named curves can be obtained with
2499 EC_get_builtin_curves().
2500 Also add a 'curve_name' member to EC_GROUP objects, which can be
2502 EC_GROUP_set_curve_name()
2503 EC_GROUP_get_curve_name()
2504 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
2506 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2507 was actually never needed) and in BN_mul(). The removal in BN_mul()
2508 required a small change in bn_mul_part_recursive() and the addition
2509 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2510 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2511 bn_sub_words() and bn_add_words() except they take arrays with
2515 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
2517 *) Cleanse PEM buffers before freeing them since they may contain
2519 [Benjamin Bennett <ben@psc.edu>]
2521 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2522 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2523 authentication-only ciphersuites.
2526 *) Since AES128 and AES256 share a single mask bit in the logic of
2527 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2528 kludge to work properly if AES128 is available and AES256 isn't.
2531 *) Expand security boundary to match 1.1.1 module.
2534 *) Remove redundant features: hash file source, editing of test vectors
2535 modify fipsld to use external fips_premain.c signature.
2538 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
2539 run algorithm test programs.
2542 *) Make algorithm test programs more tolerant of whitespace.
2545 *) Have SSL/TLS server implementation tolerate "mismatched" record
2546 protocol version while receiving ClientHello even if the
2547 ClientHello is fragmented. (The server can't insist on the
2548 particular protocol version it has chosen before the ServerHello
2549 message has informed the client about his choice.)
2552 *) Load error codes if they are not already present instead of using a
2553 static variable. This allows them to be cleanly unloaded and reloaded.
2556 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
2558 *) Introduce limits to prevent malicious keys being able to
2559 cause a denial of service. (CVE-2006-2940)
2560 [Steve Henson, Bodo Moeller]
2562 *) Fix ASN.1 parsing of certain invalid structures that can result
2563 in a denial of service. (CVE-2006-2937) [Steve Henson]
2565 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2566 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2568 *) Fix SSL client code which could crash if connecting to a
2569 malicious SSLv2 server. (CVE-2006-4343)
2570 [Tavis Ormandy and Will Drewry, Google Security Team]
2572 *) Change ciphersuite string processing so that an explicit
2573 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
2574 will no longer include "AES128-SHA"), and any other similar
2575 ciphersuite (same bitmap) from *other* protocol versions (so that
2576 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
2577 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
2578 changes from 0.9.8b and 0.9.8d.
2581 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
2583 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2584 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2586 *) Change the Unix randomness entropy gathering to use poll() when
2587 possible instead of select(), since the latter has some
2588 undesirable limitations.
2589 [Darryl Miles via Richard Levitte and Bodo Moeller]
2591 *) Disable rogue ciphersuites:
2593 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2594 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2595 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2597 The latter two were purportedly from
2598 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2601 Also deactive the remaining ciphersuites from
2602 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2603 unofficial, and the ID has long expired.
2606 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2607 dual-core machines) and other potential thread-safety issues.
2610 Changes between 0.9.7i and 0.9.7j [04 May 2006]
2612 *) Adapt fipsld and the build system to link against the validated FIPS
2613 module in FIPS mode.
2616 *) Fixes for VC++ 2005 build under Windows.
2619 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
2620 from a Windows bash shell such as MSYS. It is autodetected from the
2621 "config" script when run from a VC++ environment. Modify standard VC++
2622 build to use fipscanister.o from the GNU make build.
2625 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
2627 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
2628 The value now differs depending on if you build for FIPS or not.
2629 BEWARE! A program linked with a shared FIPSed libcrypto can't be
2630 safely run with a non-FIPSed libcrypto, as it may crash because of
2631 the difference induced by this change.
2634 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
2636 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2637 (part of SSL_OP_ALL). This option used to disable the
2638 countermeasure against man-in-the-middle protocol-version
2639 rollback in the SSL 2.0 server implementation, which is a bad
2640 idea. (CVE-2005-2969)
2642 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2643 for Information Security, National Institute of Advanced Industrial
2644 Science and Technology [AIST], Japan)]
2646 *) Minimal support for X9.31 signatures and PSS padding modes. This is
2647 mainly for FIPS compliance and not fully integrated at this stage.
2650 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
2651 the exponentiation using a fixed-length exponent. (Otherwise,
2652 the information leaked through timing could expose the secret key
2653 after many signatures; cf. Bleichenbacher's attack on DSA with
2657 *) Make a new fixed-window mod_exp implementation the default for
2658 RSA, DSA, and DH private-key operations so that the sequence of
2659 squares and multiplies and the memory access pattern are
2660 independent of the particular secret key. This will mitigate
2661 cache-timing and potential related attacks.
2663 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
2664 and this is automatically used by BN_mod_exp_mont() if the new flag
2665 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
2666 will use this BN flag for private exponents unless the flag
2667 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
2668 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
2670 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
2672 *) Change the client implementation for SSLv23_method() and
2673 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
2674 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
2675 (Previously, the SSL 2.0 backwards compatible Client Hello
2676 message format would be used even with SSL_OP_NO_SSLv2.)
2679 *) Add support for smime-type MIME parameter in S/MIME messages which some
2683 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
2684 a threadsafe manner. Modify rsa code to use new function and add calls
2685 to dsa and dh code (which had race conditions before).
2688 *) Include the fixed error library code in the C error file definitions
2689 instead of fixing them up at runtime. This keeps the error code
2690 structures constant.
2693 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
2695 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
2698 *) Fixes for newer kerberos headers. NB: the casts are needed because
2699 the 'length' field is signed on one version and unsigned on another
2700 with no (?) obvious way to tell the difference, without these VC++
2701 complains. Also the "definition" of FAR (blank) is no longer included
2702 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
2703 some needed definitions.
2706 *) Undo Cygwin change.
2709 *) Added support for proxy certificates according to RFC 3820.
2710 Because they may be a security thread to unaware applications,
2711 they must be explicitely allowed in run-time. See
2712 docs/HOWTO/proxy_certificates.txt for further information.
2715 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
2717 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
2718 server and client random values. Previously
2719 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
2720 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
2722 This change has negligible security impact because:
2724 1. Server and client random values still have 24 bytes of pseudo random
2727 2. Server and client random values are sent in the clear in the initial
2730 3. The master secret is derived using the premaster secret (48 bytes in
2731 size for static RSA ciphersuites) as well as client server and random
2734 The OpenSSL team would like to thank the UK NISCC for bringing this issue
2737 [Stephen Henson, reported by UK NISCC]
2739 *) Use Windows randomness collection on Cygwin.
2742 *) Fix hang in EGD/PRNGD query when communication socket is closed
2743 prematurely by EGD/PRNGD.
2744 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
2746 *) Prompt for pass phrases when appropriate for PKCS12 input format.
2749 *) Back-port of selected performance improvements from development
2750 branch, as well as improved support for PowerPC platforms.
2753 *) Add lots of checks for memory allocation failure, error codes to indicate
2754 failure and freeing up memory if a failure occurs.
2755 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
2757 *) Add new -passin argument to dgst.
2760 *) Perform some character comparisons of different types in X509_NAME_cmp:
2761 this is needed for some certificates that reencode DNs into UTF8Strings
2762 (in violation of RFC3280) and can't or wont issue name rollover
2766 *) Make an explicit check during certificate validation to see that
2767 the CA setting in each certificate on the chain is correct. As a
2768 side effect always do the following basic checks on extensions,
2769 not just when there's an associated purpose to the check:
2771 - if there is an unhandled critical extension (unless the user
2772 has chosen to ignore this fault)
2773 - if the path length has been exceeded (if one is set at all)
2774 - that certain extensions fit the associated purpose (if one has
2778 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
2780 *) Avoid a race condition when CRLs are checked in a multi threaded
2781 environment. This would happen due to the reordering of the revoked
2782 entries during signature checking and serial number lookup. Now the
2783 encoding is cached and the serial number sort performed under a lock.
2784 Add new STACK function sk_is_sorted().
2787 *) Add Delta CRL to the extension code.
2790 *) Various fixes to s3_pkt.c so alerts are sent properly.
2791 [David Holmes <d.holmes@f5.com>]
2793 *) Reduce the chances of duplicate issuer name and serial numbers (in
2794 violation of RFC3280) using the OpenSSL certificate creation utilities.
2795 This is done by creating a random 64 bit value for the initial serial
2796 number when a serial number file is created or when a self signed
2797 certificate is created using 'openssl req -x509'. The initial serial
2798 number file is created using 'openssl x509 -next_serial' in CA.pl
2799 rather than being initialized to 1.
2802 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
2804 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
2805 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
2806 [Joe Orton, Steve Henson]
2808 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
2810 [Joe Orton, Steve Henson]
2812 *) Make it possible to have multiple active certificates with the same
2813 subject in the CA index file. This is done only if the keyword
2814 'unique_subject' is set to 'no' in the main CA section (default
2815 if 'CA_default') of the configuration file. The value is saved
2816 with the database itself in a separate index attribute file,
2817 named like the index file with '.attr' appended to the name.
2820 *) X509 verify fixes. Disable broken certificate workarounds when
2821 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
2822 keyUsage extension present. Don't accept CRLs with unhandled critical
2823 extensions: since verify currently doesn't process CRL extensions this
2824 rejects a CRL with *any* critical extensions. Add new verify error codes
2828 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
2829 A clarification of RFC2560 will require the use of OCTET STRINGs and
2830 some implementations cannot handle the current raw format. Since OpenSSL
2831 copies and compares OCSP nonces as opaque blobs without any attempt at
2832 parsing them this should not create any compatibility issues.
2835 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
2836 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
2837 this HMAC (and other) operations are several times slower than OpenSSL
2841 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
2842 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2844 *) Use the correct content when signing type "other".
2847 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
2849 *) Fix various bugs revealed by running the NISCC test suite:
2851 Stop out of bounds reads in the ASN1 code when presented with
2852 invalid tags (CVE-2003-0543 and CVE-2003-0544).
2854 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
2856 If verify callback ignores invalid public key errors don't try to check
2857 certificate signature with the NULL public key.
2861 *) New -ignore_err option in ocsp application to stop the server
2862 exiting on the first error in a request.
2865 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
2866 if the server requested one: as stated in TLS 1.0 and SSL 3.0
2870 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
2871 extra data after the compression methods not only for TLS 1.0
2872 but also for SSL 3.0 (as required by the specification).
2873 [Bodo Moeller; problem pointed out by Matthias Loepfe]
2875 *) Change X509_certificate_type() to mark the key as exported/exportable
2876 when it's 512 *bits* long, not 512 bytes.
2879 *) Change AES_cbc_encrypt() so it outputs exact multiple of
2880 blocks during encryption.
2883 *) Various fixes to base64 BIO and non blocking I/O. On write
2884 flushes were not handled properly if the BIO retried. On read
2885 data was not being buffered properly and had various logic bugs.
2886 This also affects blocking I/O when the data being decoded is a
2890 *) Various S/MIME bugfixes and compatibility changes:
2891 output correct application/pkcs7 MIME type if
2892 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
2893 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
2894 of files as .eml work). Correctly handle very long lines in MIME
2898 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
2900 *) Countermeasure against the Klima-Pokorny-Rosa extension of
2901 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
2902 a protocol version number mismatch like a decryption error
2903 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
2906 *) Turn on RSA blinding by default in the default implementation
2907 to avoid a timing attack. Applications that don't want it can call
2908 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
2909 They would be ill-advised to do so in most cases.
2910 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
2912 *) Change RSA blinding code so that it works when the PRNG is not
2913 seeded (in this case, the secret RSA exponent is abused as
2914 an unpredictable seed -- if it is not unpredictable, there
2915 is no point in blinding anyway). Make RSA blinding thread-safe
2916 by remembering the creator's thread ID in rsa->blinding and
2917 having all other threads use local one-time blinding factors
2918 (this requires more computation than sharing rsa->blinding, but
2919 avoids excessive locking; and if an RSA object is not shared
2920 between threads, blinding will still be very fast).
2923 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
2924 ENGINE as defaults for all supported algorithms irrespective of
2925 the 'flags' parameter. 'flags' is now honoured, so applications
2926 should make sure they are passing it correctly.
2929 *) Target "mingw" now allows native Windows code to be generated in
2930 the Cygwin environment as well as with the MinGW compiler.
2933 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
2935 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
2936 via timing by performing a MAC computation even if incorrrect
2937 block cipher padding has been found. This is a countermeasure
2938 against active attacks where the attacker has to distinguish
2939 between bad padding and a MAC verification error. (CVE-2003-0078)
2941 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
2942 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
2943 Martin Vuagnoux (EPFL, Ilion)]
2945 *) Make the no-err option work as intended. The intention with no-err
2946 is not to have the whole error stack handling routines removed from
2947 libcrypto, it's only intended to remove all the function name and
2948 reason texts, thereby removing some of the footprint that may not
2949 be interesting if those errors aren't displayed anyway.
2951 NOTE: it's still possible for any application or module to have it's
2952 own set of error texts inserted. The routines are there, just not
2953 used by default when no-err is given.
2956 *) Add support for FreeBSD on IA64.
2957 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
2959 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
2960 Kerberos function mit_des_cbc_cksum(). Before this change,
2961 the value returned by DES_cbc_cksum() was like the one from
2962 mit_des_cbc_cksum(), except the bytes were swapped.
2963 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
2965 *) Allow an application to disable the automatic SSL chain building.
2966 Before this a rather primitive chain build was always performed in
2967 ssl3_output_cert_chain(): an application had no way to send the
2968 correct chain if the automatic operation produced an incorrect result.
2970 Now the chain builder is disabled if either:
2972 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
2974 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
2976 The reasoning behind this is that an application would not want the
2977 auto chain building to take place if extra chain certificates are
2978 present and it might also want a means of sending no additional
2979 certificates (for example the chain has two certificates and the
2983 *) Add the possibility to build without the ENGINE framework.
2984 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2986 *) Under Win32 gmtime() can return NULL: check return value in
2987 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
2990 *) DSA routines: under certain error conditions uninitialized BN objects
2991 could be freed. Solution: make sure initialization is performed early
2992 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
2993 Nils Larsch <nla@trustcenter.de> via PR#459)
2996 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
2997 checked on reconnect on the client side, therefore session resumption
2998 could still fail with a "ssl session id is different" error. This
2999 behaviour is masked when SSL_OP_ALL is used due to
3000 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
3001 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3002 followup to PR #377.
3005 *) IA-32 assembler support enhancements: unified ELF targets, support
3006 for SCO/Caldera platforms, fix for Cygwin shared build.
3009 *) Add support for FreeBSD on sparc64. As a consequence, support for
3010 FreeBSD on non-x86 processors is separate from x86 processors on
3011 the config script, much like the NetBSD support.
3012 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
3014 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
3016 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
3019 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
3020 code (06) was taken as the first octet of the session ID and the last
3021 octet was ignored consequently. As a result SSLv2 client side session
3022 caching could not have worked due to the session ID mismatch between
3024 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3028 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
3029 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
3033 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
3034 seems that in spite of existing for more than a year, many application
3035 author have done nothing to provide the necessary callbacks, which
3036 means that this particular engine will not work properly anywhere.
3037 This is a very unfortunate situation which forces us, in the name
3038 of usability, to give the hw_ncipher.c a static lock, which is part
3040 NOTE: This is for the 0.9.7 series ONLY. This hack will never
3041 appear in 0.9.8 or later. We EXPECT application authors to have
3042 dealt properly with this when 0.9.8 is released (unless we actually
3043 make such changes in the libcrypto locking code that changes will
3044 have to be made anyway).
3047 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
3048 octets have been read, EOF or an error occurs. Without this change
3049 some truncated ASN1 structures will not produce an error.
3052 *) Disable Heimdal support, since it hasn't been fully implemented.
3053 Still give the possibility to force the use of Heimdal, but with
3054 warnings and a request that patches get sent to openssl-dev.
3057 *) Add the VC-CE target, introduce the WINCE sysname, and add
3058 INSTALL.WCE and appropriate conditionals to make it build.
3059 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3061 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
3062 cygssl-x.y.z.dll, where x, y and z are the major, minor and
3063 edit numbers of the version.
3064 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3066 *) Introduce safe string copy and catenation functions
3067 (BUF_strlcpy() and BUF_strlcat()).
3068 [Ben Laurie (CHATS) and Richard Levitte]
3070 *) Avoid using fixed-size buffers for one-line DNs.
3071 [Ben Laurie (CHATS)]
3073 *) Add BUF_MEM_grow_clean() to avoid information leakage when
3074 resizing buffers containing secrets, and use where appropriate.
3075 [Ben Laurie (CHATS)]
3077 *) Avoid using fixed size buffers for configuration file location.
3078 [Ben Laurie (CHATS)]
3080 *) Avoid filename truncation for various CA files.
3081 [Ben Laurie (CHATS)]
3083 *) Use sizeof in preference to magic numbers.
3084 [Ben Laurie (CHATS)]
3086 *) Avoid filename truncation in cert requests.
3087 [Ben Laurie (CHATS)]
3089 *) Add assertions to check for (supposedly impossible) buffer
3091 [Ben Laurie (CHATS)]
3093 *) Don't cache truncated DNS entries in the local cache (this could
3094 potentially lead to a spoofing attack).
3095 [Ben Laurie (CHATS)]
3097 *) Fix various buffers to be large enough for hex/decimal
3098 representations in a platform independent manner.
3099 [Ben Laurie (CHATS)]
3101 *) Add CRYPTO_realloc_clean() to avoid information leakage when
3102 resizing buffers containing secrets, and use where appropriate.
3103 [Ben Laurie (CHATS)]
3105 *) Add BIO_indent() to avoid much slightly worrying code to do
3107 [Ben Laurie (CHATS)]
3109 *) Convert sprintf()/BIO_puts() to BIO_printf().
3110 [Ben Laurie (CHATS)]
3112 *) buffer_gets() could terminate with the buffer only half
3114 [Ben Laurie (CHATS)]
3116 *) Add assertions to prevent user-supplied crypto functions from
3117 overflowing internal buffers by having large block sizes, etc.
3118 [Ben Laurie (CHATS)]
3120 *) New OPENSSL_assert() macro (similar to assert(), but enabled
3122 [Ben Laurie (CHATS)]
3124 *) Eliminate unused copy of key in RC4.
3125 [Ben Laurie (CHATS)]
3127 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
3128 [Ben Laurie (CHATS)]
3130 *) Fix off-by-one error in EGD path.
3131 [Ben Laurie (CHATS)]
3133 *) If RANDFILE path is too long, ignore instead of truncating.
3134 [Ben Laurie (CHATS)]
3136 *) Eliminate unused and incorrectly sized X.509 structure
3138 [Ben Laurie (CHATS)]
3140 *) Eliminate unused and dangerous function knumber().
3141 [Ben Laurie (CHATS)]
3143 *) Eliminate unused and dangerous structure, KSSL_ERR.
3144 [Ben Laurie (CHATS)]
3146 *) Protect against overlong session ID context length in an encoded
3147 session object. Since these are local, this does not appear to be
3149 [Ben Laurie (CHATS)]
3151 *) Change from security patch (see 0.9.6e below) that did not affect
3152 the 0.9.6 release series:
3154 Remote buffer overflow in SSL3 protocol - an attacker could
3155 supply an oversized master key in Kerberos-enabled versions.
3157 [Ben Laurie (CHATS)]
3159 *) Change the SSL kerb5 codes to match RFC 2712.
3162 *) Make -nameopt work fully for req and add -reqopt switch.
3163 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
3165 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
3166 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
3168 *) Make sure tests can be performed even if the corresponding algorithms
3169 have been removed entirely. This was also the last step to make
3170 OpenSSL compilable with DJGPP under all reasonable conditions.
3171 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
3173 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
3174 to allow version independent disabling of normally unselected ciphers,
3175 which may be activated as a side-effect of selecting a single cipher.
3177 (E.g., cipher list string "RSA" enables ciphersuites that are left
3178 out of "ALL" because they do not provide symmetric encryption.
3179 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
3180 [Lutz Jaenicke, Bodo Moeller]
3182 *) Add appropriate support for separate platform-dependent build
3183 directories. The recommended way to make a platform-dependent
3184 build directory is the following (tested on Linux), maybe with
3187 # Place yourself outside of the OpenSSL source tree. In
3188 # this example, the environment variable OPENSSL_SOURCE
3189 # is assumed to contain the absolute OpenSSL source directory.
3190 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
3191 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
3192 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
3193 mkdir -p `dirname $F`
3194 ln -s $OPENSSL_SOURCE/$F $F
3197 To be absolutely sure not to disturb the source tree, a "make clean"
3198 is a good thing. If it isn't successfull, don't worry about it,
3199 it probably means the source directory is very clean.
3202 *) Make sure any ENGINE control commands make local copies of string
3203 pointers passed to them whenever necessary. Otherwise it is possible
3204 the caller may have overwritten (or deallocated) the original string
3205 data when a later ENGINE operation tries to use the stored values.
3206 [Götz Babin-Ebell <babinebell@trustcenter.de>]
3208 *) Improve diagnostics in file reading and command-line digests.
3209 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
3211 *) Add AES modes CFB and OFB to the object database. Correct an
3212 error in AES-CFB decryption.
3215 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
3216 allows existing EVP_CIPHER_CTX structures to be reused after
3217 calling EVP_*Final(). This behaviour is used by encryption
3218 BIOs and some applications. This has the side effect that
3219 applications must explicitly clean up cipher contexts with
3220 EVP_CIPHER_CTX_cleanup() or they will leak memory.
3223 *) Check the values of dna and dnb in bn_mul_recursive before calling
3224 bn_mul_comba (a non zero value means the a or b arrays do not contain
3225 n2 elements) and fallback to bn_mul_normal if either is not zero.
3228 *) Fix escaping of non-ASCII characters when using the -subj option
3229 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
3232 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
3233 form for "surname", serialNumber has no short form.
3234 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
3235 therefore remove "mail" short name for "internet 7".
3236 The OID for unique identifiers in X509 certificates is
3237 x500UniqueIdentifier, not uniqueIdentifier.
3238 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
3241 *) Add an "init" command to the ENGINE config module and auto initialize
3242 ENGINEs. Without any "init" command the ENGINE will be initialized
3243 after all ctrl commands have been executed on it. If init=1 the
3244 ENGINE is initailized at that point (ctrls before that point are run
3245 on the uninitialized ENGINE and after on the initialized one). If
3246 init=0 then the ENGINE will not be iniatialized at all.
3249 *) Fix the 'app_verify_callback' interface so that the user-defined
3250 argument is actually passed to the callback: In the
3251 SSL_CTX_set_cert_verify_callback() prototype, the callback
3252 declaration has been changed from
3255 int (*cb)(X509_STORE_CTX *,void *);
3256 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
3257 i=s->ctx->app_verify_callback(&ctx)
3258 has been changed into
3259 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
3261 To update applications using SSL_CTX_set_cert_verify_callback(),
3262 a dummy argument can be added to their callback functions.
3263 [D. K. Smetters <smetters@parc.xerox.com>]
3265 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
3266 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
3268 *) Add and OPENSSL_LOAD_CONF define which will cause
3269 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
3270 This allows older applications to transparently support certain
3271 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
3272 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
3273 load the config file and OPENSSL_add_all_algorithms_conf() which will
3274 always load it have also been added.
3277 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
3278 Adjust NIDs and EVP layer.
3279 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3281 *) Config modules support in openssl utility.
3283 Most commands now load modules from the config file,
3284 though in a few (such as version) this isn't done
3285 because it couldn't be used for anything.
3287 In the case of ca and req the config file used is
3288 the same as the utility itself: that is the -config
3289 command line option can be used to specify an
3293 *) Move default behaviour from OPENSSL_config(). If appname is NULL
3294 use "openssl_conf" if filename is NULL use default openssl config file.
3297 *) Add an argument to OPENSSL_config() to allow the use of an alternative
3298 config section name. Add a new flag to tolerate a missing config file
3299 and move code to CONF_modules_load_file().
3302 *) Support for crypto accelerator cards from Accelerated Encryption
3303 Processing, www.aep.ie. (Use engine 'aep')
3304 The support was copied from 0.9.6c [engine] and adapted/corrected
3305 to work with the new engine framework.
3306 [AEP Inc. and Richard Levitte]
3308 *) Support for SureWare crypto accelerator cards from Baltimore
3309 Technologies. (Use engine 'sureware')
3310 The support was copied from 0.9.6c [engine] and adapted
3311 to work with the new engine framework.
3314 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
3315 make the newer ENGINE framework commands for the CHIL engine work.
3316 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
3318 *) Make it possible to produce shared libraries on ReliantUNIX.
3319 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
3321 *) Add the configuration target debug-linux-ppro.
3322 Make 'openssl rsa' use the general key loading routines
3323 implemented in apps.c, and make those routines able to
3324 handle the key format FORMAT_NETSCAPE and the variant
3326 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3328 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3329 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3331 *) Add -keyform to rsautl, and document -engine.
3332 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
3334 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
3335 BIO_R_NO_SUCH_FILE error code rather than the generic
3336 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
3339 *) Add new functions
3341 ERR_peek_last_error_line
3342 ERR_peek_last_error_line_data.
3343 These are similar to
3346 ERR_peek_error_line_data,
3347 but report on the latest error recorded rather than the first one
3348 still in the error queue.
3349 [Ben Laurie, Bodo Moeller]
3351 *) default_algorithms option in ENGINE config module. This allows things
3353 default_algorithms = ALL
3354 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
3357 *) Prelminary ENGINE config module.
3360 *) New experimental application configuration code.
3363 *) Change the AES code to follow the same name structure as all other
3364 symmetric ciphers, and behave the same way. Move everything to
3365 the directory crypto/aes, thereby obsoleting crypto/rijndael.
