5 Changes between 1.0.0d and 1.0.1 [xx XXX xxxx]
7 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
11 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
12 FIPS modules versions.
15 *) Add TLS v1.2 client side support for client authentication. Keep cache
16 of handshake records longer as we don't know the hash algorithm to use
17 until after the certificate request message is received.
20 *) Initial TLS v1.2 client support. Add a default signature algorithms
21 extension including all the algorithms we support. Parse new signature
22 format in client key exchange. Relax some ECC signing restrictions for
23 TLS v1.2 as indicated in RFC5246.
26 *) Add server support for TLS v1.2 signature algorithms extension. Switch
27 to new signature format when needed using client digest preference.
28 All server ciphersuites should now work correctly in TLS v1.2. No client
29 support yet and no support for client certificates.
32 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
33 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
34 ciphersuites. At present only RSA key exchange ciphersuites work with
35 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
36 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
40 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
41 with this defined it will not be affected by any changes to ssl internal
42 structures. Add several utility functions to allow openssl application
43 to work with OPENSSL_NO_SSL_INTERN defined.
47 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
49 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
52 *) Add EC_GFp_nistp224_method(), a 64-bit optimized implementation for
53 elliptic curve NIST-P224 with constant-time single point multiplication on
54 typical inputs. EC_GROUP_new_by_curve_name() will automatically use this
55 (while EC_GROUP_new_curve_GFp() currently won't and prefers the more
56 flexible implementations).
58 The implementation requires support for the nonstandard type __uint128_t,
59 and so is disabled by default. To include this in your build of OpenSSL,
60 use -DEC_NISTP224_64_GCC_128 on the Configure (or config) command line,
61 and run "make depend" (or "make update").
62 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
64 *) Permit abbreviated handshakes when renegotiating using the function
65 SSL_renegotiate_abbreviated().
66 [Robin Seggelmann <seggelmann@fh-muenster.de>]
68 *) Add call to ENGINE_register_all_complete() to
69 ENGINE_load_builtin_engines(), so some implementations get used
70 automatically instead of needing explicit application support.
73 *) Add support for TLS key exporter as described in RFC5705.
74 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
76 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
77 a few changes are required:
79 Add SSL_OP_NO_TLSv1_1 flag.
81 Update version checking logic to handle version 1.1.
82 Add explicit IV handling (ported from DTLS code).
83 Add command line options to s_client/s_server.
86 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
88 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
89 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
91 *) Fix bug in string printing code: if *any* escaping is enabled we must
92 escape the escape character (backslash) or the resulting string is
96 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
98 *) Disable code workaround for ancient and obsolete Netscape browsers
99 and servers: an attacker can use it in a ciphersuite downgrade attack.
100 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
103 *) Fixed J-PAKE implementation error, originally discovered by
104 Sebastien Martini, further info and confirmation from Stefan
105 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
108 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
110 *) Fix extension code to avoid race conditions which can result in a buffer
111 overrun vulnerability: resumed sessions must not be modified as they can
112 be shared by multiple threads. CVE-2010-3864
115 *) Fix WIN32 build system to correctly link an ENGINE directory into
119 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
121 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
123 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
125 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
127 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
128 context. The operation can be customised via the ctrl mechanism in
129 case ENGINEs want to include additional functionality.
132 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
135 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
136 output hashes compatible with older versions of OpenSSL.
137 [Willy Weisz <weisz@vcpc.univie.ac.at>]
139 *) Fix compression algorithm handling: if resuming a session use the
140 compression algorithm of the resumed session instead of determining
141 it from client hello again. Don't allow server to change algorithm.
144 *) Add load_crls() function to apps tidying load_certs() too. Add option
145 to verify utility to allow additional CRLs to be included.
148 *) Update OCSP request code to permit adding custom headers to the request:
149 some responders need this.
152 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
154 [Julia Lawall <julia@diku.dk>]
156 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
157 needlessly dereferenced structures, used obsolete functions and
158 didn't handle all updated verify codes correctly.
161 *) Disable MD2 in the default configuration.
164 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
165 indicate the initial BIO being pushed or popped. This makes it possible
166 to determine whether the BIO is the one explicitly called or as a result
167 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
168 it handles reference counts correctly and doesn't zero out the I/O bio
169 when it is not being explicitly popped. WARNING: applications which
170 included workarounds for the old buggy behaviour will need to be modified
171 or they could free up already freed BIOs.
174 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
175 renaming to all platforms (within the 0.9.8 branch, this was
176 done conditionally on Netware platforms to avoid a name clash).
177 [Guenter <lists@gknw.net>]
179 *) Add ECDHE and PSK support to DTLS.
180 [Michael Tuexen <tuexen@fh-muenster.de>]
182 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
186 *) Add "missing" function EVP_MD_flags() (without this the only way to
187 retrieve a digest flags is by accessing the structure directly. Update
188 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
189 or cipher is registered as in the "from" argument. Print out all
190 registered digests in the dgst usage message instead of manually
191 attempting to work them out.
194 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
195 this allows the use of compression and extensions. Change default cipher
196 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
197 by default unless an application cipher string requests it.
200 *) Alter match criteria in PKCS12_parse(). It used to try to use local
201 key ids to find matching certificates and keys but some PKCS#12 files
202 don't follow the (somewhat unwritten) rules and this strategy fails.
203 Now just gather all certificates together and the first private key
204 then look for the first certificate that matches the key.
207 *) Support use of registered digest and cipher names for dgst and cipher
208 commands instead of having to add each one as a special case. So now
215 openssl dgst -sha256 foo
217 and this works for ENGINE based algorithms too.
221 *) Update Gost ENGINE to support parameter files.
222 [Victor B. Wagner <vitus@cryptocom.ru>]
224 *) Support GeneralizedTime in ca utility.
225 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
227 *) Enhance the hash format used for certificate directory links. The new
228 form uses the canonical encoding (meaning equivalent names will work
229 even if they aren't identical) and uses SHA1 instead of MD5. This form
230 is incompatible with the older format and as a result c_rehash should
231 be used to rebuild symbolic links.
234 *) Make PKCS#8 the default write format for private keys, replacing the
235 traditional format. This form is standardised, more secure and doesn't
236 include an implicit MD5 dependency.
239 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
240 committed to OpenSSL should pass this lot as a minimum.
243 *) Add session ticket override functionality for use by EAP-FAST.
244 [Jouni Malinen <j@w1.fi>]
246 *) Modify HMAC functions to return a value. Since these can be implemented
247 in an ENGINE errors can occur.
250 *) Type-checked OBJ_bsearch_ex.
253 *) Type-checked OBJ_bsearch. Also some constification necessitated
254 by type-checking. Still to come: TXT_DB, bsearch(?),
255 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
259 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
260 seconds to a tm structure directly, instead of going through OS
261 specific date routines. This avoids any issues with OS routines such
262 as the year 2038 bug. New *_adj() functions for ASN1 time structures
263 and X509_time_adj_ex() to cover the extended range. The existing
264 X509_time_adj() is still usable and will no longer have any date issues.
267 *) Delta CRL support. New use deltas option which will attempt to locate
268 and search any appropriate delta CRLs available.
270 This work was sponsored by Google.
273 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
274 code and add additional score elements. Validate alternate CRL paths
275 as part of the CRL checking and indicate a new error "CRL path validation
276 error" in this case. Applications wanting additional details can use
277 the verify callback and check the new "parent" field. If this is not
278 NULL CRL path validation is taking place. Existing applications wont
279 see this because it requires extended CRL support which is off by
282 This work was sponsored by Google.
285 *) Support for freshest CRL extension.