3366 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3368 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
3369 [Ben Laurie and Theo de Raadt]
3371 *) Add option to output public keys in req command.
3372 [Massimiliano Pala madwolf@openca.org]
3374 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
3375 (up to about 10% better than before for P-192 and P-224).
3378 *) New functions/macros
3380 SSL_CTX_set_msg_callback(ctx, cb)
3381 SSL_CTX_set_msg_callback_arg(ctx, arg)
3382 SSL_set_msg_callback(ssl, cb)
3383 SSL_set_msg_callback_arg(ssl, arg)
3385 to request calling a callback function
3387 void cb(int write_p, int version, int content_type,
3388 const void *buf, size_t len, SSL *ssl, void *arg)
3390 whenever a protocol message has been completely received
3391 (write_p == 0) or sent (write_p == 1). Here 'version' is the
3392 protocol version according to which the SSL library interprets
3393 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
3394 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
3395 the content type as defined in the SSL 3.0/TLS 1.0 protocol
3396 specification (change_cipher_spec(20), alert(21), handshake(22)).
3397 'buf' and 'len' point to the actual message, 'ssl' to the
3398 SSL object, and 'arg' is the application-defined value set by
3399 SSL[_CTX]_set_msg_callback_arg().
3401 'openssl s_client' and 'openssl s_server' have new '-msg' options
3402 to enable a callback that displays all protocol messages.
3405 *) Change the shared library support so shared libraries are built as
3406 soon as the corresponding static library is finished, and thereby get
3407 openssl and the test programs linked against the shared library.
3408 This still only happens when the keyword "shard" has been given to
3409 the configuration scripts.
3411 NOTE: shared library support is still an experimental thing, and
3412 backward binary compatibility is still not guaranteed.
3413 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
3415 *) Add support for Subject Information Access extension.
3416 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3418 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
3419 additional bytes when new memory had to be allocated, not just
3420 when reusing an existing buffer.
3423 *) New command line and configuration option 'utf8' for the req command.
3424 This allows field values to be specified as UTF8 strings.
3427 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
3428 runs for the former and machine-readable output for the latter.
3431 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
3432 of the e-mail address in the DN (i.e., it will go into a certificate
3433 extension only). The new configuration file option 'email_in_dn = no'
3434 has the same effect.
3435 [Massimiliano Pala madwolf@openca.org]
3437 *) Change all functions with names starting with des_ to be starting
3438 with DES_ instead. Add wrappers that are compatible with libdes,
3439 but are named _ossl_old_des_*. Finally, add macros that map the
3440 des_* symbols to the corresponding _ossl_old_des_* if libdes
3441 compatibility is desired. If OpenSSL 0.9.6c compatibility is
3442 desired, the des_* symbols will be mapped to DES_*, with one
3445 Since we provide two compatibility mappings, the user needs to
3446 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
3447 compatibility is desired. The default (i.e., when that macro
3448 isn't defined) is OpenSSL 0.9.6c compatibility.
3450 There are also macros that enable and disable the support of old
3451 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
3452 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
3453 are defined, the default will apply: to support the old des routines.
3455 In either case, one must include openssl/des.h to get the correct
3456 definitions. Do not try to just include openssl/des_old.h, that
3459 NOTE: This is a major break of an old API into a new one. Software
3460 authors are encouraged to switch to the DES_ style functions. Some
3461 time in the future, des_old.h and the libdes compatibility functions
3462 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
3463 default), and then completely removed.
3466 *) Test for certificates which contain unsupported critical extensions.
3467 If such a certificate is found during a verify operation it is
3468 rejected by default: this behaviour can be overridden by either
3469 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
3470 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
3471 X509_supported_extension() has also been added which returns 1 if a
3472 particular extension is supported.
3475 *) Modify the behaviour of EVP cipher functions in similar way to digests
3476 to retain compatibility with existing code.
3479 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
3480 compatibility with existing code. In particular the 'ctx' parameter does
3481 not have to be to be initialized before the call to EVP_DigestInit() and
3482 it is tidied up after a call to EVP_DigestFinal(). New function
3483 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
3484 EVP_MD_CTX_copy() changed to not require the destination to be
3485 initialized valid and new function EVP_MD_CTX_copy_ex() added which
3486 requires the destination to be valid.
3488 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
3489 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
3492 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
3493 so that complete 'Handshake' protocol structures are kept in memory
3494 instead of overwriting 'msg_type' and 'length' with 'body' data.
3497 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
3498 [Massimo Santin via Richard Levitte]
3500 *) Major restructuring to the underlying ENGINE code. This includes
3501 reduction of linker bloat, separation of pure "ENGINE" manipulation
3502 (initialisation, etc) from functionality dealing with implementations
3503 of specific crypto iterfaces. This change also introduces integrated
3504 support for symmetric ciphers and digest implementations - so ENGINEs
3505 can now accelerate these by providing EVP_CIPHER and EVP_MD
3506 implementations of their own. This is detailed in crypto/engine/README
3507 as it couldn't be adequately described here. However, there are a few
3508 API changes worth noting - some RSA, DSA, DH, and RAND functions that
3509 were changed in the original introduction of ENGINE code have now
3510 reverted back - the hooking from this code to ENGINE is now a good
3511 deal more passive and at run-time, operations deal directly with
3512 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
3513 dereferencing through an ENGINE pointer any more. Also, the ENGINE
3514 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
3515 they were not being used by the framework as there is no concept of a
3516 BIGNUM_METHOD and they could not be generalised to the new
3517 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
3518 ENGINE_cpy() has been removed as it cannot be consistently defined in
3522 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
3525 *) Change mkdef.pl to sort symbols that get the same entry number,
3526 and make sure the automatically generated functions ERR_load_*
3527 become part of libeay.num as well.
3530 *) New function SSL_renegotiate_pending(). This returns true once
3531 renegotiation has been requested (either SSL_renegotiate() call
3532 or HelloRequest/ClientHello receveived from the peer) and becomes
3533 false once a handshake has been completed.
3534 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
3535 sends a HelloRequest, but does not ensure that a handshake takes
3536 place. SSL_renegotiate_pending() is useful for checking if the
3537 client has followed the request.)
3540 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
3541 By default, clients may request session resumption even during
3542 renegotiation (if session ID contexts permit); with this option,
3543 session resumption is possible only in the first handshake.
3545 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
3546 more bits available for options that should not be part of
3547 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
3550 *) Add some demos for certificate and certificate request creation.
3553 *) Make maximum certificate chain size accepted from the peer application
3554 settable (SSL*_get/set_max_cert_list()), as proposed by
3555 "Douglas E. Engert" <deengert@anl.gov>.
3558 *) Add support for shared libraries for Unixware-7
3559 (Boyd Lynn Gerber <gerberb@zenez.com>).
3562 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
3563 be done prior to destruction. Use this to unload error strings from
3564 ENGINEs that load their own error strings. NB: This adds two new API
3565 functions to "get" and "set" this destroy handler in an ENGINE.
3568 *) Alter all existing ENGINE implementations (except "openssl" and
3569 "openbsd") to dynamically instantiate their own error strings. This
3570 makes them more flexible to be built both as statically-linked ENGINEs
3571 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
3572 Also, add stub code to each that makes building them as self-contained
3573 shared-libraries easier (see README.ENGINE).
3576 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
3577 implementations into applications that are completely implemented in
3578 self-contained shared-libraries. The "dynamic" ENGINE exposes control
3579 commands that can be used to configure what shared-library to load and
3580 to control aspects of the way it is handled. Also, made an update to
3581 the README.ENGINE file that brings its information up-to-date and
3582 provides some information and instructions on the "dynamic" ENGINE
3583 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
3586 *) Make it possible to unload ranges of ERR strings with a new
3587 "ERR_unload_strings" function.
3590 *) Add a copy() function to EVP_MD.
3593 *) Make EVP_MD routines take a context pointer instead of just the
3594 md_data void pointer.
3597 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
3598 that the digest can only process a single chunk of data
3599 (typically because it is provided by a piece of
3600 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
3601 is only going to provide a single chunk of data, and hence the
3602 framework needn't accumulate the data for oneshot drivers.
3605 *) As with "ERR", make it possible to replace the underlying "ex_data"
3606 functions. This change also alters the storage and management of global
3607 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
3608 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
3609 index counters. The API functions that use this state have been changed
3610 to take a "class_index" rather than pointers to the class's local STACK
3611 and counter, and there is now an API function to dynamically create new
3612 classes. This centralisation allows us to (a) plug a lot of the
3613 thread-safety problems that existed, and (b) makes it possible to clean
3614 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
3615 such data would previously have always leaked in application code and
3616 workarounds were in place to make the memory debugging turn a blind eye
3617 to it. Application code that doesn't use this new function will still
3618 leak as before, but their memory debugging output will announce it now
3619 rather than letting it slide.
3621 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
3622 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
3623 has a return value to indicate success or failure.
3626 *) Make it possible to replace the underlying "ERR" functions such that the
3627 global state (2 LHASH tables and 2 locks) is only used by the "default"
3628 implementation. This change also adds two functions to "get" and "set"
3629 the implementation prior to it being automatically set the first time
3630 any other ERR function takes place. Ie. an application can call "get",
3631 pass the return value to a module it has just loaded, and that module
3632 can call its own "set" function using that value. This means the
3633 module's "ERR" operations will use (and modify) the error state in the
3634 application and not in its own statically linked copy of OpenSSL code.
3637 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
3638 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
3639 the operation, and provides a more encapsulated way for external code
3640 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
3641 to use these functions rather than manually incrementing the counts.
3643 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
3646 *) Add EVP test program.
3649 *) Add symmetric cipher support to ENGINE. Expect the API to change!
3652 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
3653 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
3654 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
3655 These allow a CRL to be built without having to access X509_CRL fields
3656 directly. Modify 'ca' application to use new functions.
3659 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
3660 bug workarounds. Rollback attack detection is a security feature.
3661 The problem will only arise on OpenSSL servers when TLSv1 is not
3662 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
3663 Software authors not wanting to support TLSv1 will have special reasons
3664 for their choice and can explicitly enable this option.
3665 [Bodo Moeller, Lutz Jaenicke]
3667 *) Rationalise EVP so it can be extended: don't include a union of
3668 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
3669 (similar to those existing for EVP_CIPHER_CTX).
3674 EVP_MD_CTX_init(&md); /* new function call */
3675 EVP_DigestInit(&md, EVP_sha1());
3676 EVP_DigestUpdate(&md, in, len);
3677 EVP_DigestFinal(&md, out, NULL);
3678 EVP_MD_CTX_cleanup(&md); /* new function call */
3682 *) Make DES key schedule conform to the usual scheme, as well as
3683 correcting its structure. This means that calls to DES functions
3684 now have to pass a pointer to a des_key_schedule instead of a
3685 plain des_key_schedule (which was actually always a pointer
3688 des_key_schedule ks;
3690 des_set_key_checked(..., &ks);
3691 des_ncbc_encrypt(..., &ks, ...);
3693 (Note that a later change renames 'des_...' into 'DES_...'.)
3696 *) Initial reduction of linker bloat: the use of some functions, such as
3697 PEM causes large amounts of unused functions to be linked in due to
3698 poor organisation. For example pem_all.c contains every PEM function
3699 which has a knock on effect of linking in large amounts of (unused)
3700 ASN1 code. Grouping together similar functions and splitting unrelated
3701 functions prevents this.
3704 *) Cleanup of EVP macros.
3707 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
3708 correct _ecb suffix.
3711 *) Add initial OCSP responder support to ocsp application. The
3712 revocation information is handled using the text based index
3713 use by the ca application. The responder can either handle
3714 requests generated internally, supplied in files (for example
3715 via a CGI script) or using an internal minimal server.
3718 *) Add configuration choices to get zlib compression for TLS.
3721 *) Changes to Kerberos SSL for RFC 2712 compliance:
3722 1. Implemented real KerberosWrapper, instead of just using
3723 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
3724 2. Implemented optional authenticator field of KerberosWrapper.
3726 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
3727 and authenticator structs; see crypto/krb5/.
3729 Generalized Kerberos calls to support multiple Kerberos libraries.
3730 [Vern Staats <staatsvr@asc.hpc.mil>,
3731 Jeffrey Altman <jaltman@columbia.edu>
3732 via Richard Levitte]
3734 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
3735 already does with RSA. testdsa.h now has 'priv_key/pub_key'
3736 values for each of the key sizes rather than having just
3737 parameters (and 'speed' generating keys each time).
3740 *) Speed up EVP routines.
3743 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
3744 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
3745 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
3746 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
3748 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
3749 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
3750 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
3753 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
3755 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
3758 *) Added the OS2-EMX target.
3759 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
3761 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
3762 to support NCONF routines in extension code. New function CONF_set_nconf()
3763 to allow functions which take an NCONF to also handle the old LHASH
3764 structure: this means that the old CONF compatible routines can be
3765 retained (in particular wrt extensions) without having to duplicate the
3766 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
3769 *) Enhance the general user interface with mechanisms for inner control
3770 and with possibilities to have yes/no kind of prompts.
3773 *) Change all calls to low level digest routines in the library and
3774 applications to use EVP. Add missing calls to HMAC_cleanup() and
3775 don't assume HMAC_CTX can be copied using memcpy().
3776 [Verdon Walker <VWalker@novell.com>, Steve Henson]
3778 *) Add the possibility to control engines through control names but with
3779 arbitrary arguments instead of just a string.
3780 Change the key loaders to take a UI_METHOD instead of a callback
3781 function pointer. NOTE: this breaks binary compatibility with earlier
3782 versions of OpenSSL [engine].
3783 Adapt the nCipher code for these new conditions and add a card insertion
3787 *) Enhance the general user interface with mechanisms to better support
3788 dialog box interfaces, application-defined prompts, the possibility
3789 to use defaults (for example default passwords from somewhere else)
3790 and interrupts/cancellations.
3793 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
3794 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
3797 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
3798 tidy up some unnecessarily weird code in 'sk_new()').
3799 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
3801 *) Change the key loading routines for ENGINEs to use the same kind
3802 callback (pem_password_cb) as all other routines that need this
3806 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
3807 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
3808 than this minimum value is recommended.
3811 *) New random seeder for OpenVMS, using the system process statistics
3812 that are easily reachable.
3815 *) Windows apparently can't transparently handle global
3816 variables defined in DLLs. Initialisations such as:
3818 const ASN1_ITEM *it = &ASN1_INTEGER_it;
3820 wont compile. This is used by the any applications that need to
3821 declare their own ASN1 modules. This was fixed by adding the option
3822 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
3823 needed for static libraries under Win32.
3826 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
3827 setting of purpose and trust fields. New X509_STORE trust and
3828 purpose functions and tidy up setting in other SSL functions.
3831 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
3832 structure. These are inherited by X509_STORE_CTX when it is
3833 initialised. This allows various defaults to be set in the
3834 X509_STORE structure (such as flags for CRL checking and custom
3835 purpose or trust settings) for functions which only use X509_STORE_CTX
3836 internally such as S/MIME.
3838 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
3839 trust settings if they are not set in X509_STORE. This allows X509_STORE
3840 purposes and trust (in S/MIME for example) to override any set by default.
3842 Add command line options for CRL checking to smime, s_client and s_server
3846 *) Initial CRL based revocation checking. If the CRL checking flag(s)
3847 are set then the CRL is looked up in the X509_STORE structure and
3848 its validity and signature checked, then if the certificate is found
3849 in the CRL the verify fails with a revoked error.
3851 Various new CRL related callbacks added to X509_STORE_CTX structure.
3853 Command line options added to 'verify' application to support this.
3855 This needs some additional work, such as being able to handle multiple
3856 CRLs with different times, extension based lookup (rather than just
3857 by subject name) and ultimately more complete V2 CRL extension
3861 *) Add a general user interface API (crypto/ui/). This is designed
3862 to replace things like des_read_password and friends (backward
3863 compatibility functions using this new API are provided).
3864 The purpose is to remove prompting functions from the DES code
3865 section as well as provide for prompting through dialog boxes in
3866 a window system and the like.
3869 *) Add "ex_data" support to ENGINE so implementations can add state at a
3870 per-structure level rather than having to store it globally.
3873 *) Make it possible for ENGINE structures to be copied when retrieved by
3874 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
3875 This causes the "original" ENGINE structure to act like a template,
3876 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
3877 operational state can be localised to each ENGINE structure, despite the
3878 fact they all share the same "methods". New ENGINE structures returned in
3879 this case have no functional references and the return value is the single
3880 structural reference. This matches the single structural reference returned
3881 by ENGINE_by_id() normally, when it is incremented on the pre-existing
3885 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
3886 needs to match any other type at all we need to manually clear the
3890 *) Changes to the "openssl engine" utility to include;
3891 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
3892 about an ENGINE's available control commands.
3893 - executing control commands from command line arguments using the
3894 '-pre' and '-post' switches. '-post' is only used if '-t' is
3895 specified and the ENGINE is successfully initialised. The syntax for
3896 the individual commands are colon-separated, for example;
3897 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
3900 *) New dynamic control command support for ENGINEs. ENGINEs can now
3901 declare their own commands (numbers), names (strings), descriptions,
3902 and input types for run-time discovery by calling applications. A
3903 subset of these commands are implicitly classed as "executable"
3904 depending on their input type, and only these can be invoked through
3905 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
3906 can be based on user input, config files, etc). The distinction is
3907 that "executable" commands cannot return anything other than a boolean
3908 result and can only support numeric or string input, whereas some
3909 discoverable commands may only be for direct use through
3910 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
3911 pointers, or other custom uses. The "executable" commands are to
3912 support parameterisations of ENGINE behaviour that can be
3913 unambiguously defined by ENGINEs and used consistently across any
3914 OpenSSL-based application. Commands have been added to all the
3915 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
3916 control over shared-library paths without source code alterations.
3919 *) Changed all ENGINE implementations to dynamically allocate their
3920 ENGINEs rather than declaring them statically. Apart from this being
3921 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
3922 this also allows the implementations to compile without using the
3923 internal engine_int.h header.
3926 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
3927 'const' value. Any code that should be able to modify a RAND_METHOD
3928 should already have non-const pointers to it (ie. they should only
3929 modify their own ones).
3932 *) Made a variety of little tweaks to the ENGINE code.
3933 - "atalla" and "ubsec" string definitions were moved from header files
3934 to C code. "nuron" string definitions were placed in variables
3935 rather than hard-coded - allowing parameterisation of these values
3936 later on via ctrl() commands.
3937 - Removed unused "#if 0"'d code.
3938 - Fixed engine list iteration code so it uses ENGINE_free() to release
3939 structural references.
3940 - Constified the RAND_METHOD element of ENGINE structures.
3941 - Constified various get/set functions as appropriate and added
3942 missing functions (including a catch-all ENGINE_cpy that duplicates
3943 all ENGINE values onto a new ENGINE except reference counts/state).
3944 - Removed NULL parameter checks in get/set functions. Setting a method
3945 or function to NULL is a way of cancelling out a previously set
3946 value. Passing a NULL ENGINE parameter is just plain stupid anyway
3947 and doesn't justify the extra error symbols and code.
3948 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
3949 flags from engine_int.h to engine.h.
3950 - Changed prototypes for ENGINE handler functions (init(), finish(),
3951 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
3954 *) Implement binary inversion algorithm for BN_mod_inverse in addition
3955 to the algorithm using long division. The binary algorithm can be
3956 used only if the modulus is odd. On 32-bit systems, it is faster
3957 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
3958 roughly 5-15% for 256-bit moduli), so we use it only for moduli
3959 up to 450 bits. In 64-bit environments, the binary algorithm
3960 appears to be advantageous for much longer moduli; here we use it
3961 for moduli up to 2048 bits.
3964 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
3965 could not support the combine flag in choice fields.
3968 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
3969 extensions from a certificate request to the certificate.
3972 *) Allow multiple 'certopt' and 'nameopt' options to be separated
3973 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
3974 file: this allows the display of the certificate about to be
3975 signed to be customised, to allow certain fields to be included
3976 or excluded and extension details. The old system didn't display
3977 multicharacter strings properly, omitted fields not in the policy
3978 and couldn't display additional details such as extensions.
3981 *) Function EC_POINTs_mul for multiple scalar multiplication
3982 of an arbitrary number of elliptic curve points
3983 \sum scalars[i]*points[i],
3984 optionally including the generator defined for the EC_GROUP:
3985 scalar*generator + \sum scalars[i]*points[i].
3987 EC_POINT_mul is a simple wrapper function for the typical case
3988 that the point list has just one item (besides the optional
3992 *) First EC_METHODs for curves over GF(p):
3994 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
3995 operations and provides various method functions that can also
3996 operate with faster implementations of modular arithmetic.
3998 EC_GFp_mont_method() reuses most functions that are part of
3999 EC_GFp_simple_method, but uses Montgomery arithmetic.
4001 [Bodo Moeller; point addition and point doubling
4002 implementation directly derived from source code provided by
4003 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
4005 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
4006 crypto/ec/ec_lib.c):
4008 Curves are EC_GROUP objects (with an optional group generator)
4009 based on EC_METHODs that are built into the library.
4011 Points are EC_POINT objects based on EC_GROUP objects.
4013 Most of the framework would be able to handle curves over arbitrary
4014 finite fields, but as there are no obvious types for fields other
4015 than GF(p), some functions are limited to that for now.
4018 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
4019 that the file contains a complete HTTP response.
4022 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
4023 change the def and num file printf format specifier from "%-40sXXX"
4024 to "%-39s XXX". The latter will always guarantee a space after the
4025 field while the former will cause them to run together if the field
4026 is 40 of more characters long.
4029 *) Constify the cipher and digest 'method' functions and structures
4030 and modify related functions to take constant EVP_MD and EVP_CIPHER
4034 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
4035 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
4038 *) Modify EVP_Digest*() routines so they now return values. Although the
4039 internal software routines can never fail additional hardware versions
4043 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
4045 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
4046 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
4049 ERR_R_NESTED_ASN1_ERROR
4051 ERR_R_MISSING_ASN1_EOS
4052 were 4 .. 9, conflicting with
4053 ERR_LIB_RSA (= ERR_R_RSA_LIB)
4055 ERR_LIB_PEM (= ERR_R_PEM_LIB).
4056 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
4058 Add new error code 'ERR_R_INTERNAL_ERROR'.
4061 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
4065 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
4066 sets the subject name for a new request or supersedes the
4067 subject name in a given request. Formats that can be parsed are
4068 'CN=Some Name, OU=myOU, C=IT'
4070 'CN=Some Name/OU=myOU/C=IT'.
4072 Add options '-batch' and '-verbose' to 'openssl req'.
4073 [Massimiliano Pala <madwolf@hackmasters.net>]
4075 *) Introduce the possibility to access global variables through
4076 functions on platform were that's the best way to handle exporting
4077 global variables in shared libraries. To enable this functionality,
4078 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
4079 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
4080 is normally done by Configure or something similar).
4082 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
4083 in the source file (foo.c) like this:
4085 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
4086 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
4088 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
4089 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
4091 OPENSSL_DECLARE_GLOBAL(int,foo);
4092 #define foo OPENSSL_GLOBAL_REF(foo)
4093 OPENSSL_DECLARE_GLOBAL(double,bar);
4094 #define bar OPENSSL_GLOBAL_REF(bar)
4096 The #defines are very important, and therefore so is including the
4097 header file everywhere where the defined globals are used.
4099 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
4100 of ASN.1 items, but that structure is a bit different.
4102 The largest change is in util/mkdef.pl which has been enhanced with
4103 better and easier to understand logic to choose which symbols should
4104 go into the Windows .def files as well as a number of fixes and code
4105 cleanup (among others, algorithm keywords are now sorted
4106 lexicographically to avoid constant rewrites).
4109 *) In BN_div() keep a copy of the sign of 'num' before writing the
4110 result to 'rm' because if rm==num the value will be overwritten
4111 and produce the wrong result if 'num' is negative: this caused
4112 problems with BN_mod() and BN_nnmod().
4115 *) Function OCSP_request_verify(). This checks the signature on an
4116 OCSP request and verifies the signer certificate. The signer
4117 certificate is just checked for a generic purpose and OCSP request
4121 *) Add OCSP_check_validity() function to check the validity of OCSP
4122 responses. OCSP responses are prepared in real time and may only
4123 be a few seconds old. Simply checking that the current time lies
4124 between thisUpdate and nextUpdate max reject otherwise valid responses
4125 caused by either OCSP responder or client clock inaccuracy. Instead
4126 we allow thisUpdate and nextUpdate to fall within a certain period of
4127 the current time. The age of the response can also optionally be
4128 checked. Two new options -validity_period and -status_age added to
4132 *) If signature or public key algorithm is unrecognized print out its
4133 OID rather that just UNKNOWN.
4136 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
4137 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
4138 ID to be generated from the issuer certificate alone which can then be
4139 passed to OCSP_id_issuer_cmp().
4142 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
4143 ASN1 modules to export functions returning ASN1_ITEM pointers
4144 instead of the ASN1_ITEM structures themselves. This adds several
4145 new macros which allow the underlying ASN1 function/structure to
4146 be accessed transparently. As a result code should not use ASN1_ITEM
4147 references directly (such as &X509_it) but instead use the relevant
4148 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
4149 use of the new ASN1 code on platforms where exporting structures
4150 is problematical (for example in shared libraries) but exporting
4151 functions returning pointers to structures is not.
4154 *) Add support for overriding the generation of SSL/TLS session IDs.
4155 These callbacks can be registered either in an SSL_CTX or per SSL.
4156 The purpose of this is to allow applications to control, if they wish,
4157 the arbitrary values chosen for use as session IDs, particularly as it
4158 can be useful for session caching in multiple-server environments. A
4159 command-line switch for testing this (and any client code that wishes
4160 to use such a feature) has been added to "s_server".
4161 [Geoff Thorpe, Lutz Jaenicke]
4163 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
4164 of the form '#if defined(...) || defined(...) || ...' and
4165 '#if !defined(...) && !defined(...) && ...'. This also avoids
4166 the growing number of special cases it was previously handling.
4169 *) Make all configuration macros available for application by making
4170 sure they are available in opensslconf.h, by giving them names starting
4171 with "OPENSSL_" to avoid conflicts with other packages and by making
4172 sure e_os2.h will cover all platform-specific cases together with
4174 Additionally, it is now possible to define configuration/platform-
4175 specific names (called "system identities"). In the C code, these
4176 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
4177 macro with the name beginning with "OPENSSL_SYS_", which is determined
4178 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
4182 *) New option -set_serial to 'req' and 'x509' this allows the serial
4183 number to use to be specified on the command line. Previously self
4184 signed certificates were hard coded with serial number 0 and the
4185 CA options of 'x509' had to use a serial number in a file which was
4189 *) New options to 'ca' utility to support V2 CRL entry extensions.
4190 Currently CRL reason, invalidity date and hold instruction are
4191 supported. Add new CRL extensions to V3 code and some new objects.
4194 *) New function EVP_CIPHER_CTX_set_padding() this is used to
4195 disable standard block padding (aka PKCS#5 padding) in the EVP
4196 API, which was previously mandatory. This means that the data is
4197 not padded in any way and so the total length much be a multiple
4198 of the block size, otherwise an error occurs.
4201 *) Initial (incomplete) OCSP SSL support.
4204 *) New function OCSP_parse_url(). This splits up a URL into its host,
4205 port and path components: primarily to parse OCSP URLs. New -url
4206 option to ocsp utility.
4209 *) New nonce behavior. The return value of OCSP_check_nonce() now
4210 reflects the various checks performed. Applications can decide
4211 whether to tolerate certain situations such as an absent nonce
4212 in a response when one was present in a request: the ocsp application
4213 just prints out a warning. New function OCSP_add1_basic_nonce()
4214 this is to allow responders to include a nonce in a response even if
4215 the request is nonce-less.
4218 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
4219 skipped when using openssl x509 multiple times on a single input file,
4220 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
4223 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
4224 set string type: to handle setting ASN1_TIME structures. Fix ca
4225 utility to correctly initialize revocation date of CRLs.
4228 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
4229 the clients preferred ciphersuites and rather use its own preferences.
4230 Should help to work around M$ SGC (Server Gated Cryptography) bug in
4231 Internet Explorer by ensuring unchanged hash method during stepup.
4232 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
4235 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
4236 to aes and add a new 'exist' option to print out symbols that don't
4240 *) Additional options to ocsp utility to allow flags to be set and
4241 additional certificates supplied.
4244 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
4245 OCSP client a number of certificate to only verify the response
4249 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
4250 handle the new API. Currently only ECB, CBC modes supported. Add new
4253 Add TLS AES ciphersuites as described in RFC3268, "Advanced
4254 Encryption Standard (AES) Ciphersuites for Transport Layer
4255 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
4256 not enabled by default and were not part of the "ALL" ciphersuite
4257 alias because they were not yet official; they could be
4258 explicitly requested by specifying the "AESdraft" ciphersuite
4259 group alias. In the final release of OpenSSL 0.9.7, the group
4260 alias is called "AES" and is part of "ALL".)
4261 [Ben Laurie, Steve Henson, Bodo Moeller]
4263 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
4264 request to response.
4267 *) Functions for OCSP responders. OCSP_request_onereq_count(),
4268 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
4269 extract information from a certificate request. OCSP_response_create()
4270 creates a response and optionally adds a basic response structure.
4271 OCSP_basic_add1_status() adds a complete single response to a basic
4272 response and returns the OCSP_SINGLERESP structure just added (to allow
4273 extensions to be included for example). OCSP_basic_add1_cert() adds a
4274 certificate to a basic response and OCSP_basic_sign() signs a basic
4275 response with various flags. New helper functions ASN1_TIME_check()
4276 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
4277 (converts ASN1_TIME to GeneralizedTime).
4280 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
4281 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
4282 structure from a certificate. X509_pubkey_digest() digests the public_key
4283 contents: this is used in various key identifiers.
4286 *) Make sk_sort() tolerate a NULL argument.
4287 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
4289 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
4290 passed by the function are trusted implicitly. If any of them signed the
4291 response then it is assumed to be valid and is not verified.
4294 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
4295 to data. This was previously part of the PKCS7 ASN1 code. This
4296 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
4297 [Steve Henson, reported by Kenneth R. Robinette
4298 <support@securenetterm.com>]
4300 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
4301 routines: without these tracing memory leaks is very painful.
4302 Fix leaks in PKCS12 and PKCS7 routines.
4305 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
4306 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
4307 effectively meant GeneralizedTime would never be used. Now it
4308 is initialised to -1 but X509_time_adj() now has to check the value
4309 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
4310 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
4311 [Steve Henson, reported by Kenneth R. Robinette
4312 <support@securenetterm.com>]
4314 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
4315 result in a zero length in the ASN1_INTEGER structure which was
4316 not consistent with the structure when d2i_ASN1_INTEGER() was used
4317 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
4318 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
4319 where it did not print out a minus for negative ASN1_INTEGER.
4322 *) Add summary printout to ocsp utility. The various functions which
4323 convert status values to strings have been renamed to:
4324 OCSP_response_status_str(), OCSP_cert_status_str() and
4325 OCSP_crl_reason_str() and are no longer static. New options
4326 to verify nonce values and to disable verification. OCSP response
4327 printout format cleaned up.
4330 *) Add additional OCSP certificate checks. These are those specified
4331 in RFC2560. This consists of two separate checks: the CA of the
4332 certificate being checked must either be the OCSP signer certificate
4333 or the issuer of the OCSP signer certificate. In the latter case the
4334 OCSP signer certificate must contain the OCSP signing extended key
4335 usage. This check is performed by attempting to match the OCSP
4336 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
4337 in the OCSP_CERTID structures of the response.
4340 *) Initial OCSP certificate verification added to OCSP_basic_verify()
4341 and related routines. This uses the standard OpenSSL certificate
4342 verify routines to perform initial checks (just CA validity) and
4343 to obtain the certificate chain. Then additional checks will be
4344 performed on the chain. Currently the root CA is checked to see
4345 if it is explicitly trusted for OCSP signing. This is used to set
4346 a root CA as a global signing root: that is any certificate that
4347 chains to that CA is an acceptable OCSP signing certificate.
4350 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
4351 extensions from a separate configuration file.
4352 As when reading extensions from the main configuration file,
4353 the '-extensions ...' option may be used for specifying the
4355 [Massimiliano Pala <madwolf@comune.modena.it>]
4357 *) New OCSP utility. Allows OCSP requests to be generated or
4358 read. The request can be sent to a responder and the output
4359 parsed, outputed or printed in text form. Not complete yet:
4360 still needs to check the OCSP response validity.
4363 *) New subcommands for 'openssl ca':
4364 'openssl ca -status <serial>' prints the status of the cert with
4365 the given serial number (according to the index file).
4366 'openssl ca -updatedb' updates the expiry status of certificates
4368 [Massimiliano Pala <madwolf@comune.modena.it>]
4370 *) New '-newreq-nodes' command option to CA.pl. This is like
4371 '-newreq', but calls 'openssl req' with the '-nodes' option
4372 so that the resulting key is not encrypted.
4373 [Damien Miller <djm@mindrot.org>]
4375 *) New configuration for the GNU Hurd.
4376 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
4378 *) Initial code to implement OCSP basic response verify. This
4379 is currently incomplete. Currently just finds the signer's
4380 certificate and verifies the signature on the response.
4383 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
4384 value of OPENSSLDIR. This is available via the new '-d' option
4385 to 'openssl version', and is also included in 'openssl version -a'.
4388 *) Allowing defining memory allocation callbacks that will be given
4389 file name and line number information in additional arguments
4390 (a const char* and an int). The basic functionality remains, as
4391 well as the original possibility to just replace malloc(),
4392 realloc() and free() by functions that do not know about these
4393 additional arguments. To register and find out the current
4394 settings for extended allocation functions, the following
4395 functions are provided:
4397 CRYPTO_set_mem_ex_functions
4398 CRYPTO_set_locked_mem_ex_functions
4399 CRYPTO_get_mem_ex_functions
4400 CRYPTO_get_locked_mem_ex_functions
4402 These work the same way as CRYPTO_set_mem_functions and friends.
4403 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
4404 extended allocation function is enabled.
4405 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
4406 a conventional allocation function is enabled.
4407 [Richard Levitte, Bodo Moeller]
4409 *) Finish off removing the remaining LHASH function pointer casts.
4410 There should no longer be any prototype-casting required when using
4411 the LHASH abstraction, and any casts that remain are "bugs". See
4412 the callback types and macros at the head of lhash.h for details
4413 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
4416 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
4417 If /dev/[u]random devices are not available or do not return enough
4418 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
4420 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
4421 /etc/entropy will be queried once each in this sequence, quering stops
4422 when enough entropy was collected without querying more sockets.
4425 *) Change the Unix RAND_poll() variant to be able to poll several
4426 random devices, as specified by DEVRANDOM, until a sufficient amount
4427 of data has been collected. We spend at most 10 ms on each file
4428 (select timeout) and read in non-blocking mode. DEVRANDOM now
4429 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
4430 (previously it was just the string "/dev/urandom"), so on typical
4431 platforms the 10 ms delay will never occur.
4432 Also separate out the Unix variant to its own file, rand_unix.c.
4433 For VMS, there's a currently-empty rand_vms.c.
4436 *) Move OCSP client related routines to ocsp_cl.c. These
4437 provide utility functions which an application needing
4438 to issue a request to an OCSP responder and analyse the
4439 response will typically need: as opposed to those which an
4440 OCSP responder itself would need which will be added later.
4442 OCSP_request_sign() signs an OCSP request with an API similar
4443 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
4444 response. OCSP_response_get1_basic() extracts basic response
4445 from response. OCSP_resp_find_status(): finds and extracts status
4446 information from an OCSP_CERTID structure (which will be created
4447 when the request structure is built). These are built from lower
4448 level functions which work on OCSP_SINGLERESP structures but
4449 wont normally be used unless the application wishes to examine
4450 extensions in the OCSP response for example.
4452 Replace nonce routines with a pair of functions.
4453 OCSP_request_add1_nonce() adds a nonce value and optionally
4454 generates a random value. OCSP_check_nonce() checks the
4455 validity of the nonce in an OCSP response.
4458 *) Change function OCSP_request_add() to OCSP_request_add0_id().
4459 This doesn't copy the supplied OCSP_CERTID and avoids the
4460 need to free up the newly created id. Change return type
4461 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
4462 This can then be used to add extensions to the request.
4463 Deleted OCSP_request_new(), since most of its functionality
4464 is now in OCSP_REQUEST_new() (and the case insensitive name
4465 clash) apart from the ability to set the request name which
4466 will be added elsewhere.
4469 *) Update OCSP API. Remove obsolete extensions argument from
4470 various functions. Extensions are now handled using the new
4471 OCSP extension code. New simple OCSP HTTP function which
4472 can be used to send requests and parse the response.
4475 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
4476 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
4477 uses the special reorder version of SET OF to sort the attributes
4478 and reorder them to match the encoded order. This resolves a long
4479 standing problem: a verify on a PKCS7 structure just after signing
4480 it used to fail because the attribute order did not match the
4481 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
4482 it uses the received order. This is necessary to tolerate some broken
4483 software that does not order SET OF. This is handled by encoding
4484 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
4485 to produce the required SET OF.
4488 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
4489 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
4490 files to get correct declarations of the ASN.1 item variables.
4493 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
4494 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
4495 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
4496 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
4497 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
4498 ASN1_ITEM and no wrapper functions.
4501 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
4502 replace the old function pointer based I/O routines. Change most of
4503 the *_d2i_bio() and *_d2i_fp() functions to use these.
4506 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
4507 lines, recognice more "algorithms" that can be deselected, and make
4508 it complain about algorithm deselection that isn't recognised.
4511 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
4512 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
4513 to use new functions. Add NO_ASN1_OLD which can be set to remove
4514 some old style ASN1 functions: this can be used to determine if old
4515 code will still work when these eventually go away.
4518 *) New extension functions for OCSP structures, these follow the
4519 same conventions as certificates and CRLs.
4522 *) New function X509V3_add1_i2d(). This automatically encodes and
4523 adds an extension. Its behaviour can be customised with various
4524 flags to append, replace or delete. Various wrappers added for
4525 certifcates and CRLs.
4528 *) Fix to avoid calling the underlying ASN1 print routine when
4529 an extension cannot be parsed. Correct a typo in the
4530 OCSP_SERVICELOC extension. Tidy up print OCSP format.
4533 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
4534 entries for variables.
4537 *) Add functionality to apps/openssl.c for detecting locking
4538 problems: As the program is single-threaded, all we have
4539 to do is register a locking callback using an array for
4540 storing which locks are currently held by the program.
4543 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
4544 SSL_get_ex_data_X509_STORE_idx(), which is used in
4545 ssl_verify_cert_chain() and thus can be called at any time
4546 during TLS/SSL handshakes so that thread-safety is essential.
4547 Unfortunately, the ex_data design is not at all suited
4548 for multi-threaded use, so it probably should be abolished.
4551 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
4552 [Broadcom, tweaked and integrated by Geoff Thorpe]
4554 *) Move common extension printing code to new function
4555 X509V3_print_extensions(). Reorganise OCSP print routines and
4556 implement some needed OCSP ASN1 functions. Add OCSP extensions.
4559 *) New function X509_signature_print() to remove duplication in some
4563 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
4564 set (this was treated exactly the same as SET OF previously). This
4565 is used to reorder the STACK representing the structure to match the
4566 encoding. This will be used to get round a problem where a PKCS7
4567 structure which was signed could not be verified because the STACK
4568 order did not reflect the encoded order.
4571 *) Reimplement the OCSP ASN1 module using the new code.
4574 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
4575 for its ASN1 operations. The old style function pointers still exist
4576 for now but they will eventually go away.
4579 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
4580 completely replaces the old ASN1 functionality with a table driven
4581 encoder and decoder which interprets an ASN1_ITEM structure describing
4582 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
4583 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
4584 has also been converted to the new form.
4587 *) Change BN_mod_exp_recp so that negative moduli are tolerated
4588 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
4589 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
4590 for negative moduli.
4593 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
4594 of not touching the result's sign bit.
4597 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
4601 *) Changed the LHASH code to use prototypes for callbacks, and created
4602 macros to declare and implement thin (optionally static) functions
4603 that provide type-safety and avoid function pointer casting for the
4604 type-specific callbacks.
4607 *) Added Kerberos Cipher Suites to be used with TLS, as written in
4609 [Veers Staats <staatsvr@asc.hpc.mil>,
4610 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
4612 *) Reformat the FAQ so the different questions and answers can be divided
4613 in sections depending on the subject.
4616 *) Have the zlib compression code load ZLIB.DLL dynamically under
4620 *) New function BN_mod_sqrt for computing square roots modulo a prime
4621 (using the probabilistic Tonelli-Shanks algorithm unless
4622 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
4623 be handled deterministically).
4624 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
4626 *) Make BN_mod_inverse faster by explicitly handling small quotients
4627 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
4628 512 bits], about 30% for larger ones [1024 or 2048 bits].)
4631 *) New function BN_kronecker.
4634 *) Fix BN_gcd so that it works on negative inputs; the result is
4635 positive unless both parameters are zero.
4636 Previously something reasonably close to an infinite loop was
4637 possible because numbers could be growing instead of shrinking
4638 in the implementation of Euclid's algorithm.
4641 *) Fix BN_is_word() and BN_is_one() macros to take into account the
4642 sign of the number in question.
4644 Fix BN_is_word(a,w) to work correctly for w == 0.
4646 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
4647 because its test if the absolute value of 'a' equals 'w'.
4648 Note that BN_abs_is_word does *not* handle w == 0 reliably;
4649 it exists mostly for use in the implementations of BN_is_zero(),
4650 BN_is_one(), and BN_is_word().
4653 *) New function BN_swap.
4656 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
4657 the exponentiation functions are more likely to produce reasonable
4658 results on negative inputs.
4661 *) Change BN_mod_mul so that the result is always non-negative.
4662 Previously, it could be negative if one of the factors was negative;
4663 I don't think anyone really wanted that behaviour.
4666 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
4667 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
4668 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
4669 and add new functions:
4678 BN_mod_lshift1_quick
4682 These functions always generate non-negative results.
4684 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
4685 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
4687 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
4688 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
4689 be reduced modulo m.
4690 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
4693 The following entry accidentily appeared in the CHANGES file
4694 distributed with OpenSSL 0.9.7. The modifications described in
4695 it do *not* apply to OpenSSL 0.9.7.
4697 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4698 was actually never needed) and in BN_mul(). The removal in BN_mul()
4699 required a small change in bn_mul_part_recursive() and the addition
4700 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4701 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4702 bn_sub_words() and bn_add_words() except they take arrays with
4707 *) In 'openssl passwd', verify passwords read from the terminal
4708 unless the '-salt' option is used (which usually means that
4709 verification would just waste user's time since the resulting
4710 hash is going to be compared with some given password hash)
4711 or the new '-noverify' option is used.
4713 This is an incompatible change, but it does not affect
4714 non-interactive use of 'openssl passwd' (passwords on the command
4715 line, '-stdin' option, '-in ...' option) and thus should not
4719 *) Remove all references to RSAref, since there's no more need for it.
4722 *) Make DSO load along a path given through an environment variable
4723 (SHLIB_PATH) with shl_load().
4726 *) Constify the ENGINE code as a result of BIGNUM constification.
4727 Also constify the RSA code and most things related to it. In a
4728 few places, most notable in the depth of the ASN.1 code, ugly
4729 casts back to non-const were required (to be solved at a later
4733 *) Make it so the openssl application has all engines loaded by default.
4736 *) Constify the BIGNUM routines a little more.
4739 *) Add the following functions:
4741 ENGINE_load_cswift()
4743 ENGINE_load_atalla()
4745 ENGINE_load_builtin_engines()
4747 That way, an application can itself choose if external engines that
4748 are built-in in OpenSSL shall ever be used or not. The benefit is
4749 that applications won't have to be linked with libdl or other dso
4750 libraries unless it's really needed.
4752 Changed 'openssl engine' to load all engines on demand.
4753 Changed the engine header files to avoid the duplication of some
4754 declarations (they differed!).
4757 *) 'openssl engine' can now list capabilities.
4760 *) Better error reporting in 'openssl engine'.
4763 *) Never call load_dh_param(NULL) in s_server.
4766 *) Add engine application. It can currently list engines by name and
4767 identity, and test if they are actually available.
4770 *) Improve RPM specification file by forcing symbolic linking and making
4771 sure the installed documentation is also owned by root.root.
4772 [Damien Miller <djm@mindrot.org>]
4774 *) Give the OpenSSL applications more possibilities to make use of
4775 keys (public as well as private) handled by engines.
4778 *) Add OCSP code that comes from CertCo.
4781 *) Add VMS support for the Rijndael code.
4784 *) Added untested support for Nuron crypto accelerator.
4787 *) Add support for external cryptographic devices. This code was
4788 previously distributed separately as the "engine" branch.
4789 [Geoff Thorpe, Richard Levitte]
4791 *) Rework the filename-translation in the DSO code. It is now possible to
4792 have far greater control over how a "name" is turned into a filename
4793 depending on the operating environment and any oddities about the
4794 different shared library filenames on each system.
4797 *) Support threads on FreeBSD-elf in Configure.
4800 *) Fix for SHA1 assembly problem with MASM: it produces
4801 warnings about corrupt line number information when assembling
4802 with debugging information. This is caused by the overlapping
4804 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
4807 NCONF_get_number() has no error checking at all. As a replacement,
4808 NCONF_get_number_e() is defined (_e for "error checking") and is
4809 promoted strongly. The old NCONF_get_number is kept around for
4810 binary backward compatibility.
4811 Make it possible for methods to load from something other than a BIO,
4812 by providing a function pointer that is given a name instead of a BIO.
4813 For example, this could be used to load configuration data from an
4817 *) Fix for non blocking accept BIOs. Added new I/O special reason
4818 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
4819 with non blocking I/O was not possible because no retry code was
4820 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
4824 *) Added the beginnings of Rijndael support.
4827 *) Fix for bug in DirectoryString mask setting. Add support for
4828 X509_NAME_print_ex() in 'req' and X509_print_ex() function
4829 to allow certificate printing to more controllable, additional
4830 'certopt' option to 'x509' to allow new printing options to be
4834 *) Clean old EAY MD5 hack from e_os.h.
4837 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
4839 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4840 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4841 [Joe Orton, Steve Henson]
4843 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
4845 *) Fix additional bug revealed by the NISCC test suite:
4847 Stop bug triggering large recursion when presented with
4848 certain ASN.1 tags (CVE-2003-0851)
4851 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
4853 *) Fix various bugs revealed by running the NISCC test suite:
4855 Stop out of bounds reads in the ASN1 code when presented with
4856 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4858 If verify callback ignores invalid public key errors don't try to check
4859 certificate signature with the NULL public key.
4863 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4864 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4868 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4869 extra data after the compression methods not only for TLS 1.0
4870 but also for SSL 3.0 (as required by the specification).
4871 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4873 *) Change X509_certificate_type() to mark the key as exported/exportable
4874 when it's 512 *bits* long, not 512 bytes.
4877 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
4879 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4880 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4881 a protocol version number mismatch like a decryption error
4882 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4885 *) Turn on RSA blinding by default in the default implementation
4886 to avoid a timing attack. Applications that don't want it can call
4887 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4888 They would be ill-advised to do so in most cases.
4889 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4891 *) Change RSA blinding code so that it works when the PRNG is not
4892 seeded (in this case, the secret RSA exponent is abused as
4893 an unpredictable seed -- if it is not unpredictable, there
4894 is no point in blinding anyway). Make RSA blinding thread-safe
4895 by remembering the creator's thread ID in rsa->blinding and
4896 having all other threads use local one-time blinding factors
4897 (this requires more computation than sharing rsa->blinding, but
4898 avoids excessive locking; and if an RSA object is not shared
4899 between threads, blinding will still be very fast).
4902 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
4904 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4905 via timing by performing a MAC computation even if incorrrect
4906 block cipher padding has been found. This is a countermeasure
4907 against active attacks where the attacker has to distinguish
4908 between bad padding and a MAC verification error. (CVE-2003-0078)
4910 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4911 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4912 Martin Vuagnoux (EPFL, Ilion)]
4914 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
4916 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
4917 memory from it's contents. This is done with a counter that will
4918 place alternating values in each byte. This can be used to solve
4919 two issues: 1) the removal of calls to memset() by highly optimizing
4920 compilers, and 2) cleansing with other values than 0, since those can
4921 be read through on certain media, for example a swap space on disk.
4924 *) Bugfix: client side session caching did not work with external caching,
4925 because the session->cipher setting was not restored when reloading
4926 from the external cache. This problem was masked, when
4927 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
4928 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
4931 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
4932 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
4933 [Zeev Lieber <zeev-l@yahoo.com>]
4935 *) Undo an undocumented change introduced in 0.9.6e which caused
4936 repeated calls to OpenSSL_add_all_ciphers() and
4937 OpenSSL_add_all_digests() to be ignored, even after calling
4941 *) Change the default configuration reader to deal with last line not
4942 being properly terminated.
4945 *) Change X509_NAME_cmp() so it applies the special rules on handling
4946 DN values that are of type PrintableString, as well as RDNs of type
4947 emailAddress where the value has the type ia5String.
4948 [stefank@valicert.com via Richard Levitte]
4950 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
4951 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
4952 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
4953 the bitwise-OR of the two for use by the majority of applications
4954 wanting this behaviour, and update the docs. The documented
4955 behaviour and actual behaviour were inconsistent and had been
4956 changing anyway, so this is more a bug-fix than a behavioural
4958 [Geoff Thorpe, diagnosed by Nadav Har'El]
4960 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
4961 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
4964 *) Fix initialization code race conditions in
4965 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
4966 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
4967 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
4968 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
4969 ssl2_get_cipher_by_char(),
4970 ssl3_get_cipher_by_char().
4971 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
4973 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
4974 the cached sessions are flushed, as the remove_cb() might use ex_data
4975 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
4976 (see [openssl.org #212]).
4977 [Geoff Thorpe, Lutz Jaenicke]
4979 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
4980 length, instead of the encoding length to d2i_ASN1_OBJECT.
4983 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
4985 *) [In 0.9.6g-engine release:]
4986 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
4987 [Lynn Gazis <lgazis@rainbow.com>]
4989 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
4991 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
4992 and get fix the header length calculation.
4993 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
4994 Alon Kantor <alonk@checkpoint.com> (and others),
4997 *) Use proper error handling instead of 'assertions' in buffer
4998 overflow checks added in 0.9.6e. This prevents DoS (the
4999 assertions could call abort()).
5000 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
5002 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
5004 *) Add various sanity checks to asn1_get_length() to reject
5005 the ASN1 length bytes if they exceed sizeof(long), will appear
5006 negative or the content length exceeds the length of the
5008 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5010 *) Fix cipher selection routines: ciphers without encryption had no flags
5011 for the cipher strength set and where therefore not handled correctly
5012 by the selection routines (PR #130).
5015 *) Fix EVP_dsa_sha macro.
5019 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
5020 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
5021 that was added in OpenSSL 0.9.6d.
5023 As the countermeasure turned out to be incompatible with some
5024 broken SSL implementations, the new option is part of SSL_OP_ALL.