287 This work was sponsored by Google.
290 *) Initial indirect CRL support. Currently only supported in the CRLs
291 passed directly and not via lookup. Process certificate issuer
292 CRL entry extension and lookup CRL entries by bother issuer name
293 and serial number. Check and process CRL issuer entry in IDP extension.
295 This work was sponsored by Google.
298 *) Add support for distinct certificate and CRL paths. The CRL issuer
299 certificate is validated separately in this case. Only enabled if
300 an extended CRL support flag is set: this flag will enable additional
301 CRL functionality in future.
303 This work was sponsored by Google.
306 *) Add support for policy mappings extension.
308 This work was sponsored by Google.
311 *) Fixes to pathlength constraint, self issued certificate handling,
312 policy processing to align with RFC3280 and PKITS tests.
314 This work was sponsored by Google.
317 *) Support for name constraints certificate extension. DN, email, DNS
318 and URI types are currently supported.
320 This work was sponsored by Google.
323 *) To cater for systems that provide a pointer-based thread ID rather
324 than numeric, deprecate the current numeric thread ID mechanism and
325 replace it with a structure and associated callback type. This
326 mechanism allows a numeric "hash" to be extracted from a thread ID in
327 either case, and on platforms where pointers are larger than 'long',
328 mixing is done to help ensure the numeric 'hash' is usable even if it
329 can't be guaranteed unique. The default mechanism is to use "&errno"
330 as a pointer-based thread ID to distinguish between threads.
332 Applications that want to provide their own thread IDs should now use
333 CRYPTO_THREADID_set_callback() to register a callback that will call
334 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
336 Note that ERR_remove_state() is now deprecated, because it is tied
337 to the assumption that thread IDs are numeric. ERR_remove_state(0)
338 to free the current thread's error state should be replaced by
339 ERR_remove_thread_state(NULL).
341 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
342 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
343 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
344 application was previously providing a numeric thread callback that
345 was inappropriate for distinguishing threads, then uniqueness might
346 have been obtained with &errno that happened immediately in the
347 intermediate development versions of OpenSSL; this is no longer the
348 case, the numeric thread callback will now override the automatic use
350 [Geoff Thorpe, with help from Bodo Moeller]
352 *) Initial support for different CRL issuing certificates. This covers a
353 simple case where the self issued certificates in the chain exist and
354 the real CRL issuer is higher in the existing chain.
356 This work was sponsored by Google.
359 *) Removed effectively defunct crypto/store from the build.
362 *) Revamp of STACK to provide stronger type-checking. Still to come:
363 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
364 ASN1_STRING, CONF_VALUE.
367 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
368 RAM on SSL connections. This option can save about 34k per idle SSL.
371 *) Revamp of LHASH to provide stronger type-checking. Still to come:
372 STACK, TXT_DB, bsearch, qsort.
375 *) Initial support for Cryptographic Message Syntax (aka CMS) based
376 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
377 support for data, signedData, compressedData, digestedData and
378 encryptedData, envelopedData types included. Scripts to check against
379 RFC4134 examples draft and interop and consistency checks of many
380 content types and variants.
383 *) Add options to enc utility to support use of zlib compression BIO.
386 *) Extend mk1mf to support importing of options and assembly language
387 files from Configure script, currently only included in VC-WIN32.
388 The assembly language rules can now optionally generate the source
389 files from the associated perl scripts.
392 *) Implement remaining functionality needed to support GOST ciphersuites.
393 Interop testing has been performed using CryptoPro implementations.
394 [Victor B. Wagner <vitus@cryptocom.ru>]
396 *) s390x assembler pack.
399 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
403 *) Implement Opaque PRF Input TLS extension as specified in
404 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
405 official specification yet and no extension type assignment by
406 IANA exists, this extension (for now) will have to be explicitly
407 enabled when building OpenSSL by providing the extension number
408 to use. For example, specify an option
410 -DTLSEXT_TYPE_opaque_prf_input=0x9527
412 to the "config" or "Configure" script to enable the extension,
413 assuming extension number 0x9527 (which is a completely arbitrary
414 and unofficial assignment based on the MD5 hash of the Internet
415 Draft). Note that by doing so, you potentially lose
416 interoperability with other TLS implementations since these might
417 be using the same extension number for other purposes.
419 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
420 opaque PRF input value to use in the handshake. This will create
421 an interal copy of the length-'len' string at 'src', and will
422 return non-zero for success.
424 To get more control and flexibility, provide a callback function
427 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
428 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
432 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
435 Callback function 'cb' will be called in handshakes, and is
436 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
437 Argument 'arg' is for application purposes (the value as given to
438 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
439 be provided to the callback function). The callback function
440 has to return non-zero to report success: usually 1 to use opaque
441 PRF input just if possible, or 2 to enforce use of the opaque PRF
442 input. In the latter case, the library will abort the handshake
443 if opaque PRF input is not successfully negotiated.
445 Arguments 'peerinput' and 'len' given to the callback function
446 will always be NULL and 0 in the case of a client. A server will
447 see the client's opaque PRF input through these variables if
448 available (NULL and 0 otherwise). Note that if the server
449 provides an opaque PRF input, the length must be the same as the
450 length of the client's opaque PRF input.
452 Note that the callback function will only be called when creating
453 a new session (session resumption can resume whatever was
454 previously negotiated), and will not be called in SSL 2.0
455 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
456 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
457 for applications that need to enforce opaque PRF input.
461 *) Update ssl code to support digests other than SHA1+MD5 for handshake
464 [Victor B. Wagner <vitus@cryptocom.ru>]
466 *) Add RFC4507 support to OpenSSL. This includes the corrections in
467 RFC4507bis. The encrypted ticket format is an encrypted encoded
468 SSL_SESSION structure, that way new session features are automatically
471 If a client application caches session in an SSL_SESSION structure
472 support is transparent because tickets are now stored in the encoded
475 The SSL_CTX structure automatically generates keys for ticket
476 protection in servers so again support should be possible
477 with no application modification.
479 If a client or server wishes to disable RFC4507 support then the option
480 SSL_OP_NO_TICKET can be set.
482 Add a TLS extension debugging callback to allow the contents of any client
483 or server extensions to be examined.
485 This work was sponsored by Google.
488 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
489 OpenSSL should now compile cleanly on gcc 4.2
490 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
492 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
493 support including streaming MAC support: this is required for GOST
495 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
497 *) Add option -stream to use PKCS#7 streaming in smime utility. New
498 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
499 to output in BER and PEM format.
502 *) Experimental support for use of HMAC via EVP_PKEY interface. This
503 allows HMAC to be handled via the EVP_DigestSign*() interface. The
504 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
505 ENGINE support for HMAC keys which are unextractable. New -mac and
506 -macopt options to dgst utility.
509 *) New option -sigopt to dgst utility. Update dgst to use
510 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
511 alternative signing paramaters such as X9.31 or PSS in the dgst
515 *) Change ssl_cipher_apply_rule(), the internal function that does
516 the work each time a ciphersuite string requests enabling
517 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
518 removing ("!foo+bar") a class of ciphersuites: Now it maintains
519 the order of disabled ciphersuites such that those ciphersuites
520 that most recently went from enabled to disabled not only stay
521 in order with respect to each other, but also have higher priority
522 than other disabled ciphersuites the next time ciphersuites are
525 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
526 the same ciphersuites as with "HIGH" alone, but in a specific
527 order where the PSK ciphersuites come first (since they are the
528 most recently disabled ciphersuites when "HIGH" is parsed).