5025 SSL_OP_ALL is usually employed when compatibility with weird SSL
5026 implementations is desired (e.g. '-bugs' option to 's_client' and
5027 's_server'), so the new option is automatically set in many
5031 *) Changes in security patch:
5033 Changes marked "(CHATS)" were sponsored by the Defense Advanced
5034 Research Projects Agency (DARPA) and Air Force Research Laboratory,
5035 Air Force Materiel Command, USAF, under agreement number
5038 *) Add various sanity checks to asn1_get_length() to reject
5039 the ASN1 length bytes if they exceed sizeof(long), will appear
5040 negative or the content length exceeds the length of the
5041 supplied buffer. (CVE-2002-0659)
5042 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5044 *) Assertions for various potential buffer overflows, not known to
5046 [Ben Laurie (CHATS)]
5048 *) Various temporary buffers to hold ASCII versions of integers were
5049 too small for 64 bit platforms. (CVE-2002-0655)
5050 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
5052 *) Remote buffer overflow in SSL3 protocol - an attacker could
5053 supply an oversized session ID to a client. (CVE-2002-0656)
5054 [Ben Laurie (CHATS)]
5056 *) Remote buffer overflow in SSL2 protocol - an attacker could
5057 supply an oversized client master key. (CVE-2002-0656)
5058 [Ben Laurie (CHATS)]
5060 Changes between 0.9.6c and 0.9.6d [9 May 2002]
5062 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
5063 encoded as NULL) with id-dsa-with-sha1.
5064 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
5066 *) Check various X509_...() return values in apps/req.c.
5067 [Nils Larsch <nla@trustcenter.de>]
5069 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
5070 an end-of-file condition would erronously be flagged, when the CRLF
5071 was just at the end of a processed block. The bug was discovered when
5072 processing data through a buffering memory BIO handing the data to a
5073 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
5074 <ptsekov@syntrex.com> and Nedelcho Stanev.
5077 *) Implement a countermeasure against a vulnerability recently found
5078 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
5079 before application data chunks to avoid the use of known IVs
5080 with data potentially chosen by the attacker.
5083 *) Fix length checks in ssl3_get_client_hello().
5086 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
5087 to prevent ssl3_read_internal() from incorrectly assuming that
5088 ssl3_read_bytes() found application data while handshake
5089 processing was enabled when in fact s->s3->in_read_app_data was
5090 merely automatically cleared during the initial handshake.
5091 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
5093 *) Fix object definitions for Private and Enterprise: they were not
5094 recognized in their shortname (=lowercase) representation. Extend
5095 obj_dat.pl to issue an error when using undefined keywords instead
5096 of silently ignoring the problem (Svenning Sorensen
5097 <sss@sss.dnsalias.net>).
5100 *) Fix DH_generate_parameters() so that it works for 'non-standard'
5101 generators, i.e. generators other than 2 and 5. (Previously, the
5102 code did not properly initialise the 'add' and 'rem' values to
5103 BN_generate_prime().)
5105 In the new general case, we do not insist that 'generator' is
5106 actually a primitive root: This requirement is rather pointless;
5107 a generator of the order-q subgroup is just as good, if not
5111 *) Map new X509 verification errors to alerts. Discovered and submitted by
5112 Tom Wu <tom@arcot.com>.
5115 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
5116 returning non-zero before the data has been completely received
5117 when using non-blocking I/O.
5118 [Bodo Moeller; problem pointed out by John Hughes]
5120 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
5121 [Ben Laurie, Lutz Jaenicke]
5123 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
5124 Yoram Zahavi <YoramZ@gilian.com>).
5127 *) Add information about CygWin 1.3 and on, and preserve proper
5128 configuration for the versions before that.
5129 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5131 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
5132 check whether we deal with a copy of a session and do not delete from
5133 the cache in this case. Problem reported by "Izhar Shoshani Levi"
5134 <izhar@checkpoint.com>.
5137 *) Do not store session data into the internal session cache, if it
5138 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
5139 flag is set). Proposed by Aslam <aslam@funk.com>.
5142 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
5146 *) [In 0.9.6d-engine release:]
5147 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5148 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5150 *) Add the configuration target linux-s390x.
5151 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
5153 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
5154 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
5155 variable as an indication that a ClientHello message has been
5156 received. As the flag value will be lost between multiple
5157 invocations of ssl3_accept when using non-blocking I/O, the
5158 function may not be aware that a handshake has actually taken
5159 place, thus preventing a new session from being added to the
5162 To avoid this problem, we now set s->new_session to 2 instead of
5163 using a local variable.
5164 [Lutz Jaenicke, Bodo Moeller]
5166 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
5167 if the SSL_R_LENGTH_MISMATCH error is detected.
5168 [Geoff Thorpe, Bodo Moeller]
5170 *) New 'shared_ldflag' column in Configure platform table.
5173 *) Fix EVP_CIPHER_mode macro.
5174 ["Dan S. Camper" <dan@bti.net>]
5176 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
5177 type, we must throw them away by setting rr->length to 0.
5178 [D P Chang <dpc@qualys.com>]
5180 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
5182 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
5183 <Dominikus.Scherkl@biodata.com>. (The previous implementation
5184 worked incorrectly for those cases where range = 10..._2 and
5185 3*range is two bits longer than range.)
5188 *) Only add signing time to PKCS7 structures if it is not already
5192 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
5193 OBJ_ld_ce should be OBJ_id_ce.
5194 Also some ip-pda OIDs in crypto/objects/objects.txt were
5195 incorrect (cf. RFC 3039).
5196 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
5198 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
5199 returns early because it has nothing to do.
5200 [Andy Schneider <andy.schneider@bjss.co.uk>]
5202 *) [In 0.9.6c-engine release:]
5203 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
5204 [Andy Schneider <andy.schneider@bjss.co.uk>]
5206 *) [In 0.9.6c-engine release:]
5207 Add support for Cryptographic Appliance's keyserver technology.
5208 (Use engine 'keyclient')
5209 [Cryptographic Appliances and Geoff Thorpe]
5211 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
5212 is called via tools/c89.sh because arguments have to be
5213 rearranged (all '-L' options must appear before the first object
5215 [Richard Shapiro <rshapiro@abinitio.com>]
5217 *) [In 0.9.6c-engine release:]
5218 Add support for Broadcom crypto accelerator cards, backported
5220 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
5222 *) [In 0.9.6c-engine release:]
5223 Add support for SureWare crypto accelerator cards from
5224 Baltimore Technologies. (Use engine 'sureware')
5225 [Baltimore Technologies and Mark Cox]
5227 *) [In 0.9.6c-engine release:]
5228 Add support for crypto accelerator cards from Accelerated
5229 Encryption Processing, www.aep.ie. (Use engine 'aep')
5230 [AEP Inc. and Mark Cox]
5232 *) Add a configuration entry for gcc on UnixWare.
5233 [Gary Benson <gbenson@redhat.com>]
5235 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
5236 messages are stored in a single piece (fixed-length part and
5237 variable-length part combined) and fix various bugs found on the way.
5240 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
5241 instead. BIO_gethostbyname() does not know what timeouts are
5242 appropriate, so entries would stay in cache even when they have
5244 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
5246 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
5247 faced with a pathologically small ClientHello fragment that does
5248 not contain client_version: Instead of aborting with an error,
5249 simply choose the highest available protocol version (i.e.,
5250 TLS 1.0 unless it is disabled). In practice, ClientHello
5251 messages are never sent like this, but this change gives us
5252 strictly correct behaviour at least for TLS.
5255 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
5256 never resets s->method to s->ctx->method when called from within
5257 one of the SSL handshake functions.
5258 [Bodo Moeller; problem pointed out by Niko Baric]
5260 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
5261 (sent using the client's version number) if client_version is
5262 smaller than the protocol version in use. Also change
5263 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
5264 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
5265 the client will at least see that alert.
5268 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
5272 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
5273 client receives HelloRequest while in a handshake.
5274 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
5276 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
5277 should end in 'break', not 'goto end' which circuments various
5278 cleanups done in state SSL_ST_OK. But session related stuff
5279 must be disabled for SSL_ST_OK in the case that we just sent a
5282 Also avoid some overhead by not calling ssl_init_wbio_buffer()
5283 before just sending a HelloRequest.
5284 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
5286 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
5287 reveal whether illegal block cipher padding was found or a MAC
5288 verification error occured. (Neither SSLerr() codes nor alerts
5289 are directly visible to potential attackers, but the information
5290 may leak via logfiles.)
5292 Similar changes are not required for the SSL 2.0 implementation
5293 because the number of padding bytes is sent in clear for SSL 2.0,
5294 and the extra bytes are just ignored. However ssl/s2_pkt.c
5295 failed to verify that the purported number of padding bytes is in
5299 *) Add OpenUNIX-8 support including shared libraries
5300 (Boyd Lynn Gerber <gerberb@zenez.com>).
5303 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
5304 'wristwatch attack' using huge encoding parameters (cf.
5305 James H. Manger's CRYPTO 2001 paper). Note that the
5306 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
5307 encoding parameters and hence was not vulnerable.
5310 *) BN_sqr() bug fix.
5311 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
5313 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
5314 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
5315 followed by modular reduction.
5316 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
5318 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
5319 equivalent based on BN_pseudo_rand() instead of BN_rand().
5322 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
5323 This function was broken, as the check for a new client hello message
5324 to handle SGC did not allow these large messages.
5325 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
5328 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
5331 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
5332 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
5335 *) Rework the configuration and shared library support for Tru64 Unix.
5336 The configuration part makes use of modern compiler features and
5337 still retains old compiler behavior for those that run older versions
5338 of the OS. The shared library support part includes a variant that
5339 uses the RPATH feature, and is available through the special
5340 configuration target "alpha-cc-rpath", which will never be selected
5342 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
5344 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
5345 with the same message size as in ssl3_get_certificate_request().
5346 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
5347 messages might inadvertently be reject as too long.
5348 [Petr Lampa <lampa@fee.vutbr.cz>]
5350 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
5353 *) Modified SSL library such that the verify_callback that has been set
5354 specificly for an SSL object with SSL_set_verify() is actually being
5355 used. Before the change, a verify_callback set with this function was
5356 ignored and the verify_callback() set in the SSL_CTX at the time of
5357 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
5358 to allow the necessary settings.
5361 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
5362 explicitly to NULL, as at least on Solaris 8 this seems not always to be
5363 done automatically (in contradiction to the requirements of the C
5364 standard). This made problems when used from OpenSSH.
5367 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
5368 dh->length and always used
5370 BN_rand_range(priv_key, dh->p).
5372 BN_rand_range() is not necessary for Diffie-Hellman, and this
5373 specific range makes Diffie-Hellman unnecessarily inefficient if
5374 dh->length (recommended exponent length) is much smaller than the
5375 length of dh->p. We could use BN_rand_range() if the order of
5376 the subgroup was stored in the DH structure, but we only have
5381 BN_rand(priv_key, l, ...)
5383 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
5389 RSA_eay_public_encrypt
5390 RSA_eay_private_decrypt
5391 RSA_eay_private_encrypt (signing)
5392 RSA_eay_public_decrypt (signature verification)
5394 (default implementations for RSA_public_encrypt,
5395 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
5396 always reject numbers >= n.
5399 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
5400 to synchronize access to 'locking_thread'. This is necessary on
5401 systems where access to 'locking_thread' (an 'unsigned long'
5402 variable) is not atomic.
5405 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
5406 *before* setting the 'crypto_lock_rand' flag. The previous code had
5407 a race condition if 0 is a valid thread ID.
5408 [Travis Vitek <vitek@roguewave.com>]
5410 *) Add support for shared libraries under Irix.
5411 [Albert Chin-A-Young <china@thewrittenword.com>]
5413 *) Add configuration option to build on Linux on both big-endian and
5415 [Ralf Baechle <ralf@uni-koblenz.de>]
5417 *) Add the possibility to create shared libraries on HP-UX.
5420 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
5422 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
5423 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
5424 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
5425 PRNG state recovery was possible based on the output of
5426 one PRNG request appropriately sized to gain knowledge on
5427 'md' followed by enough consecutive 1-byte PRNG requests
5428 to traverse all of 'state'.
5430 1. When updating 'md_local' (the current thread's copy of 'md')
5431 during PRNG output generation, hash all of the previous
5432 'md_local' value, not just the half used for PRNG output.
5434 2. Make the number of bytes from 'state' included into the hash
5435 independent from the number of PRNG bytes requested.
5437 The first measure alone would be sufficient to avoid
5438 Markku-Juhani's attack. (Actually it had never occurred
5439 to me that the half of 'md_local' used for chaining was the
5440 half from which PRNG output bytes were taken -- I had always
5441 assumed that the secret half would be used.) The second
5442 measure makes sure that additional data from 'state' is never
5443 mixed into 'md_local' in small portions; this heuristically
5444 further strengthens the PRNG.
5447 *) Fix crypto/bn/asm/mips3.s.
5450 *) When only the key is given to "enc", the IV is undefined. Print out
5451 an error message in this case.
5454 *) Handle special case when X509_NAME is empty in X509 printing routines.
5457 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
5458 positive and less than q.
5461 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
5462 used: it isn't thread safe and the add_lock_callback should handle
5464 [Paul Rose <Paul.Rose@bridge.com>]
5466 *) Verify that incoming data obeys the block size in
5467 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
5471 [Ulf Möller, Bodo Möller]
5473 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
5474 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
5475 when fixing the server behaviour for backwards-compatible 'client
5476 hello' messages. (Note that the attack is impractical against
5477 SSL 3.0 and TLS 1.0 anyway because length and version checking
5478 means that the probability of guessing a valid ciphertext is
5479 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
5482 Before 0.9.5, the countermeasure (hide the error by generating a
5483 random 'decryption result') did not work properly because
5484 ERR_clear_error() was missing, meaning that SSL_get_error() would
5485 detect the supposedly ignored error.
5487 Both problems are now fixed.
5490 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
5491 (previously it was 1024).
5494 *) Fix for compatibility mode trust settings: ignore trust settings
5495 unless some valid trust or reject settings are present.
5498 *) Fix for blowfish EVP: its a variable length cipher.
5501 *) Fix various bugs related to DSA S/MIME verification. Handle missing
5502 parameters in DSA public key structures and return an error in the
5503 DSA routines if parameters are absent.
5506 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
5507 in the current directory if neither $RANDFILE nor $HOME was set.
5508 RAND_file_name() in 0.9.6a returned NULL in this case. This has
5509 caused some confusion to Windows users who haven't defined $HOME.
5510 Thus RAND_file_name() is changed again: e_os.h can define a
5511 DEFAULT_HOME, which will be used if $HOME is not set.
5512 For Windows, we use "C:"; on other platforms, we still require
5513 environment variables.
5515 *) Move 'if (!initialized) RAND_poll()' into regions protected by
5516 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
5517 having multiple threads call RAND_poll() concurrently.
5520 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
5521 combination of a flag and a thread ID variable.
5522 Otherwise while one thread is in ssleay_rand_bytes (which sets the
5523 flag), *other* threads can enter ssleay_add_bytes without obeying
5524 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
5525 that they do not hold after the first thread unsets add_do_not_lock).
5528 *) Change bctest again: '-x' expressions are not available in all
5532 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
5534 *) Fix a couple of memory leaks in PKCS7_dataDecode()
5535 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
5537 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
5538 the default extension for executables, if any. Also, make the perl
5539 scripts that use symlink() to test if it really exists and use "cp"
5540 if it doesn't. All this made OpenSSL compilable and installable in
5544 *) Fix for asn1_GetSequence() for indefinite length constructed data.
5545 If SEQUENCE is length is indefinite just set c->slen to the total
5546 amount of data available.
5547 [Steve Henson, reported by shige@FreeBSD.org]
5548 [This change does not apply to 0.9.7.]
5550 *) Change bctest to avoid here-documents inside command substitution
5551 (workaround for FreeBSD /bin/sh bug).
5552 For compatibility with Ultrix, avoid shell functions (introduced
5553 in the bctest version that searches along $PATH).
5556 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
5557 with des_encrypt() defined on some operating systems, like Solaris
5561 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
5562 On the Importance of Eliminating Errors in Cryptographic
5563 Computations, J. Cryptology 14 (2001) 2, 101-119,
5564 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
5567 *) MIPS assembler BIGNUM division bug fix.
5570 *) Disabled incorrect Alpha assembler code.
5573 *) Fix PKCS#7 decode routines so they correctly update the length
5574 after reading an EOC for the EXPLICIT tag.
5576 [This change does not apply to 0.9.7.]
5578 *) Fix bug in PKCS#12 key generation routines. This was triggered
5579 if a 3DES key was generated with a 0 initial byte. Include
5580 PKCS12_BROKEN_KEYGEN compilation option to retain the old
5581 (but broken) behaviour.
5584 *) Enhance bctest to search for a working bc along $PATH and print
5586 [Tim Rice <tim@multitalents.net> via Richard Levitte]
5588 *) Fix memory leaks in err.c: free err_data string if necessary;
5589 don't write to the wrong index in ERR_set_error_data.
5592 *) Implement ssl23_peek (analogous to ssl23_read), which previously
5596 *) Replace rdtsc with _emit statements for VC++ version 5.
5597 [Jeremy Cooper <jeremy@baymoo.org>]
5599 *) Make it possible to reuse SSLv2 sessions.
5602 *) In copy_email() check for >= 0 as a return value for
5603 X509_NAME_get_index_by_NID() since 0 is a valid index.
5604 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
5606 *) Avoid coredump with unsupported or invalid public keys by checking if
5607 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
5608 PKCS7_verify() fails with non detached data.
5611 *) Don't use getenv in library functions when run as setuid/setgid.
5612 New function OPENSSL_issetugid().
5615 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
5616 due to incorrect handling of multi-threading:
5618 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
5620 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
5622 3. Count how many times MemCheck_off() has been called so that
5623 nested use can be treated correctly. This also avoids
5624 inband-signalling in the previous code (which relied on the
5625 assumption that thread ID 0 is impossible).
5628 *) Add "-rand" option also to s_client and s_server.
5631 *) Fix CPU detection on Irix 6.x.
5632 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
5633 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
5635 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
5638 [This change does not apply to 0.9.7.]
5640 *) Use the cached encoding of an X509_NAME structure rather than
5641 copying it. This is apparently the reason for the libsafe "errors"
5642 but the code is actually correct.
5645 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
5646 Bleichenbacher's DSA attack.
5647 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
5648 to be set and top=0 forces the highest bit to be set; top=-1 is new
5649 and leaves the highest bit random.
5650 [Ulf Moeller, Bodo Moeller]
5652 *) In the NCONF_...-based implementations for CONF_... queries
5653 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
5654 a temporary CONF structure with the data component set to NULL
5655 (which gives segmentation faults in lh_retrieve).
5656 Instead, use NULL for the CONF pointer in CONF_get_string and
5657 CONF_get_number (which may use environment variables) and directly
5658 return NULL from CONF_get_section.
5661 *) Fix potential buffer overrun for EBCDIC.
5664 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
5665 keyUsage if basicConstraints absent for a CA.
5668 *) Make SMIME_write_PKCS7() write mail header values with a format that
5669 is more generally accepted (no spaces before the semicolon), since
5670 some programs can't parse those values properly otherwise. Also make
5671 sure BIO's that break lines after each write do not create invalid
5675 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
5676 macros previously used would not encode an empty SEQUENCE OF
5677 and break the signature.
5679 [This change does not apply to 0.9.7.]
5681 *) Zero the premaster secret after deriving the master secret in
5685 *) Add some EVP_add_digest_alias registrations (as found in
5686 OpenSSL_add_all_digests()) to SSL_library_init()
5687 aka OpenSSL_add_ssl_algorithms(). This provides improved
5688 compatibility with peers using X.509 certificates
5689 with unconventional AlgorithmIdentifier OIDs.
5692 *) Fix for Irix with NO_ASM.
5693 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
5695 *) ./config script fixes.
5696 [Ulf Moeller, Richard Levitte]
5698 *) Fix 'openssl passwd -1'.
5701 *) Change PKCS12_key_gen_asc() so it can cope with non null
5702 terminated strings whose length is passed in the passlen
5703 parameter, for example from PEM callbacks. This was done
5704 by adding an extra length parameter to asc2uni().
5705 [Steve Henson, reported by <oddissey@samsung.co.kr>]
5707 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
5708 call failed, free the DSA structure.
5711 *) Fix to uni2asc() to cope with zero length Unicode strings.
5712 These are present in some PKCS#12 files.
5715 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
5716 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
5717 when writing a 32767 byte record.
5718 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
5720 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
5721 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
5723 (RSA objects have a reference count access to which is protected
5724 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
5725 so they are meant to be shared between threads.)
5726 [Bodo Moeller, Geoff Thorpe; original patch submitted by
5727 "Reddie, Steven" <Steven.Reddie@ca.com>]
5729 *) Fix a deadlock in CRYPTO_mem_leaks().
5732 *) Use better test patterns in bntest.
5735 *) rand_win.c fix for Borland C.
5738 *) BN_rshift bugfix for n == 0.
5741 *) Add a 'bctest' script that checks for some known 'bc' bugs
5742 so that 'make test' does not abort just because 'bc' is broken.
5745 *) Store verify_result within SSL_SESSION also for client side to
5746 avoid potential security hole. (Re-used sessions on the client side
5747 always resulted in verify_result==X509_V_OK, not using the original
5748 result of the server certificate verification.)
5751 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
5752 SSL3_RT_APPLICATION_DATA, return 0.
5753 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
5757 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
5758 releases, have been re-implemented by renaming the previous
5759 implementations of ssl2_read and ssl3_read to ssl2_read_internal
5760 and ssl3_read_internal, respectively, and adding 'peek' parameters
5761 to them. The new ssl[23]_{read,peek} functions are calls to
5762 ssl[23]_read_internal with the 'peek' flag set appropriately.
5763 A 'peek' parameter has also been added to ssl3_read_bytes, which
5764 does the actual work for ssl3_read_internal.
5767 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
5768 the method-specific "init()" handler. Also clean up ex_data after
5769 calling the method-specific "finish()" handler. Previously, this was
5770 happening the other way round.
5773 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
5774 The previous value, 12, was not always sufficient for BN_mod_exp().
5777 *) Make sure that shared libraries get the internal name engine with
5778 the full version number and not just 0. This should mark the
5779 shared libraries as not backward compatible. Of course, this should
5780 be changed again when we can guarantee backward binary compatibility.
5783 *) Fix typo in get_cert_by_subject() in by_dir.c
5784 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
5786 *) Rework the system to generate shared libraries:
5788 - Make note of the expected extension for the shared libraries and
5789 if there is a need for symbolic links from for example libcrypto.so.0
5790 to libcrypto.so.0.9.7. There is extended info in Configure for
5793 - Make as few rebuilds of the shared libraries as possible.
5795 - Still avoid linking the OpenSSL programs with the shared libraries.
5797 - When installing, install the shared libraries separately from the
5801 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
5803 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
5804 and not in SSL_clear because the latter is also used by the
5805 accept/connect functions; previously, the settings made by
5806 SSL_set_read_ahead would be lost during the handshake.
5807 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
5809 *) Correct util/mkdef.pl to be selective about disabled algorithms.
5810 Previously, it would create entries for disableed algorithms no
5814 *) Added several new manual pages for SSL_* function.
5817 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
5819 *) In ssl23_get_client_hello, generate an error message when faced
5820 with an initial SSL 3.0/TLS record that is too small to contain the
5821 first two bytes of the ClientHello message, i.e. client_version.
5822 (Note that this is a pathologic case that probably has never happened
5823 in real life.) The previous approach was to use the version number
5824 from the record header as a substitute; but our protocol choice
5825 should not depend on that one because it is not authenticated
5826 by the Finished messages.
5829 *) More robust randomness gathering functions for Windows.
5830 [Jeffrey Altman <jaltman@columbia.edu>]
5832 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
5833 not set then we don't setup the error code for issuer check errors
5834 to avoid possibly overwriting other errors which the callback does
5835 handle. If an application does set the flag then we assume it knows
5836 what it is doing and can handle the new informational codes
5840 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
5841 a general "ANY" type, as such it should be able to decode anything
5842 including tagged types. However it didn't check the class so it would
5843 wrongly interpret tagged types in the same way as their universal
5844 counterpart and unknown types were just rejected. Changed so that the
5845 tagged and unknown types are handled in the same way as a SEQUENCE:
5846 that is the encoding is stored intact. There is also a new type
5847 "V_ASN1_OTHER" which is used when the class is not universal, in this
5848 case we have no idea what the actual type is so we just lump them all
5852 *) On VMS, stdout may very well lead to a file that is written to
5853 in a record-oriented fashion. That means that every write() will
5854 write a separate record, which will be read separately by the
5855 programs trying to read from it. This can be very confusing.
5857 The solution is to put a BIO filter in the way that will buffer
5858 text until a linefeed is reached, and then write everything a
5859 line at a time, so every record written will be an actual line,
5860 not chunks of lines and not (usually doesn't happen, but I've
5861 seen it once) several lines in one record. BIO_f_linebuffer() is
5864 Currently, it's a VMS-only method, because that's where it has
5865 been tested well enough.
5868 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
5869 it can return incorrect results.
5870 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
5871 but it was in 0.9.6-beta[12].)
5874 *) Disable the check for content being present when verifying detached
5875 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
5876 include zero length content when signing messages.
5879 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
5880 BIO_ctrl (for BIO pairs).
5883 *) Add DSO method for VMS.
5886 *) Bug fix: Montgomery multiplication could produce results with the
5890 *) Add RPM specification openssl.spec and modify it to build three
5891 packages. The default package contains applications, application
5892 documentation and run-time libraries. The devel package contains
5893 include files, static libraries and function documentation. The
5894 doc package contains the contents of the doc directory. The original
5895 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
5898 *) Add a large number of documentation files for many SSL routines.
5899 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
5901 *) Add a configuration entry for Sony News 4.
5902 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
5904 *) Don't set the two most significant bits to one when generating a
5905 random number < q in the DSA library.
5908 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
5909 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
5910 the underlying transport is blocking) if a handshake took place.
5911 (The default behaviour is needed by applications such as s_client
5912 and s_server that use select() to determine when to use SSL_read;
5913 but for applications that know in advance when to expect data, it
5914 just makes things more complicated.)
5917 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
5921 *) Add a few more EBCDIC conditionals that make `req' and `x509'
5922 work better on such systems.
5923 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5925 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
5926 Update PKCS12_parse() so it copies the friendlyName and the
5927 keyid to the certificates aux info.
5930 *) Fix bug in PKCS7_verify() which caused an infinite loop
5931 if there was more than one signature.
5932 [Sven Uszpelkat <su@celocom.de>]
5934 *) Major change in util/mkdef.pl to include extra information
5935 about each symbol, as well as presentig variables as well
5936 as functions. This change means that there's n more need
5937 to rebuild the .num files when some algorithms are excluded.
5940 *) Allow the verify time to be set by an application,
5941 rather than always using the current time.
5944 *) Phase 2 verify code reorganisation. The certificate
5945 verify code now looks up an issuer certificate by a
5946 number of criteria: subject name, authority key id
5947 and key usage. It also verifies self signed certificates
5948 by the same criteria. The main comparison function is
5949 X509_check_issued() which performs these checks.
5951 Lot of changes were necessary in order to support this
5952 without completely rewriting the lookup code.
5954 Authority and subject key identifier are now cached.
5956 The LHASH 'certs' is X509_STORE has now been replaced
5957 by a STACK_OF(X509_OBJECT). This is mainly because an
5958 LHASH can't store or retrieve multiple objects with
5959 the same hash value.
5961 As a result various functions (which were all internal
5962 use only) have changed to handle the new X509_STORE
5963 structure. This will break anything that messed round
5964 with X509_STORE internally.
5966 The functions X509_STORE_add_cert() now checks for an
5967 exact match, rather than just subject name.
5969 The X509_STORE API doesn't directly support the retrieval
5970 of multiple certificates matching a given criteria, however
5971 this can be worked round by performing a lookup first
5972 (which will fill the cache with candidate certificates)
5973 and then examining the cache for matches. This is probably
5974 the best we can do without throwing out X509_LOOKUP
5975 entirely (maybe later...).
5977 The X509_VERIFY_CTX structure has been enhanced considerably.
5979 All certificate lookup operations now go via a get_issuer()
5980 callback. Although this currently uses an X509_STORE it
5981 can be replaced by custom lookups. This is a simple way
5982 to bypass the X509_STORE hackery necessary to make this
5983 work and makes it possible to use more efficient techniques
5984 in future. A very simple version which uses a simple
5985 STACK for its trusted certificate store is also provided
5986 using X509_STORE_CTX_trusted_stack().
5988 The verify_cb() and verify() callbacks now have equivalents
5989 in the X509_STORE_CTX structure.
5991 X509_STORE_CTX also has a 'flags' field which can be used
5992 to customise the verify behaviour.
5995 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
5996 excludes S/MIME capabilities.
5999 *) When a certificate request is read in keep a copy of the
6000 original encoding of the signed data and use it when outputing
6001 again. Signatures then use the original encoding rather than
6002 a decoded, encoded version which may cause problems if the
6003 request is improperly encoded.
6006 *) For consistency with other BIO_puts implementations, call
6007 buffer_write(b, ...) directly in buffer_puts instead of calling
6010 In BIO_puts, increment b->num_write as in BIO_write.
6011 [Peter.Sylvester@EdelWeb.fr]
6013 *) Fix BN_mul_word for the case where the word is 0. (We have to use
6014 BN_zero, we may not return a BIGNUM with an array consisting of
6018 *) Avoid calling abort() from within the library when problems are
6019 detected, except if preprocessor symbols have been defined
6020 (such as REF_CHECK, BN_DEBUG etc.).
6023 *) New openssl application 'rsautl'. This utility can be
6024 used for low level RSA operations. DER public key
6025 BIO/fp routines also added.
6028 *) New Configure entry and patches for compiling on QNX 4.
6029 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
6031 *) A demo state-machine implementation was sponsored by
6032 Nuron (http://www.nuron.com/) and is now available in
6033 demos/state_machine.
6036 *) New options added to the 'dgst' utility for signature
6037 generation and verification.
6040 *) Unrecognized PKCS#7 content types are now handled via a
6041 catch all ASN1_TYPE structure. This allows unsupported
6042 types to be stored as a "blob" and an application can
6043 encode and decode it manually.
6046 *) Fix various signed/unsigned issues to make a_strex.c
6048 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
6050 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
6051 length if passed a buffer. ASN1_INTEGER_to_BN failed
6052 if passed a NULL BN and its argument was negative.
6053 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
6055 *) Modification to PKCS#7 encoding routines to output definite
6056 length encoding. Since currently the whole structures are in
6057 memory there's not real point in using indefinite length
6058 constructed encoding. However if OpenSSL is compiled with
6059 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
6062 *) Added BIO_vprintf() and BIO_vsnprintf().
6065 *) Added more prefixes to parse for in the the strings written
6066 through a logging bio, to cover all the levels that are available
6067 through syslog. The prefixes are now:
6069 PANIC, EMERG, EMR => LOG_EMERG
6070 ALERT, ALR => LOG_ALERT
6071 CRIT, CRI => LOG_CRIT
6072 ERROR, ERR => LOG_ERR
6073 WARNING, WARN, WAR => LOG_WARNING
6074 NOTICE, NOTE, NOT => LOG_NOTICE
6075 INFO, INF => LOG_INFO
6076 DEBUG, DBG => LOG_DEBUG
6078 and as before, if none of those prefixes are present at the
6079 beginning of the string, LOG_ERR is chosen.
6081 On Win32, the LOG_* levels are mapped according to this:
6083 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
6084 LOG_WARNING => EVENTLOG_WARNING_TYPE
6085 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
6089 *) Made it possible to reconfigure with just the configuration
6090 argument "reconf" or "reconfigure". The command line arguments
6091 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
6092 and are retrieved from there when reconfiguring.
6096 [Assar Westerlund <assar@sics.se>, Richard Levitte]
6098 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
6101 *) The obj_dat.pl script was messing up the sorting of object
6102 names. The reason was that it compared the quoted version
6103 of strings as a result "OCSP" > "OCSP Signing" because
6104 " > SPACE. Changed script to store unquoted versions of
6105 names and add quotes on output. It was also omitting some
6106 names from the lookup table if they were given a default
6107 value (that is if SN is missing it is given the same
6108 value as LN and vice versa), these are now added on the
6109 grounds that if an object has a name we should be able to
6110 look it up. Finally added warning output when duplicate
6111 short or long names are found.
6114 *) Changes needed for Tandem NSK.
6115 [Scott Uroff <scott@xypro.com>]
6117 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
6118 RSA_padding_check_SSLv23(), special padding was never detected
6119 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
6120 version rollback attacks was not effective.
6122 In s23_clnt.c, don't use special rollback-attack detection padding
6123 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
6124 client; similarly, in s23_srvr.c, don't do the rollback check if
6125 SSL 2.0 is the only protocol enabled in the server.
6128 *) Make it possible to get hexdumps of unprintable data with 'openssl
6129 asn1parse'. By implication, the functions ASN1_parse_dump() and
6130 BIO_dump_indent() are added.
6133 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
6134 these print out strings and name structures based on various
6135 flags including RFC2253 support and proper handling of
6136 multibyte characters. Added options to the 'x509' utility
6137 to allow the various flags to be set.
6140 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
6141 Also change the functions X509_cmp_current_time() and
6142 X509_gmtime_adj() work with an ASN1_TIME structure,
6143 this will enable certificates using GeneralizedTime in validity
6144 dates to be checked.
6147 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
6148 negative public key encodings) on by default,
6149 NO_NEG_PUBKEY_BUG can be set to disable it.
6152 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
6153 content octets. An i2c_ASN1_OBJECT is unnecessary because
6154 the encoding can be trivially obtained from the structure.
6157 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
6158 not read locks (CRYPTO_r_[un]lock).
6161 *) A first attempt at creating official support for shared
6162 libraries through configuration. I've kept it so the
6163 default is static libraries only, and the OpenSSL programs
6164 are always statically linked for now, but there are
6165 preparations for dynamic linking in place.
6166 This has been tested on Linux and Tru64.
6169 *) Randomness polling function for Win9x, as described in:
6170 Peter Gutmann, Software Generation of Practically Strong
6174 *) Fix so PRNG is seeded in req if using an already existing
6178 *) New options to smime application. -inform and -outform
6179 allow alternative formats for the S/MIME message including
6180 PEM and DER. The -content option allows the content to be
6181 specified separately. This should allow things like Netscape
6182 form signing output easier to verify.
6185 *) Fix the ASN1 encoding of tags using the 'long form'.
6188 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
6189 STRING types. These convert content octets to and from the
6190 underlying type. The actual tag and length octets are
6191 already assumed to have been read in and checked. These
6192 are needed because all other string types have virtually
6193 identical handling apart from the tag. By having versions
6194 of the ASN1 functions that just operate on content octets
6195 IMPLICIT tagging can be handled properly. It also allows
6196 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
6197 and ASN1_INTEGER are identical apart from the tag.
6200 *) Change the handling of OID objects as follows:
6202 - New object identifiers are inserted in objects.txt, following
6203 the syntax given in objects.README.
6204 - objects.pl is used to process obj_mac.num and create a new
6206 - obj_dat.pl is used to create a new obj_dat.h, using the data in
6209 This is currently kind of a hack, and the perl code in objects.pl
6210 isn't very elegant, but it works as I intended. The simplest way
6211 to check that it worked correctly is to look in obj_dat.h and
6212 check the array nid_objs and make sure the objects haven't moved
6213 around (this is important!). Additions are OK, as well as
6214 consistent name changes.
6217 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
6220 *) Addition of the command line parameter '-rand file' to 'openssl req'.
6221 The given file adds to whatever has already been seeded into the
6222 random pool through the RANDFILE configuration file option or
6223 environment variable, or the default random state file.
6226 *) mkstack.pl now sorts each macro group into lexical order.
6227 Previously the output order depended on the order the files
6228 appeared in the directory, resulting in needless rewriting
6232 *) Patches to make OpenSSL compile under Win32 again. Mostly
6233 work arounds for the VC++ problem that it treats func() as
6234 func(void). Also stripped out the parts of mkdef.pl that
6235 added extra typesafe functions: these no longer exist.
6238 *) Reorganisation of the stack code. The macros are now all
6239 collected in safestack.h . Each macro is defined in terms of
6240 a "stack macro" of the form SKM_<name>(type, a, b). The
6241 DEBUG_SAFESTACK is now handled in terms of function casts,
6242 this has the advantage of retaining type safety without the
6243 use of additional functions. If DEBUG_SAFESTACK is not defined
6244 then the non typesafe macros are used instead. Also modified the
6245 mkstack.pl script to handle the new form. Needs testing to see
6246 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
6247 the default if no major problems. Similar behaviour for ASN1_SET_OF
6248 and PKCS12_STACK_OF.
6251 *) When some versions of IIS use the 'NET' form of private key the
6252 key derivation algorithm is different. Normally MD5(password) is
6253 used as a 128 bit RC4 key. In the modified case
6254 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
6255 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
6256 as the old Netscape_RSA functions except they have an additional
6257 'sgckey' parameter which uses the modified algorithm. Also added
6258 an -sgckey command line option to the rsa utility. Thanks to
6259 Adrian Peck <bertie@ncipher.com> for posting details of the modified
6260 algorithm to openssl-dev.
6263 *) The evp_local.h macros were using 'c.##kname' which resulted in
6264 invalid expansion on some systems (SCO 5.0.5 for example).
6265 Corrected to 'c.kname'.
6266 [Phillip Porch <root@theporch.com>]
6268 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
6269 a STACK of email addresses from a certificate or request, these look
6270 in the subject name and the subject alternative name extensions and
6271 omit any duplicate addresses.
6274 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
6275 This makes DSA verification about 2 % faster.
6278 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
6279 (meaning that now 2^5 values will be precomputed, which is only 4 KB
6280 plus overhead for 1024 bit moduli).
6281 This makes exponentiations about 0.5 % faster for 1024 bit
6282 exponents (as measured by "openssl speed rsa2048").
6285 *) Rename memory handling macros to avoid conflicts with other
6287 Malloc => OPENSSL_malloc
6288 Malloc_locked => OPENSSL_malloc_locked
6289 Realloc => OPENSSL_realloc
6290 Free => OPENSSL_free
6293 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
6294 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
6297 *) CygWin32 support.
6298 [John Jarvie <jjarvie@newsguy.com>]
6300 *) The type-safe stack code has been rejigged. It is now only compiled
6301 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
6302 by default all type-specific stack functions are "#define"d back to
6303 standard stack functions. This results in more streamlined output
6304 but retains the type-safety checking possibilities of the original
6308 *) The STACK code has been cleaned up, and certain type declarations
6309 that didn't make a lot of sense have been brought in line. This has
6310 also involved a cleanup of sorts in safestack.h to more correctly
6311 map type-safe stack functions onto their plain stack counterparts.
6312 This work has also resulted in a variety of "const"ifications of
6313 lots of the code, especially "_cmp" operations which should normally
6314 be prototyped with "const" parameters anyway.
6317 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
6318 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
6319 (The PRNG state consists of two parts, the large pool 'state' and 'md',
6320 where all of 'md' is used each time the PRNG is used, but 'state'
6321 is used only indexed by a cyclic counter. As entropy may not be
6322 well distributed from the beginning, 'md' is important as a
6323 chaining variable. However, the output function chains only half
6324 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
6325 all of 'md', and seeding with STATE_SIZE dummy bytes will result
6326 in all of 'state' being rewritten, with the new values depending
6327 on virtually all of 'md'. This overcomes the 80 bit limitation.)
6330 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
6331 the handshake is continued after ssl_verify_cert_chain();
6332 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
6333 can lead to 'unexplainable' connection aborts later.
6334 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
6336 *) Major EVP API cipher revision.
6337 Add hooks for extra EVP features. This allows various cipher
6338 parameters to be set in the EVP interface. Support added for variable
6339 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
6340 setting of RC2 and RC5 parameters.
6342 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
6345 Remove lots of duplicated code from the EVP library. For example *every*
6346 cipher init() function handles the 'iv' in the same way according to the
6347 cipher mode. They also all do nothing if the 'key' parameter is NULL and
6348 for CFB and OFB modes they zero ctx->num.
6350 New functionality allows removal of S/MIME code RC2 hack.
6352 Most of the routines have the same form and so can be declared in terms
6355 By shifting this to the top level EVP_CipherInit() it can be removed from
6356 all individual ciphers. If the cipher wants to handle IVs or keys
6357 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
6360 Change lots of functions like EVP_EncryptUpdate() to now return a
6361 value: although software versions of the algorithms cannot fail
6362 any installed hardware versions can.
6365 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
6366 this option is set, tolerate broken clients that send the negotiated
6367 protocol version number instead of the requested protocol version
6371 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
6372 i.e. non-zero for export ciphersuites, zero otherwise.
6373 Previous versions had this flag inverted, inconsistent with
6374 rsa_tmp_cb (..._TMP_RSA_CB).
6375 [Bodo Moeller; problem reported by Amit Chopra]
6377 *) Add missing DSA library text string. Work around for some IIS
6378 key files with invalid SEQUENCE encoding.
6381 *) Add a document (doc/standards.txt) that list all kinds of standards
6382 and so on that are implemented in OpenSSL.
6385 *) Enhance c_rehash script. Old version would mishandle certificates
6386 with the same subject name hash and wouldn't handle CRLs at all.
6387 Added -fingerprint option to crl utility, to support new c_rehash
6391 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
6394 *) Fix for SSL server purpose checking. Server checking was
6395 rejecting certificates which had extended key usage present
6396 but no ssl client purpose.
6397 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
6399 *) Make PKCS#12 code work with no password. The PKCS#12 spec
6400 is a little unclear about how a blank password is handled.
6401 Since the password in encoded as a BMPString with terminating
6402 double NULL a zero length password would end up as just the
6403 double NULL. However no password at all is different and is
6404 handled differently in the PKCS#12 key generation code. NS
6405 treats a blank password as zero length. MSIE treats it as no
6406 password on export: but it will try both on import. We now do
6407 the same: PKCS12_parse() tries zero length and no password if
6408 the password is set to "" or NULL (NULL is now a valid password:
6409 it wasn't before) as does the pkcs12 application.
6412 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
6413 perror when PEM_read_bio_X509_REQ fails, the error message must
6414 be obtained from the error queue.
6417 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
6418 it in ERR_remove_state if appropriate, and change ERR_get_state
6419 accordingly to avoid race conditions (this is necessary because
6420 thread_hash is no longer constant once set).
6423 *) Bugfix for linux-elf makefile.one.
6426 *) RSA_get_default_method() will now cause a default
6427 RSA_METHOD to be chosen if one doesn't exist already.
6428 Previously this was only set during a call to RSA_new()
6429 or RSA_new_method(NULL) meaning it was possible for
6430 RSA_get_default_method() to return NULL.
6433 *) Added native name translation to the existing DSO code
6434 that will convert (if the flag to do so is set) filenames
6435 that are sufficiently small and have no path information
6436 into a canonical native form. Eg. "blah" converted to
6437 "libblah.so" or "blah.dll" etc.
6440 *) New function ERR_error_string_n(e, buf, len) which is like
6441 ERR_error_string(e, buf), but writes at most 'len' bytes
6442 including the 0 terminator. For ERR_error_string_n, 'buf'
6444 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
6446 *) CONF library reworked to become more general. A new CONF
6447 configuration file reader "class" is implemented as well as a
6448 new functions (NCONF_*, for "New CONF") to handle it. The now
6449 old CONF_* functions are still there, but are reimplemented to
6450 work in terms of the new functions. Also, a set of functions
6451 to handle the internal storage of the configuration data is
6452 provided to make it easier to write new configuration file
6453 reader "classes" (I can definitely see something reading a
6454 configuration file in XML format, for example), called _CONF_*,
6455 or "the configuration storage API"...
6457 The new configuration file reading functions are:
6459 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
6460 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
6462 NCONF_default, NCONF_WIN32
6464 NCONF_dump_fp, NCONF_dump_bio
6466 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
6467 NCONF_new creates a new CONF object. This works in the same way
6468 as other interfaces in OpenSSL, like the BIO interface.
6469 NCONF_dump_* dump the internal storage of the configuration file,
6470 which is useful for debugging. All other functions take the same
6471 arguments as the old CONF_* functions wth the exception of the
6472 first that must be a `CONF *' instead of a `LHASH *'.
6474 To make it easer to use the new classes with the old CONF_* functions,
6475 the function CONF_set_default_method is provided.
6478 *) Add '-tls1' option to 'openssl ciphers', which was already
6479 mentioned in the documentation but had not been implemented.
6480 (This option is not yet really useful because even the additional
6481 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
6484 *) Initial DSO code added into libcrypto for letting OpenSSL (and
6485 OpenSSL-based applications) load shared libraries and bind to
6486 them in a portable way.
6487 [Geoff Thorpe, with contributions from Richard Levitte]
6489 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
6491 *) Make sure _lrotl and _lrotr are only used with MSVC.
6493 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
6494 (the default implementation of RAND_status).
6496 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
6497 to '-clrext' (= clear extensions), as intended and documented.
6498 [Bodo Moeller; inconsistency pointed out by Michael Attili
6501 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
6502 was larger than the MD block size.
6503 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
6505 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
6506 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
6507 using the passed key: if the passed key was a private key the result
6508 of X509_print(), for example, would be to print out all the private key
6512 *) des_quad_cksum() byte order bug fix.
6513 [Ulf Möller, using the problem description in krb4-0.9.7, where
6514 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
6516 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
6518 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
6520 *) For easily testing in shell scripts whether some command
6521 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
6522 returns with exit code 0 iff no command of the given name is available.
6523 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
6524 the output goes to stdout and nothing is printed to stderr.
6525 Additional arguments are always ignored.
6527 Since for each cipher there is a command of the same name,
6528 the 'no-cipher' compilation switches can be tested this way.
6530 ('openssl no-XXX' is not able to detect pseudo-commands such
6531 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
6534 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
6537 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
6538 is set; it will be thrown away anyway because each handshake creates
6540 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
6541 to parameters -- in previous versions (since OpenSSL 0.9.3) the
6542 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
6543 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
6546 *) New s_client option -ign_eof: EOF at stdin is ignored, and
6547 'Q' and 'R' lose their special meanings (quit/renegotiate).
6548 This is part of what -quiet does; unlike -quiet, -ign_eof
6549 does not suppress any output.
6552 *) Add compatibility options to the purpose and trust code. The
6553 purpose X509_PURPOSE_ANY is "any purpose" which automatically
6554 accepts a certificate or CA, this was the previous behaviour,
6555 with all the associated security issues.
6557 X509_TRUST_COMPAT is the old trust behaviour: only and
6558 automatically trust self signed roots in certificate store. A
6559 new trust setting X509_TRUST_DEFAULT is used to specify that
6560 a purpose has no associated trust setting and it should instead
6561 use the value in the default purpose.
6564 *) Fix the PKCS#8 DSA private key code so it decodes keys again
6565 and fix a memory leak.
6568 *) In util/mkerr.pl (which implements 'make errors'), preserve
6569 reason strings from the previous version of the .c file, as
6570 the default to have only downcase letters (and digits) in
6571 automatically generated reasons codes is not always appropriate.
6574 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
6575 using strerror. Previously, ERR_reason_error_string() returned
6576 library names as reason strings for SYSerr; but SYSerr is a special
6577 case where small numbers are errno values, not library numbers.
6580 *) Add '-dsaparam' option to 'openssl dhparam' application. This
6581 converts DSA parameters into DH parameters. (When creating parameters,
6582 DSA_generate_parameters is used.)
6585 *) Include 'length' (recommended exponent length) in C code generated
6586 by 'openssl dhparam -C'.
6589 *) The second argument to set_label in perlasm was already being used
6590 so couldn't be used as a "file scope" flag. Moved to third argument
6594 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
6595 instead of RAND_bytes for encryption IVs and salts.
6598 *) Include RAND_status() into RAND_METHOD instead of implementing
6599 it only for md_rand.c Otherwise replacing the PRNG by calling
6600 RAND_set_rand_method would be impossible.
6603 *) Don't let DSA_generate_key() enter an infinite loop if the random
6604 number generation fails.
6607 *) New 'rand' application for creating pseudo-random output.
6610 *) Added configuration support for Linux/IA64
6611 [Rolf Haberrecker <rolf@suse.de>]
6613 *) Assembler module support for Mingw32.
6616 *) Shared library support for HPUX (in shlib/).
6617 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
6619 *) Shared library support for Solaris gcc.
6620 [Lutz Behnke <behnke@trustcenter.de>]
6622 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
6624 *) PKCS7_encrypt() was adding text MIME headers twice because they
6625 were added manually and by SMIME_crlf_copy().
6628 *) In bntest.c don't call BN_rand with zero bits argument.
6629 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
6631 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
6632 case was implemented. This caused BN_div_recp() to fail occasionally.
6635 *) Add an optional second argument to the set_label() in the perl
6636 assembly language builder. If this argument exists and is set
6637 to 1 it signals that the assembler should use a symbol whose
6638 scope is the entire file, not just the current function. This
6639 is needed with MASM which uses the format label:: for this scope.
6640 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
6642 *) Change the ASN1 types so they are typedefs by default. Before
6643 almost all types were #define'd to ASN1_STRING which was causing
6644 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
6648 *) Change names of new functions to the new get1/get0 naming
6649 convention: After 'get1', the caller owns a reference count
6650 and has to call ..._free; 'get0' returns a pointer to some
6651 data structure without incrementing reference counters.
6652 (Some of the existing 'get' functions increment a reference
6653 counter, some don't.)
6654 Similarly, 'set1' and 'add1' functions increase reference
6655 counters or duplicate objects.
6658 *) Allow for the possibility of temp RSA key generation failure:
6659 the code used to assume it always worked and crashed on failure.
6662 *) Fix potential buffer overrun problem in BIO_printf().
6663 [Ulf Möller, using public domain code by Patrick Powell; problem
6664 pointed out by David Sacerdote <das33@cornell.edu>]
6666 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
6667 RAND_egd() and RAND_status(). In the command line application,
6668 the EGD socket can be specified like a seed file using RANDFILE
6672 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
6673 Some CAs (e.g. Verisign) distribute certificates in this form.
6676 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
6677 list to exclude them. This means that no special compilation option
6678 is needed to use anonymous DH: it just needs to be included in the
6682 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
6683 EVP_MD_type. The old functionality is available in a new macro called
6684 EVP_MD_md(). Change code that uses it and update docs.
6687 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
6688 where the 'void *' argument is replaced by a function pointer argument.
6689 Previously 'void *' was abused to point to functions, which works on
6690 many platforms, but is not correct. As these functions are usually
6691 called by macros defined in OpenSSL header files, most source code
6692 should work without changes.
6695 *) <openssl/opensslconf.h> (which is created by Configure) now contains
6696 sections with information on -D... compiler switches used for
6697 compiling the library so that applications can see them. To enable
6698 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
6699 must be defined. E.g.,
6700 #define OPENSSL_ALGORITHM_DEFINES
6701 #include <openssl/opensslconf.h>
6702 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
6703 [Richard Levitte, Ulf and Bodo Möller]
6705 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
6709 *) Change the 'other' type in certificate aux info to a STACK_OF
6710 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
6711 the required ASN1 format: arbitrary types determined by an OID.