530 Also, change ssl_create_cipher_list() (using this new
531 funcionality) such that between otherwise identical
532 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
536 *) Change ssl_create_cipher_list() so that it automatically
537 arranges the ciphersuites in reasonable order before starting
538 to process the rule string. Thus, the definition for "DEFAULT"
539 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
540 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
541 This makes it much easier to arrive at a reasonable default order
542 in applications for which anonymous ciphers are OK (meaning
543 that you can't actually use DEFAULT).
544 [Bodo Moeller; suggested by Victor Duchovni]
546 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
547 processing) into multiple integers instead of setting
548 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
549 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
550 (These masks as well as the individual bit definitions are hidden
551 away into the non-exported interface ssl/ssl_locl.h, so this
552 change to the definition of the SSL_CIPHER structure shouldn't
553 affect applications.) This give us more bits for each of these
554 categories, so there is no longer a need to coagulate AES128 and
555 AES256 into a single algorithm bit, and to coagulate Camellia128
556 and Camellia256 into a single algorithm bit, which has led to all
559 Thus, among other things, the kludge introduced in 0.9.7m and
560 0.9.8e for masking out AES256 independently of AES128 or masking
561 out Camellia256 independently of AES256 is not needed here in 0.9.9.
563 With the change, we also introduce new ciphersuite aliases that
564 so far were missing: "AES128", "AES256", "CAMELLIA128", and
568 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
569 Use the leftmost N bytes of the signature input if the input is
570 larger than the prime q (with N being the size in bytes of q).
573 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
574 it yet and it is largely untested.
577 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
580 *) Initial incomplete changes to avoid need for function casts in OpenSSL
581 some compilers (gcc 4.2 and later) reject their use. Safestack is
582 reimplemented. Update ASN1 to avoid use of legacy functions.
585 *) Win32/64 targets are linked with Winsock2.
588 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
589 to external functions. This can be used to increase CRL handling
590 efficiency especially when CRLs are very large by (for example) storing
591 the CRL revoked certificates in a database.
594 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
595 new CRLs added to a directory can be used. New command line option
596 -verify_return_error to s_client and s_server. This causes real errors
597 to be returned by the verify callback instead of carrying on no matter
598 what. This reflects the way a "real world" verify callback would behave.
601 *) GOST engine, supporting several GOST algorithms and public key formats.
602 Kindly donated by Cryptocom.
605 *) Partial support for Issuing Distribution Point CRL extension. CRLs
606 partitioned by DP are handled but no indirect CRL or reason partitioning
607 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
608 selected via a scoring technique which handles IDP and AKID in CRLs.
611 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
612 will ultimately be used for all verify operations: this will remove the
613 X509_STORE dependency on certificate verification and allow alternative
614 lookup methods. X509_STORE based implementations of these two callbacks.
617 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
618 Modify get_crl() to find a valid (unexpired) CRL if possible.
621 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
622 this would be called X509_CRL_cmp() but that name is already used by
623 a function that just compares CRL issuer names. Cache several CRL
624 extensions in X509_CRL structure and cache CRLDP in X509.
627 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
628 this maps equivalent X509_NAME structures into a consistent structure.
629 Name comparison can then be performed rapidly using memcmp().
632 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
636 *) Allow digests to supply their own micalg string for S/MIME type using
637 the ctrl EVP_MD_CTRL_MICALG.
640 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
641 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
642 ctrl. It can then customise the structure before and/or after signing
646 *) New function OBJ_add_sigid() to allow application defined signature OIDs
647 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
648 to free up any added signature OIDs.
651 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
652 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
653 digest and cipher tables. New options added to openssl utility:
654 list-message-digest-algorithms and list-cipher-algorithms.
657 *) Change the array representation of binary polynomials: the list
658 of degrees of non-zero coefficients is now terminated with -1.
659 Previously it was terminated with 0, which was also part of the
660 value; thus, the array representation was not applicable to
661 polynomials where t^0 has coefficient zero. This change makes
662 the array representation useful in a more general context.
665 *) Various modifications and fixes to SSL/TLS cipher string
666 handling. For ECC, the code now distinguishes between fixed ECDH
667 with RSA certificates on the one hand and with ECDSA certificates
668 on the other hand, since these are separate ciphersuites. The
669 unused code for Fortezza ciphersuites has been removed.
671 For consistency with EDH, ephemeral ECDH is now called "EECDH"
672 (not "ECDHE"). For consistency with the code for DH
673 certificates, use of ECDH certificates is now considered ECDH
674 authentication, not RSA or ECDSA authentication (the latter is
675 merely the CA's signing algorithm and not actively used in the
678 The temporary ciphersuite alias "ECCdraft" is no longer
679 available, and ECC ciphersuites are no longer excluded from "ALL"
680 and "DEFAULT". The following aliases now exist for RFC 4492
681 ciphersuites, most of these by analogy with the DH case:
683 kECDHr - ECDH cert, signed with RSA
684 kECDHe - ECDH cert, signed with ECDSA
685 kECDH - ECDH cert (signed with either RSA or ECDSA)
686 kEECDH - ephemeral ECDH
687 ECDH - ECDH cert or ephemeral ECDH
693 AECDH - anonymous ECDH
694 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
698 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
699 Use correct micalg parameters depending on digest(s) in signed message.
702 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
703 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
706 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
707 an engine to register a method. Add ENGINE lookups for methods and
708 functional reference processing.
711 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
712 EVP_{Sign,Verify}* which allow an application to customise the signature
716 *) New -resign option to smime utility. This adds one or more signers
717 to an existing PKCS#7 signedData structure. Also -md option to use an
718 alternative message digest algorithm for signing.
721 *) Tidy up PKCS#7 routines and add new functions to make it easier to
722 create PKCS7 structures containing multiple signers. Update smime
723 application to support multiple signers.
726 *) New -macalg option to pkcs12 utility to allow setting of an alternative
730 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
731 Reorganize PBE internals to lookup from a static table using NIDs,
732 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
733 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
734 PRF which will be automatically used with PBES2.
737 *) Replace the algorithm specific calls to generate keys in "req" with the
741 *) Update PKCS#7 enveloped data routines to use new API. This is now
742 supported by any public key method supporting the encrypt operation. A
743 ctrl is added to allow the public key algorithm to examine or modify
744 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
748 *) Add a ctrl to asn1 method to allow a public key algorithm to express
749 a default digest type to use. In most cases this will be SHA1 but some
750 algorithms (such as GOST) need to specify an alternative digest. The
751 return value indicates how strong the prefernce is 1 means optional and
752 2 is mandatory (that is it is the only supported type). Modify
753 ASN1_item_sign() to accept a NULL digest argument to indicate it should
754 use the default md. Update openssl utilities to use the default digest
755 type for signing if it is not explicitly indicated.
758 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
759 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
760 signing method from the key type. This effectively removes the link
761 between digests and public key types.
764 *) Add an OID cross reference table and utility functions. Its purpose is to
765 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
766 rsaEncryption. This will allow some of the algorithm specific hackery
767 needed to use the correct OID to be removed.
770 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
771 structures for PKCS7_sign(). They are now set up by the relevant public
775 *) Add provisional EC pkey method with support for ECDSA and ECDH.
778 *) Add support for key derivation (agreement) in the API, DH method and
782 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
783 public and private key formats. As a side effect these add additional
784 command line functionality not previously available: DSA signatures can be
785 generated and verified using pkeyutl and DH key support and generation in
790 [Oliver Tappe <zooey@hirschkaefer.de>]
792 *) New make target "install_html_docs" installs HTML renditions of the
794 [Oliver Tappe <zooey@hirschkaefer.de>]
796 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
797 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
798 support key and parameter generation and add initial key generation
799 functionality for RSA.
802 *) Add functions for main EVP_PKEY_method operations. The undocumented
803 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
804 EVP_PKEY_{encrypt,decrypt}_old.
807 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
808 key API, doesn't do much yet.
811 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
812 public key algorithms. New option to openssl utility:
813 "list-public-key-algorithms" to print out info.
816 *) Implement the Supported Elliptic Curves Extension for
817 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
820 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
821 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
824 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
825 utilities such as rsa, dsa, dsaparam etc except they process any key
829 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
830 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
831 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
835 *) Initial support for pluggable public key ASN1.
836 De-spaghettify the public key ASN1 handling. Move public and private
837 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
838 algorithm specific handling to a single module within the relevant
839 algorithm directory. Add functions to allow (near) opaque processing
840 of public and private key structures.
843 *) Implement the Supported Point Formats Extension for
844 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
847 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
848 for the psk identity [hint] and the psk callback functions to the
849 SSL_SESSION, SSL and SSL_CTX structure.
852 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
856 SSL_CTX_use_psk_identity_hint
857 SSL_get_psk_identity_hint
859 SSL_use_psk_identity_hint
861 [Mika Kousa and Pasi Eronen of Nokia Corporation]
863 *) Add RFC 3161 compliant time stamp request creation, response generation
864 and response verification functionality.
865 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
867 *) Add initial support for TLS extensions, specifically for the server_name
868 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
869 have new members for a host name. The SSL data structure has an
870 additional member SSL_CTX *initial_ctx so that new sessions can be
871 stored in that context to allow for session resumption, even after the
872 SSL has been switched to a new SSL_CTX in reaction to a client's
873 server_name extension.
875 New functions (subject to change):
878 SSL_get_servername_type()
881 New CTRL codes and macros (subject to change):
883 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
884 - SSL_CTX_set_tlsext_servername_callback()
885 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
886 - SSL_CTX_set_tlsext_servername_arg()
887 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
889 openssl s_client has a new '-servername ...' option.
891 openssl s_server has new options '-servername_host ...', '-cert2 ...',
892 '-key2 ...', '-servername_fatal' (subject to change). This allows
893 testing the HostName extension for a specific single host name ('-cert'
894 and '-key' remain fallbacks for handshakes without HostName
895 negotiation). If the unrecogninzed_name alert has to be sent, this by
896 default is a warning; it becomes fatal with the '-servername_fatal'
899 [Peter Sylvester, Remy Allais, Christophe Renou]
901 *) Whirlpool hash implementation is added.
904 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
905 bn(64,32). Because of instruction set limitations it doesn't have
906 any negative impact on performance. This was done mostly in order
907 to make it possible to share assembler modules, such as bn_mul_mont
908 implementations, between 32- and 64-bit builds without hassle.
911 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
912 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
916 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
917 dedicated Montgomery multiplication procedure, is introduced.
918 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
919 "64-bit" performance on certain 32-bit targets.
922 *) New option SSL_OP_NO_COMP to disable use of compression selectively
923 in SSL structures. New SSL ctrl to set maximum send fragment size.
924 Save memory by seeting the I/O buffer sizes dynamically instead of
925 using the maximum available value.
928 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
929 in addition to the text details.
932 *) Very, very preliminary EXPERIMENTAL support for printing of general
933 ASN1 structures. This currently produces rather ugly output and doesn't
934 handle several customised structures at all.
937 *) Integrated support for PVK file format and some related formats such
938 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
939 these in the 'rsa' and 'dsa' utilities.
942 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
945 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
946 place for the (very old) "NETSCAPE" format certificates which are now
947 handled using new ASN1 code equivalents.
950 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
951 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
952 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
955 *) Modify CRL distribution points extension code to print out previously
956 unsupported fields. Enhance extension setting code to allow setting of
960 *) Add print and set support for Issuing Distribution Point CRL extension.
963 *) Change 'Configure' script to enable Camellia by default.
966 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
968 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
969 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
971 *) Fix bug in string printing code: if *any* escaping is enabled we must
972 escape the escape character (backslash) or the resulting string is
976 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
978 *) Disable code workaround for ancient and obsolete Netscape browsers
979 and servers: an attacker can use it in a ciphersuite downgrade attack.
980 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
983 *) Fixed J-PAKE implementation error, originally discovered by
984 Sebastien Martini, further info and confirmation from Stefan
985 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
988 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
990 *) Fix extension code to avoid race conditions which can result in a buffer
991 overrun vulnerability: resumed sessions must not be modified as they can
992 be shared by multiple threads. CVE-2010-3864
995 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
998 *) Don't reencode certificate when calculating signature: cache and use
999 the original encoding instead. This makes signature verification of
1000 some broken encodings work correctly.
1003 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
1004 is also one of the inputs.
1005 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1007 *) Don't repeatedly append PBE algorithms to table if they already exist.
1008 Sort table on each new add. This effectively makes the table read only
1009 after all algorithms are added and subsequent calls to PKCS12_pbe_add
1013 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
1015 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
1018 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
1019 access or freeing data twice (CVE-2010-0742)
1020 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
1022 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
1023 common in certificates and some applications which only call
1024 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
1028 Reduce copying into .apps and .test in makevms.com
1029 Don't try to use blank CA certificate in CA.com
1030 Allow use of C files from original directories in maketests.com
1031 [Steven M. Schweda" <sms@antinode.info>]
1033 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1035 *) When rejecting SSL/TLS records due to an incorrect version number, never
1036 update s->server with a new major version number. As of
1037 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1038 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1039 the previous behavior could result in a read attempt at NULL when
1040 receiving specific incorrect SSL/TLS records once record payload
1041 protection is active. (CVE-2010-0740)
1042 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1044 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1045 could be crashed if the relevant tables were not present (e.g. chrooted).
1046 [Tomas Hoger <thoger@redhat.com>]
1048 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1050 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1051 [Martin Olsson, Neel Mehta]
1053 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1054 accommodate for stack sorting, always a write lock!).
1057 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1058 excessive delays in the RAND_poll(): over a minute. As a workaround
1059 include a time check in the inner Heap32Next loop too.
1062 *) The code that handled flushing of data in SSL/TLS originally used the
1063 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1064 the problem outlined in PR#1949. The fix suggested there however can
1065 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1066 of Apache). So instead simplify the code to flush unconditionally.
1067 This should be fine since flushing with no data to flush is a no op.
1070 *) Handle TLS versions 2.0 and later properly and correctly use the
1071 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1072 off ancient servers have a habit of sticking around for a while...
1075 *) Modify compression code so it frees up structures without using the
1076 ex_data callbacks. This works around a problem where some applications
1077 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1078 restarting) then use compression (e.g. SSL with compression) later.
1079 This results in significant per-connection memory leaks and
1080 has caused some security issues including CVE-2008-1678 and
1084 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1085 change when encrypting or decrypting.
1088 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1089 connect and renegotiate with servers which do not support RI.
1090 Until RI is more widely deployed this option is enabled by default.
1093 *) Add "missing" ssl ctrls to clear options and mode.