6714 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
6715 argument to 'req'. This is not because the function is newer or
6716 better than others it just uses the work 'NEW' in the certificate
6717 request header lines. Some software needs this.
6720 *) Reorganise password command line arguments: now passwords can be
6721 obtained from various sources. Delete the PEM_cb function and make
6722 it the default behaviour: i.e. if the callback is NULL and the
6723 usrdata argument is not NULL interpret it as a null terminated pass
6724 phrase. If usrdata and the callback are NULL then the pass phrase
6725 is prompted for as usual.
6728 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
6729 the support is automatically enabled. The resulting binaries will
6730 autodetect the card and use it if present.
6731 [Ben Laurie and Compaq Inc.]
6733 *) Work around for Netscape hang bug. This sends certificate request
6734 and server done in one record. Since this is perfectly legal in the
6735 SSL/TLS protocol it isn't a "bug" option and is on by default. See
6736 the bugs/SSLv3 entry for more info.
6739 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
6742 *) Add -rand argument to smime and pkcs12 applications and read/write
6746 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
6749 *) Add command line password options to the remaining applications.
6752 *) Bug fix for BN_div_recp() for numerators with an even number of
6756 *) More tests in bntest.c, and changed test_bn output.
6759 *) ./config recognizes MacOS X now.
6762 *) Bug fix for BN_div() when the first words of num and divsor are
6763 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
6766 *) Add support for various broken PKCS#8 formats, and command line
6767 options to produce them.
6770 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
6771 get temporary BIGNUMs from a BN_CTX.
6774 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
6778 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
6779 include a #define from the old name to the new. The original intent
6780 was that statically linked binaries could for example just call
6781 SSLeay_add_all_ciphers() to just add ciphers to the table and not
6782 link with digests. This never worked becayse SSLeay_add_all_digests()
6783 and SSLeay_add_all_ciphers() were in the same source file so calling
6784 one would link with the other. They are now in separate source files.
6787 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
6790 *) Use a less unusual form of the Miller-Rabin primality test (it used
6791 a binary algorithm for exponentiation integrated into the Miller-Rabin
6792 loop, our standard modexp algorithms are faster).
6795 *) Support for the EBCDIC character set completed.
6796 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
6798 *) Source code cleanups: use const where appropriate, eliminate casts,
6799 use void * instead of char * in lhash.
6802 *) Bugfix: ssl3_send_server_key_exchange was not restartable
6803 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
6804 this the server could overwrite ephemeral keys that the client
6808 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
6809 using 50 iterations of the Rabin-Miller test.
6811 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
6812 iterations of the Rabin-Miller test as required by the appendix
6813 to FIPS PUB 186[-1]) instead of DSA_is_prime.
6814 As BN_is_prime_fasttest includes trial division, DSA parameter
6815 generation becomes much faster.
6817 This implies a change for the callback functions in DSA_is_prime
6818 and DSA_generate_parameters: The callback function is called once
6819 for each positive witness in the Rabin-Miller test, not just
6820 occasionally in the inner loop; and the parameters to the
6821 callback function now provide an iteration count for the outer
6822 loop rather than for the current invocation of the inner loop.
6823 DSA_generate_parameters additionally can call the callback
6824 function with an 'iteration count' of -1, meaning that a
6825 candidate has passed the trial division test (when q is generated
6826 from an application-provided seed, trial division is skipped).
6829 *) New function BN_is_prime_fasttest that optionally does trial
6830 division before starting the Rabin-Miller test and has
6831 an additional BN_CTX * argument (whereas BN_is_prime always
6832 has to allocate at least one BN_CTX).
6833 'callback(1, -1, cb_arg)' is called when a number has passed the
6834 trial division stage.
6837 *) Fix for bug in CRL encoding. The validity dates weren't being handled
6841 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
6844 *) New function BN_pseudo_rand().
6847 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
6848 bignum version of BN_from_montgomery() with the working code from
6849 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
6853 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
6854 made it impossible to use the same SSL_SESSION data structure in
6855 SSL2 clients in multiple threads.
6858 *) The return value of RAND_load_file() no longer counts bytes obtained
6859 by stat(). RAND_load_file(..., -1) is new and uses the complete file
6860 to seed the PRNG (previously an explicit byte count was required).
6861 [Ulf Möller, Bodo Möller]
6863 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
6864 used (char *) instead of (void *) and had casts all over the place.
6867 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
6870 *) Retain source code compatibility for BN_prime_checks macro:
6871 BN_is_prime(..., BN_prime_checks, ...) now uses
6872 BN_prime_checks_for_size to determine the appropriate number of
6873 Rabin-Miller iterations.
6876 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
6877 DH_CHECK_P_NOT_SAFE_PRIME.
6878 (Check if this is true? OpenPGP calls them "strong".)
6881 *) Merge the functionality of "dh" and "gendh" programs into a new program
6882 "dhparam". The old programs are retained for now but will handle DH keys
6883 (instead of parameters) in future.
6886 *) Make the ciphers, s_server and s_client programs check the return values
6887 when a new cipher list is set.
6890 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
6891 ciphers. Before when the 56bit ciphers were enabled the sorting was
6894 The syntax for the cipher sorting has been extended to support sorting by
6895 cipher-strength (using the strength_bits hard coded in the tables).
6896 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
6898 Fix a bug in the cipher-command parser: when supplying a cipher command
6899 string with an "undefined" symbol (neither command nor alphanumeric
6900 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
6901 an error is flagged.
6903 Due to the strength-sorting extension, the code of the
6904 ssl_create_cipher_list() function was completely rearranged. I hope that
6905 the readability was also increased :-)
6906 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6908 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
6909 for the first serial number and places 2 in the serial number file. This
6910 avoids problems when the root CA is created with serial number zero and
6911 the first user certificate has the same issuer name and serial number
6915 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
6916 the new code. Add documentation for this stuff.
6919 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
6920 X509_*() to X509at_*() on the grounds that they don't handle X509
6921 structures and behave in an analagous way to the X509v3 functions:
6922 they shouldn't be called directly but wrapper functions should be used
6925 So we also now have some wrapper functions that call the X509at functions
6926 when passed certificate requests. (TO DO: similar things can be done with
6927 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
6928 things. Some of these need some d2i or i2d and print functionality
6929 because they handle more complex structures.)
6932 *) Add missing #ifndefs that caused missing symbols when building libssl
6933 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
6934 NO_RSA in ssl/s2*.c.
6935 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
6937 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
6938 has a return value which indicates the quality of the random data
6939 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
6940 error queue. New function RAND_pseudo_bytes() generates output that is
6941 guaranteed to be unique but not unpredictable. RAND_add is like
6942 RAND_seed, but takes an extra argument for an entropy estimate
6943 (RAND_seed always assumes full entropy).
6946 *) Do more iterations of Rabin-Miller probable prime test (specifically,
6947 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
6948 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
6949 in crypto/bn/bn_prime.c for the complete table). This guarantees a
6950 false-positive rate of at most 2^-80 for random input.
6953 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
6956 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
6957 in the 0.9.5 release), this returns the chain
6958 from an X509_CTX structure with a dup of the stack and all
6959 the X509 reference counts upped: so the stack will exist
6960 after X509_CTX_cleanup() has been called. Modify pkcs12.c
6963 Also make SSL_SESSION_print() print out the verify return
6967 *) Add manpage for the pkcs12 command. Also change the default
6968 behaviour so MAC iteration counts are used unless the new
6969 -nomaciter option is used. This improves file security and
6970 only older versions of MSIE (4.0 for example) need it.
6973 *) Honor the no-xxx Configure options when creating .DEF files.
6976 *) Add PKCS#10 attributes to field table: challengePassword,
6977 unstructuredName and unstructuredAddress. These are taken from
6978 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
6979 international characters are used.
6981 More changes to X509_ATTRIBUTE code: allow the setting of types
6982 based on strings. Remove the 'loc' parameter when adding
6983 attributes because these will be a SET OF encoding which is sorted
6987 *) Initial changes to the 'req' utility to allow request generation
6988 automation. This will allow an application to just generate a template
6989 file containing all the field values and have req construct the
6992 Initial support for X509_ATTRIBUTE handling. Stacks of these are
6993 used all over the place including certificate requests and PKCS#7
6994 structures. They are currently handled manually where necessary with
6995 some primitive wrappers for PKCS#7. The new functions behave in a
6996 manner analogous to the X509 extension functions: they allow
6997 attributes to be looked up by NID and added.
6999 Later something similar to the X509V3 code would be desirable to
7000 automatically handle the encoding, decoding and printing of the
7001 more complex types. The string types like challengePassword can
7002 be handled by the string table functions.
7004 Also modified the multi byte string table handling. Now there is
7005 a 'global mask' which masks out certain types. The table itself
7006 can use the flag STABLE_NO_MASK to ignore the mask setting: this
7007 is useful when for example there is only one permissible type
7008 (as in countryName) and using the mask might result in no valid
7012 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
7013 SSL_get_peer_finished to allow applications to obtain the latest
7014 Finished messages sent to the peer or expected from the peer,
7015 respectively. (SSL_get_peer_finished is usually the Finished message
7016 actually received from the peer, otherwise the protocol will be aborted.)
7018 As the Finished message are message digests of the complete handshake
7019 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
7020 be used for external authentication procedures when the authentication
7021 provided by SSL/TLS is not desired or is not enough.
7024 *) Enhanced support for Alpha Linux is added. Now ./config checks if
7025 the host supports BWX extension and if Compaq C is present on the
7026 $PATH. Just exploiting of the BWX extension results in 20-30%
7027 performance kick for some algorithms, e.g. DES and RC4 to mention
7028 a couple. Compaq C in turn generates ~20% faster code for MD5 and
7032 *) Add support for MS "fast SGC". This is arguably a violation of the
7033 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
7034 weak crypto and after checking the certificate is SGC a second one
7035 with strong crypto. MS SGC stops the first handshake after receiving
7036 the server certificate message and sends a second client hello. Since
7037 a server will typically do all the time consuming operations before
7038 expecting any further messages from the client (server key exchange
7039 is the most expensive) there is little difference between the two.
7041 To get OpenSSL to support MS SGC we have to permit a second client
7042 hello message after we have sent server done. In addition we have to
7043 reset the MAC if we do get this second client hello.
7046 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
7047 if a DER encoded private key is RSA or DSA traditional format. Changed
7048 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
7049 format DER encoded private key. Newer code should use PKCS#8 format which
7050 has the key type encoded in the ASN1 structure. Added DER private key
7051 support to pkcs8 application.
7054 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
7055 ciphersuites has been selected (as required by the SSL 3/TLS 1
7056 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
7057 is set, we interpret this as a request to violate the specification
7058 (the worst that can happen is a handshake failure, and 'correct'
7059 behaviour would result in a handshake failure anyway).
7062 *) In SSL_CTX_add_session, take into account that there might be multiple
7063 SSL_SESSION structures with the same session ID (e.g. when two threads
7064 concurrently obtain them from an external cache).
7065 The internal cache can handle only one SSL_SESSION with a given ID,
7066 so if there's a conflict, we now throw out the old one to achieve
7070 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
7071 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
7072 some routines that use cipher OIDs: some ciphers do not have OIDs
7073 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
7077 *) Simplify the trust setting structure and code. Now we just have
7078 two sequences of OIDs for trusted and rejected settings. These will
7079 typically have values the same as the extended key usage extension
7080 and any application specific purposes.
7082 The trust checking code now has a default behaviour: it will just
7083 check for an object with the same NID as the passed id. Functions can
7084 be provided to override either the default behaviour or the behaviour
7085 for a given id. SSL client, server and email already have functions
7086 in place for compatibility: they check the NID and also return "trusted"
7087 if the certificate is self signed.
7090 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
7091 traditional format into an EVP_PKEY structure.
7094 *) Add a password callback function PEM_cb() which either prompts for
7095 a password if usr_data is NULL or otherwise assumes it is a null
7096 terminated password. Allow passwords to be passed on command line
7097 environment or config files in a few more utilities.
7100 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
7101 keys. Add some short names for PKCS#8 PBE algorithms and allow them
7102 to be specified on the command line for the pkcs8 and pkcs12 utilities.
7103 Update documentation.
7106 *) Support for ASN1 "NULL" type. This could be handled before by using
7107 ASN1_TYPE but there wasn't any function that would try to read a NULL
7108 and produce an error if it couldn't. For compatibility we also have
7109 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
7110 don't allocate anything because they don't need to.
7113 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
7115 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
7117 *) Rebuild of the memory allocation routines used by OpenSSL code and
7118 possibly others as well. The purpose is to make an interface that
7119 provide hooks so anyone can build a separate set of allocation and
7120 deallocation routines to be used by OpenSSL, for example memory
7121 pool implementations, or something else, which was previously hard
7122 since Malloc(), Realloc() and Free() were defined as macros having
7123 the values malloc, realloc and free, respectively (except for Win32
7124 compilations). The same is provided for memory debugging code.
7125 OpenSSL already comes with functionality to find memory leaks, but
7126 this gives people a chance to debug other memory problems.
7128 With these changes, a new set of functions and macros have appeared:
7130 CRYPTO_set_mem_debug_functions() [F]
7131 CRYPTO_get_mem_debug_functions() [F]
7132 CRYPTO_dbg_set_options() [F]
7133 CRYPTO_dbg_get_options() [F]
7134 CRYPTO_malloc_debug_init() [M]
7136 The memory debug functions are NULL by default, unless the library
7137 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
7138 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
7139 gives the standard debugging functions that come with OpenSSL) or
7140 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
7141 provided by the library user) must be used. When the standard
7142 debugging functions are used, CRYPTO_dbg_set_options can be used to
7143 request additional information:
7144 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
7145 the CRYPTO_MDEBUG_xxx macro when compiling the library.
7147 Also, things like CRYPTO_set_mem_functions will always give the
7148 expected result (the new set of functions is used for allocation
7149 and deallocation) at all times, regardless of platform and compiler
7152 To finish it up, some functions that were never use in any other
7153 way than through macros have a new API and new semantic:
7156 CRYPTO_dbg_realloc()
7159 All macros of value have retained their old syntax.
7160 [Richard Levitte and Bodo Moeller]
7162 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
7163 ordering of SMIMECapabilities wasn't in "strength order" and there
7164 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
7168 *) Some ASN1 types with illegal zero length encoding (INTEGER,
7169 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
7170 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
7172 *) Merge in my S/MIME library for OpenSSL. This provides a simple
7173 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
7174 functionality to handle multipart/signed properly) and a utility
7175 called 'smime' to call all this stuff. This is based on code I
7176 originally wrote for Celo who have kindly allowed it to be
7177 included in OpenSSL.
7180 *) Add variants des_set_key_checked and des_set_key_unchecked of
7181 des_set_key (aka des_key_sched). Global variable des_check_key
7182 decides which of these is called by des_set_key; this way
7183 des_check_key behaves as it always did, but applications and
7184 the library itself, which was buggy for des_check_key == 1,
7185 have a cleaner way to pick the version they need.
7188 *) New function PKCS12_newpass() which changes the password of a
7192 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
7193 dynamic mix. In both cases the ids can be used as an index into the
7194 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
7195 functions so they accept a list of the field values and the
7196 application doesn't need to directly manipulate the X509_TRUST
7200 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
7204 *) Modify the way the V3 extension code looks up extensions. This now
7205 works in a similar way to the object code: we have some "standard"
7206 extensions in a static table which is searched with OBJ_bsearch()
7207 and the application can add dynamic ones if needed. The file
7208 crypto/x509v3/ext_dat.h now has the info: this file needs to be
7209 updated whenever a new extension is added to the core code and kept
7210 in ext_nid order. There is a simple program 'tabtest.c' which checks
7211 this. New extensions are not added too often so this file can readily
7212 be maintained manually.
7214 There are two big advantages in doing things this way. The extensions
7215 can be looked up immediately and no longer need to be "added" using
7216 X509V3_add_standard_extensions(): this function now does nothing.
7217 [Side note: I get *lots* of email saying the extension code doesn't
7218 work because people forget to call this function]
7219 Also no dynamic allocation is done unless new extensions are added:
7220 so if we don't add custom extensions there is no need to call
7221 X509V3_EXT_cleanup().
7224 *) Modify enc utility's salting as follows: make salting the default. Add a
7225 magic header, so unsalted files fail gracefully instead of just decrypting
7226 to garbage. This is because not salting is a big security hole, so people
7227 should be discouraged from doing it.
7230 *) Fixes and enhancements to the 'x509' utility. It allowed a message
7231 digest to be passed on the command line but it only used this
7232 parameter when signing a certificate. Modified so all relevant
7233 operations are affected by the digest parameter including the
7234 -fingerprint and -x509toreq options. Also -x509toreq choked if a
7235 DSA key was used because it didn't fix the digest.
7238 *) Initial certificate chain verify code. Currently tests the untrusted
7239 certificates for consistency with the verify purpose (which is set
7240 when the X509_STORE_CTX structure is set up) and checks the pathlength.
7242 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
7243 this is because it will reject chains with invalid extensions whereas
7244 every previous version of OpenSSL and SSLeay made no checks at all.
7246 Trust code: checks the root CA for the relevant trust settings. Trust
7247 settings have an initial value consistent with the verify purpose: e.g.
7248 if the verify purpose is for SSL client use it expects the CA to be
7249 trusted for SSL client use. However the default value can be changed to
7250 permit custom trust settings: one example of this would be to only trust
7251 certificates from a specific "secure" set of CAs.
7253 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
7254 which should be used for version portability: especially since the
7255 verify structure is likely to change more often now.
7257 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
7258 to set them. If not set then assume SSL clients will verify SSL servers
7261 Two new options to the verify program: -untrusted allows a set of
7262 untrusted certificates to be passed in and -purpose which sets the
7263 intended purpose of the certificate. If a purpose is set then the
7264 new chain verify code is used to check extension consistency.
7267 *) Support for the authority information access extension.
7270 *) Modify RSA and DSA PEM read routines to transparently handle
7271 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
7272 public keys in a format compatible with certificate
7273 SubjectPublicKeyInfo structures. Unfortunately there were already
7274 functions called *_PublicKey_* which used various odd formats so
7275 these are retained for compatibility: however the DSA variants were
7276 never in a public release so they have been deleted. Changed dsa/rsa
7277 utilities to handle the new format: note no releases ever handled public
7278 keys so we should be OK.
7280 The primary motivation for this change is to avoid the same fiasco
7281 that dogs private keys: there are several incompatible private key
7282 formats some of which are standard and some OpenSSL specific and
7283 require various evil hacks to allow partial transparent handling and
7284 even then it doesn't work with DER formats. Given the option anything
7285 other than PKCS#8 should be dumped: but the other formats have to
7286 stay in the name of compatibility.
7288 With public keys and the benefit of hindsight one standard format
7289 is used which works with EVP_PKEY, RSA or DSA structures: though
7290 it clearly returns an error if you try to read the wrong kind of key.
7292 Added a -pubkey option to the 'x509' utility to output the public key.
7293 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
7294 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
7295 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
7296 that do the same as the EVP_PKEY_assign_*() except they up the
7297 reference count of the added key (they don't "swallow" the
7301 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
7302 CRLs would fail if the file contained no certificates or no CRLs:
7303 added a new function to read in both types and return the number
7304 read: this means that if none are read it will be an error. The
7305 DER versions of the certificate and CRL reader would always fail
7306 because it isn't possible to mix certificates and CRLs in DER format
7307 without choking one or the other routine. Changed this to just read
7308 a certificate: this is the best we can do. Also modified the code
7309 in apps/verify.c to take notice of return codes: it was previously
7310 attempting to read in certificates from NULL pointers and ignoring
7311 any errors: this is one reason why the cert and CRL reader seemed
7312 to work. It doesn't check return codes from the default certificate
7313 routines: these may well fail if the certificates aren't installed.
7316 *) Code to support otherName option in GeneralName.
7319 *) First update to verify code. Change the verify utility
7320 so it warns if it is passed a self signed certificate:
7321 for consistency with the normal behaviour. X509_verify
7322 has been modified to it will now verify a self signed
7323 certificate if *exactly* the same certificate appears
7324 in the store: it was previously impossible to trust a
7325 single self signed certificate. This means that:
7326 openssl verify ss.pem
7327 now gives a warning about a self signed certificate but
7328 openssl verify -CAfile ss.pem ss.pem
7332 *) For servers, store verify_result in SSL_SESSION data structure
7333 (and add it to external session representation).
7334 This is needed when client certificate verifications fails,
7335 but an application-provided verification callback (set by
7336 SSL_CTX_set_cert_verify_callback) allows accepting the session
7337 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
7338 but returns 1): When the session is reused, we have to set
7339 ssl->verify_result to the appropriate error code to avoid
7341 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
7343 *) Fix a bug in the new PKCS#7 code: it didn't consider the
7344 case in PKCS7_dataInit() where the signed PKCS7 structure
7345 didn't contain any existing data because it was being created.
7346 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
7348 *) Add a salt to the key derivation routines in enc.c. This
7349 forms the first 8 bytes of the encrypted file. Also add a
7350 -S option to allow a salt to be input on the command line.
7353 *) New function X509_cmp(). Oddly enough there wasn't a function
7354 to compare two certificates. We do this by working out the SHA1
7355 hash and comparing that. X509_cmp() will be needed by the trust
7359 *) SSL_get1_session() is like SSL_get_session(), but increments
7360 the reference count in the SSL_SESSION returned.
7361 [Geoff Thorpe <geoff@eu.c2.net>]
7363 *) Fix for 'req': it was adding a null to request attributes.
7364 Also change the X509_LOOKUP and X509_INFO code to handle
7365 certificate auxiliary information.
7368 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
7372 *) Add the possibility to add extra information to the memory leak
7373 detecting output, to form tracebacks, showing from where each
7374 allocation was originated: CRYPTO_push_info("constant string") adds
7375 the string plus current file name and line number to a per-thread
7376 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
7377 is like calling CYRPTO_pop_info() until the stack is empty.
7378 Also updated memory leak detection code to be multi-thread-safe.
7381 *) Add options -text and -noout to pkcs7 utility and delete the
7382 encryption options which never did anything. Update docs.
7385 *) Add options to some of the utilities to allow the pass phrase
7386 to be included on either the command line (not recommended on
7387 OSes like Unix) or read from the environment. Update the
7388 manpages and fix a few bugs.
7391 *) Add a few manpages for some of the openssl commands.
7394 *) Fix the -revoke option in ca. It was freeing up memory twice,
7395 leaking and not finding already revoked certificates.
7398 *) Extensive changes to support certificate auxiliary information.
7399 This involves the use of X509_CERT_AUX structure and X509_AUX
7400 functions. An X509_AUX function such as PEM_read_X509_AUX()
7401 can still read in a certificate file in the usual way but it
7402 will also read in any additional "auxiliary information". By
7403 doing things this way a fair degree of compatibility can be
7404 retained: existing certificates can have this information added
7405 using the new 'x509' options.
7407 Current auxiliary information includes an "alias" and some trust
7408 settings. The trust settings will ultimately be used in enhanced
7409 certificate chain verification routines: currently a certificate
7410 can only be trusted if it is self signed and then it is trusted
7414 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
7415 The problem was that one of the replacement routines had not been working
7416 since SSLeay releases. For now the offending routine has been replaced
7417 with non-optimised assembler. Even so, this now gives around 95%
7418 performance improvement for 1024 bit RSA signs.
7421 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
7422 handling. Most clients have the effective key size in bits equal to
7423 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
7424 A few however don't do this and instead use the size of the decrypted key
7425 to determine the RC2 key length and the AlgorithmIdentifier to determine
7426 the effective key length. In this case the effective key length can still
7427 be 40 bits but the key length can be 168 bits for example. This is fixed
7428 by manually forcing an RC2 key into the EVP_PKEY structure because the
7429 EVP code can't currently handle unusual RC2 key sizes: it always assumes
7430 the key length and effective key length are equal.
7433 *) Add a bunch of functions that should simplify the creation of
7434 X509_NAME structures. Now you should be able to do:
7435 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
7436 and have it automatically work out the correct field type and fill in
7437 the structures. The more adventurous can try:
7438 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
7439 and it will (hopefully) work out the correct multibyte encoding.
7442 *) Change the 'req' utility to use the new field handling and multibyte
7443 copy routines. Before the DN field creation was handled in an ad hoc
7444 way in req, ca, and x509 which was rather broken and didn't support
7445 BMPStrings or UTF8Strings. Since some software doesn't implement
7446 BMPStrings or UTF8Strings yet, they can be enabled using the config file
7447 using the dirstring_type option. See the new comment in the default
7448 openssl.cnf for more info.
7451 *) Make crypto/rand/md_rand.c more robust:
7452 - Assure unique random numbers after fork().
7453 - Make sure that concurrent threads access the global counter and
7454 md serializably so that we never lose entropy in them
7455 or use exactly the same state in multiple threads.
7456 Access to the large state is not always serializable because
7457 the additional locking could be a performance killer, and
7458 md should be large enough anyway.
7461 *) New file apps/app_rand.c with commonly needed functionality
7462 for handling the random seed file.
7464 Use the random seed file in some applications that previously did not:
7466 dsaparam -genkey (which also ignored its '-rand' option),
7469 x509 (when signing).
7470 Except on systems with /dev/urandom, it is crucial to have a random
7471 seed file at least for key creation, DSA signing, and for DH exchanges;
7472 for RSA signatures we could do without one.
7474 gendh and gendsa (unlike genrsa) used to read only the first byte
7475 of each file listed in the '-rand' option. The function as previously
7476 found in genrsa is now in app_rand.c and is used by all programs
7477 that support '-rand'.
7480 *) In RAND_write_file, use mode 0600 for creating files;
7481 don't just chmod when it may be too late.