1096 *) If client attempts to renegotiate and doesn't support RI respond with
1097 a no_renegotiation alert as required by RFC5746. Some renegotiating
1098 TLS clients will continue a connection gracefully when they receive
1099 the alert. Unfortunately OpenSSL mishandled this alert and would hang
1100 waiting for a server hello which it will never receive. Now we treat a
1101 received no_renegotiation alert as a fatal error. This is because
1102 applications requesting a renegotiation might well expect it to succeed
1103 and would have no code in place to handle the server denying it so the
1104 only safe thing to do is to terminate the connection.
1107 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
1108 peer supports secure renegotiation and 0 otherwise. Print out peer
1109 renegotiation support in s_client/s_server.
1112 *) Replace the highly broken and deprecated SPKAC certification method with
1113 the updated NID creation version. This should correctly handle UTF8.
1116 *) Implement RFC5746. Re-enable renegotiation but require the extension
1117 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
1118 turns out to be a bad idea. It has been replaced by
1119 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
1120 SSL_CTX_set_options(). This is really not recommended unless you
1121 know what you are doing.
1122 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
1124 *) Fixes to stateless session resumption handling. Use initial_ctx when
1125 issuing and attempting to decrypt tickets in case it has changed during
1126 servername handling. Use a non-zero length session ID when attempting
1127 stateless session resumption: this makes it possible to determine if
1128 a resumption has occurred immediately after receiving server hello
1129 (several places in OpenSSL subtly assume this) instead of later in
1133 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
1134 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
1135 fixes for a few places where the return code is not checked
1137 [Julia Lawall <julia@diku.dk>]
1139 *) Add --strict-warnings option to Configure script to include devteam
1140 warnings in other configurations.
1143 *) Add support for --libdir option and LIBDIR variable in makefiles. This
1144 makes it possible to install openssl libraries in locations which
1145 have names other than "lib", for example "/usr/lib64" which some
1147 [Steve Henson, based on patch from Jeremy Utley]
1149 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
1150 X690 8.9.12 and can produce some misleading textual output of OIDs.
1151 [Steve Henson, reported by Dan Kaminsky]
1153 *) Delete MD2 from algorithm tables. This follows the recommendation in
1154 several standards that it is not used in new applications due to
1155 several cryptographic weaknesses. For binary compatibility reasons
1156 the MD2 API is still compiled in by default.
1159 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
1163 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
1164 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
1166 [Guenter <lists@gknw.net>]
1168 *) Fix the server certificate chain building code to use X509_verify_cert(),
1169 it used to have an ad-hoc builder which was unable to cope with anything
1170 other than a simple chain.
1171 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
1173 *) Don't check self signed certificate signatures in X509_verify_cert()
1174 by default (a flag can override this): it just wastes time without
1175 adding any security. As a useful side effect self signed root CAs
1176 with non-FIPS digests are now usable in FIPS mode.
1179 *) In dtls1_process_out_of_seq_message() the check if the current message
1180 is already buffered was missing. For every new message was memory
1181 allocated, allowing an attacker to perform an denial of service attack
1182 with sending out of seq handshake messages until there is no memory
1183 left. Additionally every future messege was buffered, even if the
1184 sequence number made no sense and would be part of another handshake.
1185 So only messages with sequence numbers less than 10 in advance will be
1186 buffered. (CVE-2009-1378)
1187 [Robin Seggelmann, discovered by Daniel Mentz]
1189 *) Records are buffered if they arrive with a future epoch to be
1190 processed after finishing the corresponding handshake. There is
1191 currently no limitation to this buffer allowing an attacker to perform
1192 a DOS attack with sending records with future epochs until there is no
1193 memory left. This patch adds the pqueue_size() function to detemine
1194 the size of a buffer and limits the record buffer to 100 entries.
1196 [Robin Seggelmann, discovered by Daniel Mentz]
1198 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
1199 parent structure is freed. (CVE-2009-1379)
1202 *) Handle non-blocking I/O properly in SSL_shutdown() call.
1203 [Darryl Miles <darryl-mailinglists@netbauds.net>]
1206 [Ilya O. <vrghost@gmail.com>]
1208 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
1210 *) Disable renegotiation completely - this fixes a severe security
1211 problem (CVE-2009-3555) at the cost of breaking all
1212 renegotiation. Renegotiation can be re-enabled by setting
1213 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
1214 run-time. This is really not recommended unless you know what
1218 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
1220 *) Don't set val to NULL when freeing up structures, it is freed up by
1221 underlying code. If sizeof(void *) > sizeof(long) this can result in
1222 zeroing past the valid field. (CVE-2009-0789)
1223 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
1225 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
1226 checked correctly. This would allow some invalid signed attributes to
1227 appear to verify correctly. (CVE-2009-0591)
1228 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1230 *) Reject UniversalString and BMPString types with invalid lengths. This
1231 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
1232 a legal length. (CVE-2009-0590)
1235 *) Set S/MIME signing as the default purpose rather than setting it
1236 unconditionally. This allows applications to override it at the store
1240 *) Permit restricted recursion of ASN1 strings. This is needed in practice
1241 to handle some structures.
1244 *) Improve efficiency of mem_gets: don't search whole buffer each time
1246 [Jeremy Shapiro <jnshapir@us.ibm.com>]
1248 *) New -hex option for openssl rand.
1251 *) Print out UTF8String and NumericString when parsing ASN1.
1254 *) Support NumericString type for name components.
1257 *) Allow CC in the environment to override the automatically chosen
1258 compiler. Note that nothing is done to ensure flags work with the
1262 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
1264 *) Properly check EVP_VerifyFinal() and similar return values
1266 [Ben Laurie, Bodo Moeller, Google Security Team]
1268 *) Enable TLS extensions by default.
1271 *) Allow the CHIL engine to be loaded, whether the application is
1272 multithreaded or not. (This does not release the developer from the
1273 obligation to set up the dynamic locking callbacks.)
1274 [Sander Temme <sander@temme.net>]
1276 *) Use correct exit code if there is an error in dgst command.
1277 [Steve Henson; problem pointed out by Roland Dirlewanger]
1279 *) Tweak Configure so that you need to say "experimental-jpake" to enable
1280 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
1283 *) Add experimental JPAKE support, including demo authentication in
1284 s_client and s_server.
1287 *) Set the comparison function in v3_addr_canonize().
1288 [Rob Austein <sra@hactrn.net>]
1290 *) Add support for XMPP STARTTLS in s_client.
1291 [Philip Paeps <philip@freebsd.org>]
1293 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
1294 to ensure that even with this option, only ciphersuites in the
1295 server's preference list will be accepted. (Note that the option
1296 applies only when resuming a session, so the earlier behavior was
1297 just about the algorithm choice for symmetric cryptography.)
1300 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
1302 *) Fix NULL pointer dereference if a DTLS server received
1303 ChangeCipherSpec as first record (CVE-2009-1386).
1306 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
1307 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
1310 *) The fix in 0.9.8c that supposedly got rid of unsafe
1311 double-checked locking was incomplete for RSA blinding,
1312 addressing just one layer of what turns out to have been
1313 doubly unsafe triple-checked locking.
1315 So now fix this for real by retiring the MONT_HELPER macro
1316 in crypto/rsa/rsa_eay.c.
1318 [Bodo Moeller; problem pointed out by Marius Schilder]
1320 *) Various precautionary measures:
1322 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
1324 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
1325 (NB: This would require knowledge of the secret session ticket key
1326 to exploit, in which case you'd be SOL either way.)
1328 - Change bn_nist.c so that it will properly handle input BIGNUMs
1329 outside the expected range.
1331 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
1334 [Neel Mehta, Bodo Moeller]
1336 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
1337 the load fails. Useful for distros.
1338 [Ben Laurie and the FreeBSD team]
1340 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
1343 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
1346 *) Expand ENGINE to support engine supplied SSL client certificate functions.
1348 This work was sponsored by Logica.
1351 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
1352 keystores. Support for SSL/TLS client authentication too.
1353 Not compiled unless enable-capieng specified to Configure.
1355 This work was sponsored by Logica.
1358 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
1359 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
1360 attribute creation routines such as certifcate requests and PKCS#12
1364 Changes between 0.9.8g and 0.9.8h [28 May 2008]
1366 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
1367 handshake which could lead to a cilent crash as found using the
1368 Codenomicon TLS test suite (CVE-2008-1672)
1369 [Steve Henson, Mark Cox]
1371 *) Fix double free in TLS server name extensions which could lead to
1372 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
1375 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
1377 Clear the error queue to ensure that error entries left from
1378 older function calls do not interfere with the correct operation.
1379 [Lutz Jaenicke, Erik de Castro Lopo]
1381 *) Remove root CA certificates of commercial CAs:
1383 The OpenSSL project does not recommend any specific CA and does not
1384 have any policy with respect to including or excluding any CA.
1385 Therefore it does not make any sense to ship an arbitrary selection
1386 of root CA certificates with the OpenSSL software.
1389 *) RSA OAEP patches to fix two separate invalid memory reads.
1390 The first one involves inputs when 'lzero' is greater than
1391 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
1392 before the beginning of from). The second one involves inputs where
1393 the 'db' section contains nothing but zeroes (there is a one-byte
1394 invalid read after the end of 'db').
1395 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1397 *) Partial backport from 0.9.9-dev:
1399 Introduce bn_mul_mont (dedicated Montgomery multiplication
1400 procedure) as a candidate for BIGNUM assembler implementation.
1401 While 0.9.9-dev uses assembler for various architectures, only
1402 x86_64 is available by default here in the 0.9.8 branch, and
1403 32-bit x86 is available through a compile-time setting.
1405 To try the 32-bit x86 assembler implementation, use Configure
1406 option "enable-montasm" (which exists only for this backport).
1408 As "enable-montasm" for 32-bit x86 disclaims code stability
1409 anyway, in this constellation we activate additional code
1410 backported from 0.9.9-dev for further performance improvements,
1411 namely BN_from_montgomery_word. (To enable this otherwise,
1412 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
1414 [Andy Polyakov (backport partially by Bodo Moeller)]
1416 *) Add TLS session ticket callback. This allows an application to set
1417 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
1418 values. This is useful for key rollover for example where several key
1419 sets may exist with different names.
1422 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
1423 This was broken until now in 0.9.8 releases, such that the only way
1424 a registered ENGINE could be used (assuming it initialises
1425 successfully on the host) was to explicitly set it as the default
1426 for the relevant algorithms. This is in contradiction with 0.9.7
1427 behaviour and the documentation. With this fix, when an ENGINE is
1428 registered into a given algorithm's table of implementations, the
1429 'uptodate' flag is reset so that auto-discovery will be used next
1430 time a new context for that algorithm attempts to select an
1432 [Ian Lister (tweaked by Geoff Thorpe)]
1434 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
1435 implemention in the following ways:
1437 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
1440 Lack of BER streaming support means one pass streaming processing is
1441 only supported if data is detached: setting the streaming flag is
1442 ignored for embedded content.
1444 CMS support is disabled by default and must be explicitly enabled
1445 with the enable-cms configuration option.
1448 *) Update the GMP engine glue to do direct copies between BIGNUM and
1449 mpz_t when openssl and GMP use the same limb size. Otherwise the
1450 existing "conversion via a text string export" trick is still used.
1451 [Paul Sheer <paulsheer@gmail.com>]
1453 *) Zlib compression BIO. This is a filter BIO which compressed and
1454 uncompresses any data passed through it.
1457 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
1458 RFC3394 compatible AES key wrapping.
1461 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
1462 sets string data without copying. X509_ALGOR_set0() and
1463 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
1464 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
1465 from an X509_ATTRIBUTE structure optionally checking it occurs only
1466 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
1470 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
1471 to get the expected BN_FLG_CONSTTIME behavior.
1472 [Bodo Moeller (Google)]
1476 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
1477 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
1478 - added some more tests to do_tests.pl
1479 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
1480 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
1481 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
1482 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
1483 - various changes to netware.pl to enable gcc-cross builds on Win32
1485 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
1486 - various changes to fix missing prototype warnings
1487 - fixed x86nasm.pl to create correct asm files for NASM COFF output
1488 - added AES, WHIRLPOOL and CPUID assembler code to build files
1489 - added missing AES assembler make rules to mk1mf.pl
1490 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
1491 [Guenter Knauf <eflash@gmx.net>]
1493 *) Implement certificate status request TLS extension defined in RFC3546.
1494 A client can set the appropriate parameters and receive the encoded
1495 OCSP response via a callback. A server can query the supplied parameters
1496 and set the encoded OCSP response in the callback. Add simplified examples
1497 to s_client and s_server.
1500 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
1502 *) Fix various bugs:
1503 + Binary incompatibility of ssl_ctx_st structure
1504 + DTLS interoperation with non-compliant servers
1505 + Don't call get_session_cb() without proposed session
1506 + Fix ia64 assembler code
1507 [Andy Polyakov, Steve Henson]
1509 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
1511 *) DTLS Handshake overhaul. There were longstanding issues with
1512 OpenSSL DTLS implementation, which were making it impossible for
1513 RFC 4347 compliant client to communicate with OpenSSL server.
1514 Unfortunately just fixing these incompatibilities would "cut off"
1515 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
1516 server keeps tolerating non RFC compliant syntax. The opposite is
1517 not true, 0.9.8f client can not communicate with earlier server.
1518 This update even addresses CVE-2007-4995.
1521 *) Changes to avoid need for function casts in OpenSSL: some compilers
1522 (gcc 4.2 and later) reject their use.
1523 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
1526 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1527 RFC4507bis. The encrypted ticket format is an encrypted encoded
1528 SSL_SESSION structure, that way new session features are automatically
1531 If a client application caches session in an SSL_SESSION structure
1532 support is transparent because tickets are now stored in the encoded
1535 The SSL_CTX structure automatically generates keys for ticket
1536 protection in servers so again support should be possible
1537 with no application modification.
1539 If a client or server wishes to disable RFC4507 support then the option
1540 SSL_OP_NO_TICKET can be set.
1542 Add a TLS extension debugging callback to allow the contents of any client
1543 or server extensions to be examined.
1545 This work was sponsored by Google.
1548 *) Add initial support for TLS extensions, specifically for the server_name
1549 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1550 have new members for a host name. The SSL data structure has an
1551 additional member SSL_CTX *initial_ctx so that new sessions can be
1552 stored in that context to allow for session resumption, even after the
1553 SSL has been switched to a new SSL_CTX in reaction to a client's
1554 server_name extension.
1556 New functions (subject to change):
1558 SSL_get_servername()
1559 SSL_get_servername_type()
1562 New CTRL codes and macros (subject to change):
1564 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1565 - SSL_CTX_set_tlsext_servername_callback()
1566 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1567 - SSL_CTX_set_tlsext_servername_arg()
1568 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1570 openssl s_client has a new '-servername ...' option.