7484 *) Report an error from X509_STORE_load_locations
7485 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
7488 *) New function ASN1_mbstring_copy() this copies a string in either
7489 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
7490 into an ASN1_STRING type. A mask of permissible types is passed
7491 and it chooses the "minimal" type to use or an error if not type
7495 *) Add function equivalents to the various macros in asn1.h. The old
7496 macros are retained with an M_ prefix. Code inside the library can
7497 use the M_ macros. External code (including the openssl utility)
7498 should *NOT* in order to be "shared library friendly".
7501 *) Add various functions that can check a certificate's extensions
7502 to see if it usable for various purposes such as SSL client,
7503 server or S/MIME and CAs of these types. This is currently
7504 VERY EXPERIMENTAL but will ultimately be used for certificate chain
7505 verification. Also added a -purpose flag to x509 utility to
7506 print out all the purposes.
7509 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
7513 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
7514 for, obtain and decode and extension and obtain its critical flag.
7515 This allows all the necessary extension code to be handled in a
7516 single function call.
7519 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
7520 platforms. See crypto/rc4/rc4_enc.c for further details.
7523 *) New -noout option to asn1parse. This causes no output to be produced
7524 its main use is when combined with -strparse and -out to extract data
7525 from a file (which may not be in ASN.1 format).
7528 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
7529 when producing the local key id.
7530 [Richard Levitte <levitte@stacken.kth.se>]
7532 *) New option -dhparam in s_server. This allows a DH parameter file to be
7533 stated explicitly. If it is not stated then it tries the first server
7534 certificate file. The previous behaviour hard coded the filename
7538 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
7539 a public key to be input or output. For example:
7540 openssl rsa -in key.pem -pubout -out pubkey.pem
7541 Also added necessary DSA public key functions to handle this.
7544 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
7545 in the message. This was handled by allowing
7546 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
7547 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
7549 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
7550 to the end of the strings whereas this didn't. This would cause problems
7551 if strings read with d2i_ASN1_bytes() were later modified.
7552 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
7554 *) Fix for base64 decode bug. When a base64 bio reads only one line of
7555 data and it contains EOF it will end up returning an error. This is
7556 caused by input 46 bytes long. The cause is due to the way base64
7557 BIOs find the start of base64 encoded data. They do this by trying a
7558 trial decode on each line until they find one that works. When they
7559 do a flag is set and it starts again knowing it can pass all the
7560 data directly through the decoder. Unfortunately it doesn't reset
7561 the context it uses. This means that if EOF is reached an attempt
7562 is made to pass two EOFs through the context and this causes the
7563 resulting error. This can also cause other problems as well. As is
7564 usual with these problems it takes *ages* to find and the fix is
7565 trivial: move one line.
7566 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
7568 *) Ugly workaround to get s_client and s_server working under Windows. The
7569 old code wouldn't work because it needed to select() on sockets and the
7570 tty (for keypresses and to see if data could be written). Win32 only
7571 supports select() on sockets so we select() with a 1s timeout on the
7572 sockets and then see if any characters are waiting to be read, if none
7573 are present then we retry, we also assume we can always write data to
7574 the tty. This isn't nice because the code then blocks until we've
7575 received a complete line of data and it is effectively polling the
7576 keyboard at 1s intervals: however it's quite a bit better than not
7577 working at all :-) A dedicated Windows application might handle this
7578 with an event loop for example.
7581 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
7582 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
7583 will be called when RSA_sign() and RSA_verify() are used. This is useful
7584 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
7585 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
7586 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
7587 This necessitated the support of an extra signature type NID_md5_sha1
7588 for SSL signatures and modifications to the SSL library to use it instead
7589 of calling RSA_public_decrypt() and RSA_private_encrypt().
7592 *) Add new -verify -CAfile and -CApath options to the crl program, these
7593 will lookup a CRL issuers certificate and verify the signature in a
7594 similar way to the verify program. Tidy up the crl program so it
7595 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
7596 less strict. It will now permit CRL extensions even if it is not
7597 a V2 CRL: this will allow it to tolerate some broken CRLs.
7600 *) Initialize all non-automatic variables each time one of the openssl
7601 sub-programs is started (this is necessary as they may be started
7602 multiple times from the "OpenSSL>" prompt).
7603 [Lennart Bang, Bodo Moeller]
7605 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
7606 removing all other RSA functionality (this is what NO_RSA does). This
7607 is so (for example) those in the US can disable those operations covered
7608 by the RSA patent while allowing storage and parsing of RSA keys and RSA
7612 *) Non-copying interface to BIO pairs.
7613 (still largely untested)
7616 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
7617 ASCII string. This was handled independently in various places before.
7620 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
7621 UTF8 strings a character at a time.
7624 *) Use client_version from client hello to select the protocol
7625 (s23_srvr.c) and for RSA client key exchange verification
7626 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
7629 *) Add various utility functions to handle SPKACs, these were previously
7630 handled by poking round in the structure internals. Added new function
7631 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
7632 print, verify and generate SPKACs. Based on an original idea from
7633 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
7636 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
7639 *) Allow the config file extension section to be overwritten on the
7640 command line. Based on an original idea from Massimiliano Pala
7641 <madwolf@comune.modena.it>. The new option is called -extensions
7642 and can be applied to ca, req and x509. Also -reqexts to override
7643 the request extensions in req and -crlexts to override the crl extensions
7647 *) Add new feature to the SPKAC handling in ca. Now you can include
7648 the same field multiple times by preceding it by "XXXX." for example:
7651 this is the same syntax as used in the req config file.
7654 *) Allow certificate extensions to be added to certificate requests. These
7655 are specified in a 'req_extensions' option of the req section of the
7656 config file. They can be printed out with the -text option to req but
7657 are otherwise ignored at present.
7660 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
7661 data read consists of only the final block it would not decrypted because
7662 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
7663 A misplaced 'break' also meant the decrypted final block might not be
7664 copied until the next read.
7667 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
7668 a few extra parameters to the DH structure: these will be useful if
7669 for example we want the value of 'q' or implement X9.42 DH.
7672 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
7673 provides hooks that allow the default DSA functions or functions on a
7674 "per key" basis to be replaced. This allows hardware acceleration and
7675 hardware key storage to be handled without major modification to the
7676 library. Also added low level modexp hooks and CRYPTO_EX structure and
7677 associated functions.
7680 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
7681 as "read only": it can't be written to and the buffer it points to will
7682 not be freed. Reading from a read only BIO is much more efficient than
7683 a normal memory BIO. This was added because there are several times when
7684 an area of memory needs to be read from a BIO. The previous method was
7685 to create a memory BIO and write the data to it, this results in two
7686 copies of the data and an O(n^2) reading algorithm. There is a new
7687 function BIO_new_mem_buf() which creates a read only memory BIO from
7688 an area of memory. Also modified the PKCS#7 routines to use read only
7692 *) Bugfix: ssl23_get_client_hello did not work properly when called in
7693 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
7694 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
7695 but a retry condition occured while trying to read the rest.
7698 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
7699 NID_pkcs7_encrypted by default: this was wrong since this should almost
7700 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
7701 the encrypted data type: this is a more sensible place to put it and it
7702 allows the PKCS#12 code to be tidied up that duplicated this
7706 *) Changed obj_dat.pl script so it takes its input and output files on
7707 the command line. This should avoid shell escape redirection problems
7711 *) Initial support for certificate extension requests, these are included
7712 in things like Xenroll certificate requests. Included functions to allow
7713 extensions to be obtained and added.
7716 *) -crlf option to s_client and s_server for sending newlines as
7717 CRLF (as required by many protocols).
7720 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
7722 *) Install libRSAglue.a when OpenSSL is built with RSAref.
7723 [Ralf S. Engelschall]
7725 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
7726 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
7728 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
7732 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
7733 DH parameters/keys (q is lost during that conversion, but the resulting
7734 DH parameters contain its length).
7736 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
7737 much faster than DH_generate_parameters (which creates parameters
7738 where p = 2*q + 1), and also the smaller q makes DH computations
7739 much more efficient (160-bit exponentiation instead of 1024-bit
7740 exponentiation); so this provides a convenient way to support DHE
7741 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
7742 utter importance to use
7743 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
7745 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
7746 when such DH parameters are used, because otherwise small subgroup
7747 attacks may become possible!
7750 *) Avoid memory leak in i2d_DHparams.
7753 *) Allow the -k option to be used more than once in the enc program:
7754 this allows the same encrypted message to be read by multiple recipients.
7757 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
7758 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
7759 it will always use the numerical form of the OID, even if it has a short
7763 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
7764 method only got called if p,q,dmp1,dmq1,iqmp components were present,
7765 otherwise bn_mod_exp was called. In the case of hardware keys for example
7766 no private key components need be present and it might store extra data
7767 in the RSA structure, which cannot be accessed from bn_mod_exp.
7768 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
7769 private key operations.
7772 *) Added support for SPARC Linux.
7775 *) pem_password_cb function type incompatibly changed from
7776 typedef int pem_password_cb(char *buf, int size, int rwflag);
7778 ....(char *buf, int size, int rwflag, void *userdata);
7779 so that applications can pass data to their callbacks:
7780 The PEM[_ASN1]_{read,write}... functions and macros now take an
7781 additional void * argument, which is just handed through whenever
7782 the password callback is called.
7783 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
7785 New function SSL_CTX_set_default_passwd_cb_userdata.
7787 Compatibility note: As many C implementations push function arguments
7788 onto the stack in reverse order, the new library version is likely to
7789 interoperate with programs that have been compiled with the old
7790 pem_password_cb definition (PEM_whatever takes some data that
7791 happens to be on the stack as its last argument, and the callback
7792 just ignores this garbage); but there is no guarantee whatsoever that
7795 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
7796 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
7797 problems not only on Windows, but also on some Unix platforms.
7798 To avoid problematic command lines, these definitions are now in an
7799 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
7800 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
7803 *) MIPS III/IV assembler module is reimplemented.
7806 *) More DES library cleanups: remove references to srand/rand and
7807 delete an unused file.
7810 *) Add support for the the free Netwide assembler (NASM) under Win32,
7811 since not many people have MASM (ml) and it can be hard to obtain.
7812 This is currently experimental but it seems to work OK and pass all
7813 the tests. Check out INSTALL.W32 for info.
7816 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
7817 without temporary keys kept an extra copy of the server key,
7818 and connections with temporary keys did not free everything in case
7822 *) New function RSA_check_key and new openssl rsa option -check
7823 for verifying the consistency of RSA keys.
7824 [Ulf Moeller, Bodo Moeller]
7826 *) Various changes to make Win32 compile work:
7827 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
7828 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
7829 comparison" warnings.
7830 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
7833 *) Add a debugging option to PKCS#5 v2 key generation function: when
7834 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
7835 derived keys are printed to stderr.
7838 *) Copy the flags in ASN1_STRING_dup().
7839 [Roman E. Pavlov <pre@mo.msk.ru>]
7841 *) The x509 application mishandled signing requests containing DSA
7842 keys when the signing key was also DSA and the parameters didn't match.
7844 It was supposed to omit the parameters when they matched the signing key:
7845 the verifying software was then supposed to automatically use the CA's
7846 parameters if they were absent from the end user certificate.
7848 Omitting parameters is no longer recommended. The test was also
7849 the wrong way round! This was probably due to unusual behaviour in
7850 EVP_cmp_parameters() which returns 1 if the parameters match.
7851 This meant that parameters were omitted when they *didn't* match and
7852 the certificate was useless. Certificates signed with 'ca' didn't have
7854 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
7856 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
7857 The interface is as follows:
7858 Applications can use
7859 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
7860 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
7861 "off" is now the default.
7862 The library internally uses
7863 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
7864 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
7865 to disable memory-checking temporarily.
7867 Some inconsistent states that previously were possible (and were
7868 even the default) are now avoided.
7870 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
7871 with each memory chunk allocated; this is occasionally more helpful
7872 than just having a counter.
7874 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
7876 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
7880 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
7881 which largely parallels "options", but is for changing API behaviour,
7882 whereas "options" are about protocol behaviour.
7883 Initial "mode" flags are:
7885 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
7886 a single record has been written.
7887 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
7888 retries use the same buffer location.
7889 (But all of the contents must be
7893 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
7896 *) Fix problems with no-hmac etc.
7897 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
7899 *) New functions RSA_get_default_method(), RSA_set_method() and
7900 RSA_get_method(). These allows replacement of RSA_METHODs without having
7901 to mess around with the internals of an RSA structure.
7904 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
7905 Also really enable memory leak checks in openssl.c and in some
7907 [Chad C. Mulligan, Bodo Moeller]
7909 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
7910 up the length of negative integers. This has now been simplified to just
7911 store the length when it is first determined and use it later, rather
7912 than trying to keep track of where data is copied and updating it to
7914 [Steve Henson, reported by Brien Wheeler
7915 <bwheeler@authentica-security.com>]
7917 *) Add a new function PKCS7_signatureVerify. This allows the verification
7918 of a PKCS#7 signature but with the signing certificate passed to the
7919 function itself. This contrasts with PKCS7_dataVerify which assumes the
7920 certificate is present in the PKCS#7 structure. This isn't always the
7921 case: certificates can be omitted from a PKCS#7 structure and be
7922 distributed by "out of band" means (such as a certificate database).
7925 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
7926 function prototypes in pem.h, also change util/mkdef.pl to add the
7927 necessary function names.
7930 *) mk1mf.pl (used by Windows builds) did not properly read the
7931 options set by Configure in the top level Makefile, and Configure
7932 was not even able to write more than one option correctly.
7933 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
7936 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
7937 file to be loaded from a BIO or FILE pointer. The BIO version will
7938 for example allow memory BIOs to contain config info.
7941 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
7942 Whoever hopes to achieve shared-library compatibility across versions
7943 must use this, not the compile-time macro.
7944 (Exercise 0.9.4: Which is the minimum library version required by
7946 Note: All this applies only to multi-threaded programs, others don't
7950 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
7951 through a BIO pair triggered the default case, i.e.
7952 SSLerr(...,SSL_R_UNKNOWN_STATE).
7955 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
7956 can use the SSL library even if none of the specific BIOs is
7960 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
7961 for the encoded length.
7962 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
7964 *) Add initial documentation of the X509V3 functions.
7967 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
7968 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
7969 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
7970 secure PKCS#8 private key format with a high iteration count.
7973 *) Fix determination of Perl interpreter: A perl or perl5
7974 _directory_ in $PATH was also accepted as the interpreter.
7975 [Ralf S. Engelschall]
7977 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
7978 wrong with it but it was very old and did things like calling
7979 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
7983 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
7984 to use the new extension code.
7987 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
7988 with macros. This should make it easier to change their form, add extra
7989 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
7993 *) Add to configuration table a new entry that can specify an alternative
7994 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
7995 according to Mark Crispin <MRC@Panda.COM>.
7999 *) DES CBC did not update the IV. Weird.
8002 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
8003 Changing the behaviour of the former might break existing programs --
8004 where IV updating is needed, des_ncbc_encrypt can be used.
8007 *) When bntest is run from "make test" it drives bc to check its
8008 calculations, as well as internally checking them. If an internal check
8009 fails, it needs to cause bc to give a non-zero result or make test carries
8010 on without noticing the failure. Fixed.
8013 *) DES library cleanups.
8016 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
8017 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
8018 ciphers. NOTE: although the key derivation function has been verified
8019 against some published test vectors it has not been extensively tested
8020 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
8024 *) Instead of "mkdir -p", which is not fully portable, use new
8025 Perl script "util/mkdir-p.pl".
8028 *) Rewrite the way password based encryption (PBE) is handled. It used to
8029 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
8030 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
8031 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
8032 the 'parameter' field of the AlgorithmIdentifier is passed to the
8033 underlying key generation function so it must do its own ASN1 parsing.
8034 This has also changed the EVP_PBE_CipherInit() function which now has a
8035 'parameter' argument instead of literal salt and iteration count values
8036 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
8039 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
8040 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
8041 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
8042 KEY" because this clashed with PKCS#8 unencrypted string. Since this
8043 value was just used as a "magic string" and not used directly its
8044 value doesn't matter.
8047 *) Introduce some semblance of const correctness to BN. Shame C doesn't
8051 *) "linux-sparc64" configuration (ultrapenguin).
8052 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
8053 "linux-sparc" configuration.
8054 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
8056 *) config now generates no-xxx options for missing ciphers.
8059 *) Support the EBCDIC character set (work in progress).
8060 File ebcdic.c not yet included because it has a different license.
8061 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8063 *) Support BS2000/OSD-POSIX.
8064 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8066 *) Make callbacks for key generation use void * instead of char *.
8069 *) Make S/MIME samples compile (not yet tested).
8072 *) Additional typesafe stacks.
8075 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
8079 Changes between 0.9.3 and 0.9.3a [29 May 1999]
8081 *) New configuration variant "sco5-gcc".
8083 *) Updated some demos.
8084 [Sean O Riordain, Wade Scholine]
8086 *) Add missing BIO_free at exit of pkcs12 application.
8089 *) Fix memory leak in conf.c.
8092 *) Updates for Win32 to assembler version of MD5.
8095 *) Set #! path to perl in apps/der_chop to where we found it
8096 instead of using a fixed path.
8099 *) SHA library changes for irix64-mips4-cc.
8102 *) Improvements for VMS support.
8106 Changes between 0.9.2b and 0.9.3 [24 May 1999]
8108 *) Bignum library bug fix. IRIX 6 passes "make test" now!
8109 This also avoids the problems with SC4.2 and unpatched SC5.
8110 [Andy Polyakov <appro@fy.chalmers.se>]
8112 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
8113 These are required because of the typesafe stack would otherwise break
8114 existing code. If old code used a structure member which used to be STACK
8115 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
8116 sk_num or sk_value it would produce an error because the num, data members
8117 are not present in STACK_OF. Now it just produces a warning. sk_set
8118 replaces the old method of assigning a value to sk_value
8119 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
8120 that does this will no longer work (and should use sk_set instead) but
8121 this could be regarded as a "questionable" behaviour anyway.
8124 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
8125 correctly handle encrypted S/MIME data.
8128 *) Change type of various DES function arguments from des_cblock
8129 (which means, in function argument declarations, pointer to char)
8130 to des_cblock * (meaning pointer to array with 8 char elements),
8131 which allows the compiler to do more typechecking; it was like
8132 that back in SSLeay, but with lots of ugly casts.
8134 Introduce new type const_des_cblock.
8137 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
8138 problems: find RecipientInfo structure that matches recipient certificate
8139 and initialise the ASN1 structures properly based on passed cipher.
8142 *) Belatedly make the BN tests actually check the results.
8145 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
8146 to and from BNs: it was completely broken. New compilation option
8147 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
8148 key elements as negative integers.
8151 *) Reorganize and speed up MD5.
8152 [Andy Polyakov <appro@fy.chalmers.se>]
8155 [Richard Levitte <richard@levitte.org>]
8157 *) New option -out to asn1parse to allow the parsed structure to be
8158 output to a file. This is most useful when combined with the -strparse
8159 option to examine the output of things like OCTET STRINGS.
8162 *) Make SSL library a little more fool-proof by not requiring any longer
8163 that SSL_set_{accept,connect}_state be called before
8164 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
8165 in many applications because usually everything *appeared* to work as
8166 intended anyway -- now it really works as intended).
8169 *) Move openssl.cnf out of lib/.
8172 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
8173 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
8174 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
8175 [Ralf S. Engelschall]
8177 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
8178 handle PKCS#7 enveloped data properly.
8179 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
8181 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
8182 copying pointers. The cert_st handling is changed by this in
8183 various ways (and thus what used to be known as ctx->default_cert
8184 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
8185 any longer when s->cert does not give us what we need).
8186 ssl_cert_instantiate becomes obsolete by this change.
8187 As soon as we've got the new code right (possibly it already is?),
8188 we have solved a couple of bugs of the earlier code where s->cert
8189 was used as if it could not have been shared with other SSL structures.
8191 Note that using the SSL API in certain dirty ways now will result
8192 in different behaviour than observed with earlier library versions:
8193 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
8194 does not influence s as it used to.
8196 In order to clean up things more thoroughly, inside SSL_SESSION
8197 we don't use CERT any longer, but a new structure SESS_CERT
8198 that holds per-session data (if available); currently, this is
8199 the peer's certificate chain and, for clients, the server's certificate
8200 and temporary key. CERT holds only those values that can have
8201 meaningful defaults in an SSL_CTX.
8204 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
8205 from the internal representation. Various PKCS#7 fixes: remove some
8206 evil casts and set the enc_dig_alg field properly based on the signing
8210 *) Allow PKCS#12 password to be set from the command line or the
8211 environment. Let 'ca' get its config file name from the environment
8212 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
8216 *) Allow certificate policies extension to use an IA5STRING for the
8217 organization field. This is contrary to the PKIX definition but
8218 VeriSign uses it and IE5 only recognises this form. Document 'x509'
8222 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
8223 without disallowing inline assembler and the like for non-pedantic builds.
8226 *) Support Borland C++ builder.
8227 [Janez Jere <jj@void.si>, modified by Ulf Möller]
8232 *) SHA-1 cleanups and performance enhancements.
8233 [Andy Polyakov <appro@fy.chalmers.se>]
8235 *) Sparc v8plus assembler for the bignum library.
8236 [Andy Polyakov <appro@fy.chalmers.se>]
8238 *) Accept any -xxx and +xxx compiler options in Configure.
8241 *) Update HPUX configuration.
8244 *) Add missing sk_<type>_unshift() function to safestack.h
8245 [Ralf S. Engelschall]
8247 *) New function SSL_CTX_use_certificate_chain_file that sets the
8248 "extra_cert"s in addition to the certificate. (This makes sense
8249 only for "PEM" format files, as chains as a whole are not
8253 *) Support verify_depth from the SSL API.
8254 x509_vfy.c had what can be considered an off-by-one-error:
8255 Its depth (which was not part of the external interface)
8256 was actually counting the number of certificates in a chain;
8257 now it really counts the depth.
8260 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
8261 instead of X509err, which often resulted in confusing error
8262 messages since the error codes are not globally unique
8263 (e.g. an alleged error in ssl3_accept when a certificate
8264 didn't match the private key).
8266 *) New function SSL_CTX_set_session_id_context that allows to set a default
8267 value (so that you don't need SSL_set_session_id_context for each
8268 connection using the SSL_CTX).
8271 *) OAEP decoding bug fix.
8274 *) Support INSTALL_PREFIX for package builders, as proposed by
8278 *) New Configure options "threads" and "no-threads". For systems
8279 where the proper compiler options are known (currently Solaris
8280 and Linux), "threads" is the default.
8283 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
8286 *) Install various scripts to $(OPENSSLDIR)/misc, not to
8287 $(INSTALLTOP)/bin -- they shouldn't clutter directories
8288 such as /usr/local/bin.
8291 *) "make linux-shared" to build shared libraries.
8292 [Niels Poppe <niels@netbox.org>]
8294 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
8297 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
8298 extension adding in x509 utility.
8301 *) Remove NOPROTO sections and error code comments.
8304 *) Partial rewrite of the DEF file generator to now parse the ANSI
8308 *) New Configure options --prefix=DIR and --openssldir=DIR.
8311 *) Complete rewrite of the error code script(s). It is all now handled
8312 by one script at the top level which handles error code gathering,
8313 header rewriting and C source file generation. It should be much better
8314 than the old method: it now uses a modified version of Ulf's parser to
8315 read the ANSI prototypes in all header files (thus the old K&R definitions
8316 aren't needed for error creation any more) and do a better job of
8317 translating function codes into names. The old 'ASN1 error code imbedded
8318 in a comment' is no longer necessary and it doesn't use .err files which
8319 have now been deleted. Also the error code call doesn't have to appear all
8320 on one line (which resulted in some large lines...).
8323 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
8326 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
8327 0 (which usually indicates a closed connection), but continue reading.
8330 *) Fix some race conditions.
8333 *) Add support for CRL distribution points extension. Add Certificate
8334 Policies and CRL distribution points documentation.
8337 *) Move the autogenerated header file parts to crypto/opensslconf.h.
8340 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
8341 8 of keying material. Merlin has also confirmed interop with this fix
8342 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
8343 [Merlin Hughes <merlin@baltimore.ie>]
8345 *) Fix lots of warnings.
8346 [Richard Levitte <levitte@stacken.kth.se>]
8348 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
8349 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
8350 [Richard Levitte <levitte@stacken.kth.se>]
8352 *) Fix problems with sizeof(long) == 8.
8353 [Andy Polyakov <appro@fy.chalmers.se>]
8355 *) Change functions to ANSI C.
8358 *) Fix typos in error codes.
8359 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
8361 *) Remove defunct assembler files from Configure.
8364 *) SPARC v8 assembler BIGNUM implementation.
8365 [Andy Polyakov <appro@fy.chalmers.se>]
8367 *) Support for Certificate Policies extension: both print and set.
8368 Various additions to support the r2i method this uses.
8371 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
8372 return a const string when you are expecting an allocated buffer.
8375 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
8376 types DirectoryString and DisplayText.
8379 *) Add code to allow r2i extensions to access the configuration database,
8380 add an LHASH database driver and add several ctx helper functions.
8383 *) Fix an evil bug in bn_expand2() which caused various BN functions to
8384 fail when they extended the size of a BIGNUM.
8387 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
8388 support typesafe stack.
8391 *) Fix typo in SSL_[gs]et_options().
8392 [Nils Frostberg <nils@medcom.se>]
8394 *) Delete various functions and files that belonged to the (now obsolete)
8395 old X509V3 handling code.
8398 *) New Configure option "rsaref".
8401 *) Don't auto-generate pem.h.
8404 *) Introduce type-safe ASN.1 SETs.