1572 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1573 '-key2 ...', '-servername_fatal' (subject to change). This allows
1574 testing the HostName extension for a specific single host name ('-cert'
1575 and '-key' remain fallbacks for handshakes without HostName
1576 negotiation). If the unrecogninzed_name alert has to be sent, this by
1577 default is a warning; it becomes fatal with the '-servername_fatal'
1580 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
1582 *) Add AES and SSE2 assembly language support to VC++ build.
1585 *) Mitigate attack on final subtraction in Montgomery reduction.
1588 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
1589 (which previously caused an internal error).
1592 *) Squeeze another 10% out of IGE mode when in != out.
1595 *) AES IGE mode speedup.
1596 [Dean Gaudet (Google)]
1598 *) Add the Korean symmetric 128-bit cipher SEED (see
1599 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
1600 add SEED ciphersuites from RFC 4162:
1602 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
1603 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
1604 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
1605 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
1607 To minimize changes between patchlevels in the OpenSSL 0.9.8
1608 series, SEED remains excluded from compilation unless OpenSSL
1609 is configured with 'enable-seed'.
1610 [KISA, Bodo Moeller]
1612 *) Mitigate branch prediction attacks, which can be practical if a
1613 single processor is shared, allowing a spy process to extract
1614 information. For detailed background information, see
1615 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
1616 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
1617 and Necessary Software Countermeasures"). The core of the change
1618 are new versions BN_div_no_branch() and
1619 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
1620 respectively, which are slower, but avoid the security-relevant
1621 conditional branches. These are automatically called by BN_div()
1622 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
1623 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
1624 remove a conditional branch.
1626 BN_FLG_CONSTTIME is the new name for the previous
1627 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
1628 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
1629 in the exponent causes BN_mod_exp_mont() to use the alternative
1630 implementation in BN_mod_exp_mont_consttime().) The old name
1631 remains as a deprecated alias.
1633 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
1634 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
1635 constant-time implementations for more than just exponentiation.
1636 Here too the old name is kept as a deprecated alias.
1638 BN_BLINDING_new() will now use BN_dup() for the modulus so that
1639 the BN_BLINDING structure gets an independent copy of the
1640 modulus. This means that the previous "BIGNUM *m" argument to
1641 BN_BLINDING_new() and to BN_BLINDING_create_param() now
1642 essentially becomes "const BIGNUM *m", although we can't actually
1643 change this in the header file before 0.9.9. It allows
1644 RSA_setup_blinding() to use BN_with_flags() on the modulus to
1645 enable BN_FLG_CONSTTIME.
1647 [Matthew D Wood (Intel Corp)]
1649 *) In the SSL/TLS server implementation, be strict about session ID
1650 context matching (which matters if an application uses a single
1651 external cache for different purposes). Previously,
1652 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
1653 set. This did ensure strict client verification, but meant that,
1654 with applications using a single external cache for quite
1655 different requirements, clients could circumvent ciphersuite
1656 restrictions for a given session ID context by starting a session
1657 in a different context.
1660 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
1661 a ciphersuite string such as "DEFAULT:RSA" cannot enable
1662 authentication-only ciphersuites.
1665 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
1666 not complete and could lead to a possible single byte overflow
1667 (CVE-2007-5135) [Ben Laurie]
1669 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
1671 *) Since AES128 and AES256 (and similarly Camellia128 and
1672 Camellia256) share a single mask bit in the logic of
1673 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
1674 kludge to work properly if AES128 is available and AES256 isn't
1675 (or if Camellia128 is available and Camellia256 isn't).
1678 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
1679 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
1680 When a point or a seed is encoded in a BIT STRING, we need to
1681 prevent the removal of trailing zero bits to get the proper DER
1682 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
1683 of a NamedBitList, for which trailing 0 bits need to be removed.)
1686 *) Have SSL/TLS server implementation tolerate "mismatched" record
1687 protocol version while receiving ClientHello even if the
1688 ClientHello is fragmented. (The server can't insist on the
1689 particular protocol version it has chosen before the ServerHello
1690 message has informed the client about his choice.)
1693 *) Add RFC 3779 support.
1694 [Rob Austein for ARIN, Ben Laurie]
1696 *) Load error codes if they are not already present instead of using a
1697 static variable. This allows them to be cleanly unloaded and reloaded.
1698 Improve header file function name parsing.
1701 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
1702 or CAPABILITY handshake as required by RFCs.
1705 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
1707 *) Introduce limits to prevent malicious keys being able to
1708 cause a denial of service. (CVE-2006-2940)
1709 [Steve Henson, Bodo Moeller]
1711 *) Fix ASN.1 parsing of certain invalid structures that can result
1712 in a denial of service. (CVE-2006-2937) [Steve Henson]
1714 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
1715 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
1717 *) Fix SSL client code which could crash if connecting to a
1718 malicious SSLv2 server. (CVE-2006-4343)
1719 [Tavis Ormandy and Will Drewry, Google Security Team]
1721 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
1722 match only those. Before that, "AES256-SHA" would be interpreted
1723 as a pattern and match "AES128-SHA" too (since AES128-SHA got
1724 the same strength classification in 0.9.7h) as we currently only
1725 have a single AES bit in the ciphersuite description bitmap.
1726 That change, however, also applied to ciphersuite strings such as
1727 "RC4-MD5" that intentionally matched multiple ciphersuites --
1728 namely, SSL 2.0 ciphersuites in addition to the more common ones
1729 from SSL 3.0/TLS 1.0.
1731 So we change the selection algorithm again: Naming an explicit
1732 ciphersuite selects this one ciphersuite, and any other similar
1733 ciphersuite (same bitmap) from *other* protocol versions.
1734 Thus, "RC4-MD5" again will properly select both the SSL 2.0
1735 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
1737 Since SSL 2.0 does not have any ciphersuites for which the
1738 128/256 bit distinction would be relevant, this works for now.
1739 The proper fix will be to use different bits for AES128 and
1740 AES256, which would have avoided the problems from the beginning;
1741 however, bits are scarce, so we can only do this in a new release
1742 (not just a patchlevel) when we can change the SSL_CIPHER
1743 definition to split the single 'unsigned long mask' bitmap into
1744 multiple values to extend the available space.
1748 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
1750 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
1751 (CVE-2006-4339) [Ben Laurie and Google Security Team]
1753 *) Add AES IGE and biIGE modes.
1756 *) Change the Unix randomness entropy gathering to use poll() when
1757 possible instead of select(), since the latter has some
1758 undesirable limitations.
1759 [Darryl Miles via Richard Levitte and Bodo Moeller]
1761 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
1762 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
1763 cannot be implicitly activated as part of, e.g., the "AES" alias.
1764 However, please upgrade to OpenSSL 0.9.9[-dev] for
1765 non-experimental use of the ECC ciphersuites to get TLS extension
1766 support, which is required for curve and point format negotiation
1767 to avoid potential handshake problems.
1770 *) Disable rogue ciphersuites:
1772 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
1773 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
1774 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
1776 The latter two were purportedly from
1777 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
1780 Also deactivate the remaining ciphersuites from
1781 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
1782 unofficial, and the ID has long expired.
1785 *) Fix RSA blinding Heisenbug (problems sometimes occured on
1786 dual-core machines) and other potential thread-safety issues.