8407 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
8408 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
8410 *) Introduce type-safe STACKs. This will almost certainly break lots of code
8411 that links with OpenSSL (well at least cause lots of warnings), but fear
8412 not: the conversion is trivial, and it eliminates loads of evil casts. A
8413 few STACKed things have been converted already. Feel free to convert more.
8414 In the fullness of time, I'll do away with the STACK type altogether.
8417 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
8418 specified in <certfile> by updating the entry in the index.txt file.
8419 This way one no longer has to edit the index.txt file manually for
8420 revoking a certificate. The -revoke option does the gory details now.
8421 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
8423 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
8424 `-text' option at all and this way the `-noout -text' combination was
8425 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
8426 [Ralf S. Engelschall]
8428 *) Make sure a corresponding plain text error message exists for the
8429 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
8430 verify callback function determined that a certificate was revoked.
8431 [Ralf S. Engelschall]
8433 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
8434 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
8435 all available cipers including rc5, which was forgotten until now.
8436 In order to let the testing shell script know which algorithms
8437 are available, a new (up to now undocumented) command
8438 "openssl list-cipher-commands" is used.
8441 *) Bugfix: s_client occasionally would sleep in select() when
8442 it should have checked SSL_pending() first.
8445 *) New functions DSA_do_sign and DSA_do_verify to provide access to
8446 the raw DSA values prior to ASN.1 encoding.
8449 *) Tweaks to Configure
8450 [Niels Poppe <niels@netbox.org>]
8452 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
8456 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
8459 *) New config option to avoid instructions that are illegal on the 80386.
8460 The default code is faster, but requires at least a 486.
8463 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
8464 SSL2_SERVER_VERSION (not used at all) macros, which are now the
8465 same as SSL2_VERSION anyway.
8468 *) New "-showcerts" option for s_client.
8471 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
8472 application. Various cleanups and fixes.
8475 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
8476 modify error routines to work internally. Add error codes and PBE init
8477 to library startup routines.
8480 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
8481 packing functions to asn1 and evp. Changed function names and error
8482 codes along the way.
8485 *) PKCS12 integration: and so it begins... First of several patches to
8486 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
8487 objects to objects.h
8490 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
8491 and display support for Thawte strong extranet extension.
8494 *) Add LinuxPPC support.
8495 [Jeff Dubrule <igor@pobox.org>]
8497 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
8498 bn_div_words in alpha.s.
8499 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
8501 *) Make sure the RSA OAEP test is skipped under -DRSAref because
8502 OAEP isn't supported when OpenSSL is built with RSAref.
8503 [Ulf Moeller <ulf@fitug.de>]
8505 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
8506 so they no longer are missing under -DNOPROTO.
8507 [Soren S. Jorvang <soren@t.dk>]
8510 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
8512 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
8513 doesn't work when the session is reused. Coming soon!
8516 *) Fix a security hole, that allows sessions to be reused in the wrong
8517 context thus bypassing client cert protection! All software that uses
8518 client certs and session caches in multiple contexts NEEDS PATCHING to
8519 allow session reuse! A fuller solution is in the works.
8520 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
8522 *) Some more source tree cleanups (removed obsolete files
8523 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
8524 permission on "config" script to be executable) and a fix for the INSTALL
8526 [Ulf Moeller <ulf@fitug.de>]
8528 *) Remove some legacy and erroneous uses of malloc, free instead of
8530 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
8532 *) Make rsa_oaep_test return non-zero on error.
8533 [Ulf Moeller <ulf@fitug.de>]
8535 *) Add support for native Solaris shared libraries. Configure
8536 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
8537 if someone would make that last step automatic.
8538 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
8540 *) ctx_size was not built with the right compiler during "make links". Fixed.
8543 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
8544 except NULL ciphers". This means the default cipher list will no longer
8545 enable NULL ciphers. They need to be specifically enabled e.g. with
8546 the string "DEFAULT:eNULL".
8549 *) Fix to RSA private encryption routines: if p < q then it would
8550 occasionally produce an invalid result. This will only happen with
8551 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
8554 *) Be less restrictive and allow also `perl util/perlpath.pl
8555 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
8556 because this way one can also use an interpreter named `perl5' (which is
8557 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
8558 installed as `perl').
8559 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8561 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
8562 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8564 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
8565 advapi32.lib to Win32 build and change the pem test comparision
8566 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
8567 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
8568 and crypto/des/ede_cbcm_enc.c.
8571 *) DES quad checksum was broken on big-endian architectures. Fixed.
8574 *) Comment out two functions in bio.h that aren't implemented. Fix up the
8575 Win32 test batch file so it (might) work again. The Win32 test batch file
8576 is horrible: I feel ill....
8579 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
8580 in e_os.h. Audit of header files to check ANSI and non ANSI
8581 sections: 10 functions were absent from non ANSI section and not exported
8582 from Windows DLLs. Fixed up libeay.num for new functions.
8585 *) Make `openssl version' output lines consistent.
8586 [Ralf S. Engelschall]
8588 *) Fix Win32 symbol export lists for BIO functions: Added
8589 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
8590 to ms/libeay{16,32}.def.
8591 [Ralf S. Engelschall]
8593 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
8594 fine under Unix and passes some trivial tests I've now added. But the
8595 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
8596 added to make sure no one expects that this stuff really works in the
8597 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
8598 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
8600 [Ralf S. Engelschall]
8602 *) Fix the generation of two part addresses in perl.
8603 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
8605 *) Add config entry for Linux on MIPS.
8606 [John Tobey <jtobey@channel1.com>]
8608 *) Make links whenever Configure is run, unless we are on Windoze.
8611 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
8612 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
8616 *) Add a useful kludge to allow package maintainers to specify compiler and
8617 other platforms details on the command line without having to patch the
8618 Configure script everytime: One now can use ``perl Configure
8619 <id>:<details>'', i.e. platform ids are allowed to have details appended
8620 to them (seperated by colons). This is treated as there would be a static
8621 pre-configured entry in Configure's %table under key <id> with value
8622 <details> and ``perl Configure <id>'' is called. So, when you want to
8623 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
8624 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
8625 now, which overrides the FreeBSD-elf entry on-the-fly.
8626 [Ralf S. Engelschall]
8628 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
8631 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
8632 on the `perl Configure ...' command line. This way one can compile
8633 OpenSSL libraries with Position Independent Code (PIC) which is needed
8634 for linking it into DSOs.
8635 [Ralf S. Engelschall]
8637 *) Remarkably, export ciphers were totally broken and no-one had noticed!
8641 *) Cleaned up the LICENSE document: The official contact for any license
8642 questions now is the OpenSSL core team under openssl-core@openssl.org.
8643 And add a paragraph about the dual-license situation to make sure people
8644 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
8645 to the OpenSSL toolkit.
8646 [Ralf S. Engelschall]
8648 *) General source tree makefile cleanups: Made `making xxx in yyy...'
8649 display consistent in the source tree and replaced `/bin/rm' by `rm'.
8650 Additonally cleaned up the `make links' target: Remove unnecessary
8651 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
8652 to speed processing and no longer clutter the display with confusing
8653 stuff. Instead only the actually done links are displayed.
8654 [Ralf S. Engelschall]
8656 *) Permit null encryption ciphersuites, used for authentication only. It used
8657 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
8658 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
8662 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
8663 signed attributes when verifying signatures (this would break them),
8664 the detached data encoding was wrong and public keys obtained using
8665 X509_get_pubkey() weren't freed.
8668 *) Add text documentation for the BUFFER functions. Also added a work around
8669 to a Win95 console bug. This was triggered by the password read stuff: the
8670 last character typed gets carried over to the next fread(). If you were
8671 generating a new cert request using 'req' for example then the last
8672 character of the passphrase would be CR which would then enter the first
8676 *) Added the new `Includes OpenSSL Cryptography Software' button as
8677 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
8678 button and can be used by applications based on OpenSSL to show the
8679 relationship to the OpenSSL project.
8680 [Ralf S. Engelschall]
8682 *) Remove confusing variables in function signatures in files
8683 ssl/ssl_lib.c and ssl/ssl.h.
8684 [Lennart Bong <lob@kulthea.stacken.kth.se>]
8686 *) Don't install bss_file.c under PREFIX/include/
8687 [Lennart Bong <lob@kulthea.stacken.kth.se>]
8689 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
8690 functions that return function pointers and has support for NT specific
8691 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
8692 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
8693 unsigned to signed types: this was killing the Win32 compile.
8696 *) Add new certificate file to stack functions,
8697 SSL_add_dir_cert_subjects_to_stack() and
8698 SSL_add_file_cert_subjects_to_stack(). These largely supplant
8699 SSL_load_client_CA_file(), and can be used to add multiple certs easily
8700 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
8701 This means that Apache-SSL and similar packages don't have to mess around
8702 to add as many CAs as they want to the preferred list.
8705 *) Experiment with doxygen documentation. Currently only partially applied to
8707 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
8708 openssl.doxy as the configuration file.
8711 *) Get rid of remaining C++-style comments which strict C compilers hate.
8712 [Ralf S. Engelschall, pointed out by Carlos Amengual]
8714 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
8715 compiled in by default: it has problems with large keys.
8718 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
8719 DH private keys and/or callback functions which directly correspond to
8720 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
8721 is needed for applications which have to configure certificates on a
8722 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
8724 For the RSA certificate situation is makes no difference, but
8725 for the DSA certificate situation this fixes the "no shared cipher"
8726 problem where the OpenSSL cipher selection procedure failed because the
8727 temporary keys were not overtaken from the context and the API provided
8728 no way to reconfigure them.
8729 The new functions now let applications reconfigure the stuff and they
8730 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
8731 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
8732 non-public-API function ssl_cert_instantiate() is used as a helper
8733 function and also to reduce code redundancy inside ssl_rsa.c.
8734 [Ralf S. Engelschall]
8736 *) Move s_server -dcert and -dkey options out of the undocumented feature
8737 area because they are useful for the DSA situation and should be
8738 recognized by the users.
8739 [Ralf S. Engelschall]
8741 *) Fix the cipher decision scheme for export ciphers: the export bits are
8742 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
8743 SSL_EXP_MASK. So, the original variable has to be used instead of the
8744 already masked variable.
8745 [Richard Levitte <levitte@stacken.kth.se>]
8747 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
8748 [Richard Levitte <levitte@stacken.kth.se>]
8750 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
8751 from `int' to `unsigned int' because it's a length and initialized by
8752 EVP_DigestFinal() which expects an `unsigned int *'.
8753 [Richard Levitte <levitte@stacken.kth.se>]
8755 *) Don't hard-code path to Perl interpreter on shebang line of Configure
8756 script. Instead use the usual Shell->Perl transition trick.
8757 [Ralf S. Engelschall]
8759 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
8760 (in addition to RSA certificates) to match the behaviour of `openssl dsa
8761 -noout -modulus' as it's already the case for `openssl rsa -noout
8762 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
8763 currently the public key is printed (a decision which was already done by
8764 `openssl dsa -modulus' in the past) which serves a similar purpose.
8765 Additionally the NO_RSA no longer completely removes the whole -modulus
8766 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
8768 [Ralf S. Engelschall]
8770 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
8771 BIO. See the source (crypto/evp/bio_ok.c) for more info.
8772 [Arne Ansper <arne@ats.cyber.ee>]
8774 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
8775 to be added. Now both 'req' and 'ca' can use new objects defined in the
8779 *) Add cool BIO that does syslog (or event log on NT).
8780 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
8782 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
8783 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
8784 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
8785 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
8788 *) Add preliminary config info for new extension code.
8791 *) Make RSA_NO_PADDING really use no padding.
8792 [Ulf Moeller <ulf@fitug.de>]
8794 *) Generate errors when private/public key check is done.
8797 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
8798 for some CRL extensions and new objects added.
8801 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
8802 key usage extension and fuller support for authority key id.
8805 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
8806 padding method for RSA, which is recommended for new applications in PKCS
8807 #1 v2.0 (RFC 2437, October 1998).
8808 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
8809 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
8810 against Bleichbacher's attack on RSA.
8811 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
8814 *) Updates to the new SSL compression code
8815 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8817 *) Fix so that the version number in the master secret, when passed
8818 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
8819 (because the server will not accept higher), that the version number
8820 is 0x03,0x01, not 0x03,0x00
8821 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8823 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
8824 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
8825 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
8828 *) Support for RAW extensions where an arbitrary extension can be
8829 created by including its DER encoding. See apps/openssl.cnf for
8833 *) Make sure latest Perl versions don't interpret some generated C array
8834 code as Perl array code in the crypto/err/err_genc.pl script.
8835 [Lars Weber <3weber@informatik.uni-hamburg.de>]
8837 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
8838 not many people have the assembler. Various Win32 compilation fixes and
8839 update to the INSTALL.W32 file with (hopefully) more accurate Win32
8843 *) Modify configure script 'Configure' to automatically create crypto/date.h
8844 file under Win32 and also build pem.h from pem.org. New script
8845 util/mkfiles.pl to create the MINFO file on environments that can't do a
8846 'make files': perl util/mkfiles.pl >MINFO should work.
8849 *) Major rework of DES function declarations, in the pursuit of correctness
8850 and purity. As a result, many evil casts evaporated, and some weirdness,
8851 too. You may find this causes warnings in your code. Zapping your evil
8852 casts will probably fix them. Mostly.
8855 *) Fix for a typo in asn1.h. Bug fix to object creation script
8856 obj_dat.pl. It considered a zero in an object definition to mean
8857 "end of object": none of the objects in objects.h have any zeros
8858 so it wasn't spotted.
8859 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
8861 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
8862 Masking (CBCM). In the absence of test vectors, the best I have been able
8863 to do is check that the decrypt undoes the encrypt, so far. Send me test
8864 vectors if you have them.
8867 *) Correct calculation of key length for export ciphers (too much space was
8868 allocated for null ciphers). This has not been tested!
8871 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
8872 message is now correct (it understands "crypto" and "ssl" on its
8873 command line). There is also now an "update" option. This will update
8874 the util/ssleay.num and util/libeay.num files with any new functions.
8876 perl util/mkdef.pl crypto ssl update
8877 it will update them.
8880 *) Overhauled the Perl interface (perl/*):
8881 - ported BN stuff to OpenSSL's different BN library
8882 - made the perl/ source tree CVS-aware
8883 - renamed the package from SSLeay to OpenSSL (the files still contain
8884 their history because I've copied them in the repository)
8885 - removed obsolete files (the test scripts will be replaced
8886 by better Test::Harness variants in the future)
8887 [Ralf S. Engelschall]
8889 *) First cut for a very conservative source tree cleanup:
8890 1. merge various obsolete readme texts into doc/ssleay.txt
8891 where we collect the old documents and readme texts.
8892 2. remove the first part of files where I'm already sure that we no
8893 longer need them because of three reasons: either they are just temporary
8894 files which were left by Eric or they are preserved original files where
8895 I've verified that the diff is also available in the CVS via "cvs diff
8896 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
8897 the crypto/md/ stuff).
8898 [Ralf S. Engelschall]
8900 *) More extension code. Incomplete support for subject and issuer alt
8901 name, issuer and authority key id. Change the i2v function parameters
8902 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
8903 what that's for :-) Fix to ASN1 macro which messed up
8904 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
8907 *) Preliminary support for ENUMERATED type. This is largely copied from the
8911 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
8912 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8914 *) Make sure `make rehash' target really finds the `openssl' program.
8915 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8917 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
8918 like to hear about it if this slows down other processors.
8921 *) Add CygWin32 platform information to Configure script.
8922 [Alan Batie <batie@aahz.jf.intel.com>]
8924 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
8925 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
8927 *) New program nseq to manipulate netscape certificate sequences
8930 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
8934 *) Fixes to BN code. Previously the default was to define BN_RECURSION
8935 but the BN code had some problems that would cause failures when
8936 doing certificate verification and some other functions.
8937 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8939 *) Add ASN1 and PEM code to support netscape certificate sequences.
8942 *) Add ASN1 and PEM code to support netscape certificate sequences.
8945 *) Add several PKIX and private extended key usage OIDs.
8948 *) Modify the 'ca' program to handle the new extension code. Modify
8949 openssl.cnf for new extension format, add comments.
8952 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
8953 and add a sample to openssl.cnf so req -x509 now adds appropriate
8957 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
8958 error code, add initial support to X509_print() and x509 application.
8961 *) Takes a deep breath and start addding X509 V3 extension support code. Add
8962 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
8963 stuff is currently isolated and isn't even compiled yet.
8966 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
8967 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
8968 Removed the versions check from X509 routines when loading extensions:
8969 this allows certain broken certificates that don't set the version
8970 properly to be processed.
8973 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
8974 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
8975 can still be regenerated with "make depend".
8978 *) Spelling mistake in C version of CAST-128.
8979 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
8981 *) Changes to the error generation code. The perl script err-code.pl
8982 now reads in the old error codes and retains the old numbers, only
8983 adding new ones if necessary. It also only changes the .err files if new
8984 codes are added. The makefiles have been modified to only insert errors
8985 when needed (to avoid needlessly modifying header files). This is done
8986 by only inserting errors if the .err file is newer than the auto generated
8987 C file. To rebuild all the error codes from scratch (the old behaviour)
8988 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
8989 or delete all the .err files.
8992 *) CAST-128 was incorrectly implemented for short keys. The C version has
8993 been fixed, but is untested. The assembler versions are also fixed, but
8994 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
8995 to regenerate it if needed.
8996 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
8997 Hagino <itojun@kame.net>]
8999 *) File was opened incorrectly in randfile.c.
9000 [Ulf Möller <ulf@fitug.de>]
9002 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
9003 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
9004 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
9005 al: it's just almost always a UTCTime. Note this patch adds new error
9006 codes so do a "make errors" if there are problems.
9009 *) Correct Linux 1 recognition in config.
9010 [Ulf Möller <ulf@fitug.de>]
9012 *) Remove pointless MD5 hash when using DSA keys in ca.
9013 [Anonymous <nobody@replay.com>]
9015 *) Generate an error if given an empty string as a cert directory. Also
9016 generate an error if handed NULL (previously returned 0 to indicate an
9017 error, but didn't set one).
9018 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
9020 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
9023 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
9024 parameters. This was causing a warning which killed off the Win32 compile.
9027 *) Remove C++ style comments from crypto/bn/bn_local.h.
9028 [Neil Costigan <neil.costigan@celocom.com>]
9030 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
9031 based on a text string, looking up short and long names and finally
9032 "dot" format. The "dot" format stuff didn't work. Added new function
9033 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
9034 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
9035 OID is not part of the table.
9038 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
9039 X509_LOOKUP_by_alias().
9042 *) Sort openssl functions by name.
9045 *) Get the gendsa program working (hopefully) and add it to app list. Remove
9046 encryption from sample DSA keys (in case anyone is interested the password
9050 *) Make _all_ *_free functions accept a NULL pointer.
9051 [Frans Heymans <fheymans@isaserver.be>]
9053 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
9055 [Anonymous <nobody@replay.com>]
9057 *) s_server should send the CAfile as acceptable CAs, not its own cert.
9058 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9060 *) Don't blow it for numeric -newkey arguments to apps/req.
9061 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9063 *) Temp key "for export" tests were wrong in s3_srvr.c.
9064 [Anonymous <nobody@replay.com>]
9066 *) Add prototype for temp key callback functions
9067 SSL_CTX_set_tmp_{rsa,dh}_callback().
9070 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
9071 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
9074 *) X509_name_add_entry() freed the wrong thing after an error.
9075 [Arne Ansper <arne@ats.cyber.ee>]
9077 *) rsa_eay.c would attempt to free a NULL context.
9078 [Arne Ansper <arne@ats.cyber.ee>]
9080 *) BIO_s_socket() had a broken should_retry() on Windoze.
9081 [Arne Ansper <arne@ats.cyber.ee>]
9083 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
9084 [Arne Ansper <arne@ats.cyber.ee>]
9086 *) Make sure the already existing X509_STORE->depth variable is initialized
9087 in X509_STORE_new(), but document the fact that this variable is still
9088 unused in the certificate verification process.
9089 [Ralf S. Engelschall]
9091 *) Fix the various library and apps files to free up pkeys obtained from
9092 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
9095 *) Fix reference counting in X509_PUBKEY_get(). This makes
9096 demos/maurice/example2.c work, amongst others, probably.
9097 [Steve Henson and Ben Laurie]
9099 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
9100 `openssl' and second, the shortcut symlinks for the `openssl <command>'
9101 are no longer created. This way we have a single and consistent command
9102 line interface `openssl <command>', similar to `cvs <command>'.
9103 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
9105 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
9106 BIT STRING wrapper always have zero unused bits.
9109 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
9112 *) Make the top-level INSTALL documentation easier to understand.
9115 *) Makefiles updated to exit if an error occurs in a sub-directory
9116 make (including if user presses ^C) [Paul Sutton]
9118 *) Make Montgomery context stuff explicit in RSA data structure.
9121 *) Fix build order of pem and err to allow for generated pem.h.
9124 *) Fix renumbering bug in X509_NAME_delete_entry().
9127 *) Enhanced the err-ins.pl script so it makes the error library number
9128 global and can add a library name. This is needed for external ASN1 and
9129 other error libraries.
9132 *) Fixed sk_insert which never worked properly.
9135 *) Fix ASN1 macros so they can handle indefinite length construted
9136 EXPLICIT tags. Some non standard certificates use these: they can now
9140 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
9141 into a single doc/ssleay.txt bundle. This way the information is still
9142 preserved but no longer messes up this directory. Now it's new room for
9143 the new set of documenation files.
9144 [Ralf S. Engelschall]
9146 *) SETs were incorrectly DER encoded. This was a major pain, because they
9147 shared code with SEQUENCEs, which aren't coded the same. This means that
9148 almost everything to do with SETs or SEQUENCEs has either changed name or
9149 number of arguments.
9150 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
9152 *) Fix test data to work with the above.
9155 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
9156 was already fixed by Eric for 0.9.1 it seems.
9157 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
9159 *) Autodetect FreeBSD3.
9162 *) Fix various bugs in Configure. This affects the following platforms:
9166 unixware-2.0-pentium
9170 *) Eliminate generated files from CVS. Reorder tests to regenerate files
9171 before they are needed.
9174 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
9178 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
9180 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
9181 changed SSLeay to OpenSSL in version strings.
9182 [Ralf S. Engelschall]
9184 *) Some fixups to the top-level documents.
9187 *) Fixed the nasty bug where rsaref.h was not found under compile-time
9188 because the symlink to include/ was missing.
9189 [Ralf S. Engelschall]
9191 *) Incorporated the popular no-RSA/DSA-only patches
9192 which allow to compile a RSA-free SSLeay.
9193 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
9195 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
9196 when "ssleay" is still not found.
9197 [Ralf S. Engelschall]
9199 *) Added more platforms to Configure: Cray T3E, HPUX 11,
9200 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
9202 *) Updated the README file.
9203 [Ralf S. Engelschall]
9205 *) Added various .cvsignore files in the CVS repository subdirs
9206 to make a "cvs update" really silent.
9207 [Ralf S. Engelschall]
9209 *) Recompiled the error-definition header files and added
9210 missing symbols to the Win32 linker tables.
9211 [Ralf S. Engelschall]
9213 *) Cleaned up the top-level documents;
9214 o new files: CHANGES and LICENSE
9215 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
9216 o merged COPYRIGHT into LICENSE
9217 o removed obsolete TODO file
9218 o renamed MICROSOFT to INSTALL.W32
9219 [Ralf S. Engelschall]
9221 *) Removed dummy files from the 0.9.1b source tree:
9222 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
9223 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
9224 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
9225 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
9226 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
9227 [Ralf S. Engelschall]
9229 *) Added various platform portability fixes.
9232 *) The Genesis of the OpenSSL rpject:
9233 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
9234 Young and Tim J. Hudson created while they were working for C2Net until
9236 [The OpenSSL Project]
9239 Changes between 0.9.0b and 0.9.1b [not released]
9241 *) Updated a few CA certificates under certs/
9244 *) Changed some BIGNUM api stuff.
9247 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
9248 DGUX x86, Linux Alpha, etc.
9251 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
9252 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
9256 *) Add -strparse option to asn1pars program which parses nested
9258 [Dr Stephen Henson <shenson@bigfoot.com>]
9260 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
9263 *) DSA fix for "ca" program.
9266 *) Added "-genkey" option to "dsaparam" program.
9269 *) Added RIPE MD160 (rmd160) message digest.
9272 *) Added -a (all) option to "ssleay version" command.
9275 *) Added PLATFORM define which is the id given to Configure.
9278 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
9281 *) Extended the ASN.1 parser routines.
9284 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
9287 *) Added a BN_CTX to the BN library.
9290 *) Fixed the weak key values in DES library
9293 *) Changed API in EVP library for cipher aliases.
9296 *) Added support for RC2/64bit cipher.
9299 *) Converted the lhash library to the crypto/mem.c functions.
9302 *) Added more recognized ASN.1 object ids.
9305 *) Added more RSA padding checks for SSL/TLS.
9308 *) Added BIO proxy/filter functionality.
9311 *) Added extra_certs to SSL_CTX which can be used
9312 send extra CA certificates to the client in the CA cert chain sending
9313 process. It can be configured with SSL_CTX_add_extra_chain_cert().
9316 *) Now Fortezza is denied in the authentication phase because
9317 this is key exchange mechanism is not supported by SSLeay at all.
9320 *) Additional PKCS1 checks.
9323 *) Support the string "TLSv1" for all TLS v1 ciphers.
9326 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
9327 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
9330 *) Fixed a few memory leaks.
9333 *) Fixed various code and comment typos.
9336 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
9337 bytes sent in the client random.
9338 [Edward Bishop <ebishop@spyglass.com>]
projects / openssl.git / blob