1789 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
1790 versions), which is now available for royalty-free use
1791 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
1792 Also, add Camellia TLS ciphersuites from RFC 4132.
1794 To minimize changes between patchlevels in the OpenSSL 0.9.8
1795 series, Camellia remains excluded from compilation unless OpenSSL
1796 is configured with 'enable-camellia'.
1799 *) Disable the padding bug check when compression is in use. The padding
1800 bug check assumes the first packet is of even length, this is not
1801 necessarily true if compresssion is enabled and can result in false
1802 positives causing handshake failure. The actual bug test is ancient
1803 code so it is hoped that implementations will either have fixed it by
1804 now or any which still have the bug do not support compression.
1807 Changes between 0.9.8a and 0.9.8b [04 May 2006]
1809 *) When applying a cipher rule check to see if string match is an explicit
1810 cipher suite and only match that one cipher suite if it is.
1813 *) Link in manifests for VC++ if needed.
1814 [Austin Ziegler <halostatue@gmail.com>]
1816 *) Update support for ECC-based TLS ciphersuites according to
1817 draft-ietf-tls-ecc-12.txt with proposed changes (but without
1818 TLS extensions, which are supported starting with the 0.9.9
1819 branch, not in the OpenSSL 0.9.8 branch).
1822 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
1823 opaque EVP_CIPHER_CTX handling.
1826 *) Fixes and enhancements to zlib compression code. We now only use
1827 "zlib1.dll" and use the default __cdecl calling convention on Win32
1828 to conform with the standards mentioned here:
1829 http://www.zlib.net/DLL_FAQ.txt
1830 Static zlib linking now works on Windows and the new --with-zlib-include
1831 --with-zlib-lib options to Configure can be used to supply the location
1832 of the headers and library. Gracefully handle case where zlib library
1836 *) Several fixes and enhancements to the OID generation code. The old code
1837 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
1838 handle numbers larger than ULONG_MAX, truncated printing and had a
1839 non standard OBJ_obj2txt() behaviour.
1842 *) Add support for building of engines under engine/ as shared libraries
1843 under VC++ build system.
1846 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
1847 Hopefully, we will not see any false combination of paths any more.
1850 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
1852 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1853 (part of SSL_OP_ALL). This option used to disable the
1854 countermeasure against man-in-the-middle protocol-version
1855 rollback in the SSL 2.0 server implementation, which is a bad
1856 idea. (CVE-2005-2969)
1858 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
1859 for Information Security, National Institute of Advanced Industrial
1860 Science and Technology [AIST], Japan)]
1862 *) Add two function to clear and return the verify parameter flags.
1865 *) Keep cipherlists sorted in the source instead of sorting them at
1866 runtime, thus removing the need for a lock.
1869 *) Avoid some small subgroup attacks in Diffie-Hellman.
1870 [Nick Mathewson and Ben Laurie]
1872 *) Add functions for well-known primes.
1875 *) Extended Windows CE support.
1876 [Satoshi Nakamura and Andy Polyakov]
1878 *) Initialize SSL_METHOD structures at compile time instead of during
1879 runtime, thus removing the need for a lock.
1882 *) Make PKCS7_decrypt() work even if no certificate is supplied by
1883 attempting to decrypt each encrypted key in turn. Add support to
1887 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
1889 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
1892 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
1895 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
1896 key into the same file any more.
1899 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
1902 *) Add -utf8 command line and config file option to 'ca'.
1903 [Stefan <stf@udoma.org]
1905 *) Removed the macro des_crypt(), as it seems to conflict with some
1906 libraries. Use DES_crypt().
1909 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
1910 involves renaming the source and generated shared-libs for
1911 both. The engines will accept the corrected or legacy ids
1912 ('ncipher' and '4758_cca' respectively) when binding. NB,
1913 this only applies when building 'shared'.
1914 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
1916 *) Add attribute functions to EVP_PKEY structure. Modify
1917 PKCS12_create() to recognize a CSP name attribute and
1918 use it. Make -CSP option work again in pkcs12 utility.
1921 *) Add new functionality to the bn blinding code:
1922 - automatic re-creation of the BN_BLINDING parameters after
1923 a fixed number of uses (currently 32)
1924 - add new function for parameter creation
1925 - introduce flags to control the update behaviour of the
1926 BN_BLINDING parameters
1927 - hide BN_BLINDING structure
1928 Add a second BN_BLINDING slot to the RSA structure to improve
1929 performance when a single RSA object is shared among several
1933 *) Add support for DTLS.
1934 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
1936 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
1937 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
1940 *) Remove buggy and incompletet DH cert support from
1941 ssl/ssl_rsa.c and ssl/s3_both.c
1944 *) Use SHA-1 instead of MD5 as the default digest algorithm for
1945 the apps/openssl applications.
1948 *) Compile clean with "-Wall -Wmissing-prototypes
1949 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
1950 DEBUG_SAFESTACK must also be set.
1953 *) Change ./Configure so that certain algorithms can be disabled by default.
1954 The new counterpiece to "no-xxx" is "enable-xxx".
1956 The patented RC5 and MDC2 algorithms will now be disabled unless
1957 "enable-rc5" and "enable-mdc2", respectively, are specified.
1959 (IDEA remains enabled despite being patented. This is because IDEA
1960 is frequently required for interoperability, and there is no license
1961 fee for non-commercial use. As before, "no-idea" can be used to
1962 avoid this algorithm.)
1966 *) Add processing of proxy certificates (see RFC 3820). This work was
1967 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
1968 EGEE (Enabling Grids for E-science in Europe).
1971 *) RC4 performance overhaul on modern architectures/implementations, such
1972 as Intel P4, IA-64 and AMD64.
1975 *) New utility extract-section.pl. This can be used specify an alternative
1976 section number in a pod file instead of having to treat each file as
1977 a separate case in Makefile. This can be done by adding two lines to the
1980 =for comment openssl_section:XXX
1982 The blank line is mandatory.
1986 *) New arguments -certform, -keyform and -pass for s_client and s_server
1987 to allow alternative format key and certificate files and passphrase
1991 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
1992 update associated structures and add various utility functions.
1994 Add new policy related verify parameters, include policy checking in
1995 standard verify code. Enhance 'smime' application with extra parameters
1996 to support policy checking and print out.
1999 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2000 Nehemiah processors. These extensions support AES encryption in hardware
2001 as well as RNG (though RNG support is currently disabled).
2002 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2004 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2007 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2008 [Andy Polyakov and a number of other people]
2010 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2011 implementation contributed by IBM.
2012 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2014 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2015 exponent rather than 'unsigned long'. There is a corresponding change to
2016 the new 'rsa_keygen' element of the RSA_METHOD structure.
2017 [Jelte Jansen, Geoff Thorpe]
2019 *) Functionality for creating the initial serial number file is now
2020 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2022 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2023 number file to 1, which is bound to cause problems. To avoid
2024 the problems while respecting compatibility between different 0.9.7
2025 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2026 CA.pl for serial number initialization. With the new release 0.9.8,
2027 we can fix the problem directly in the 'ca' utility.)
2030 *) Reduced header interdepencies by declaring more opaque objects in
2031 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2032 give fewer recursive includes, which could break lazy source code - so
2033 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2034 developers should define this symbol when building and using openssl to
2035 ensure they track the recommended behaviour, interfaces, [etc], but
2036 backwards-compatible behaviour prevails when this isn't defined.
2039 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2042 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2043 This will generate a random key of the appropriate length based on the
2044 cipher context. The EVP_CIPHER can provide its own random key generation
2045 routine to support keys of a specific form. This is used in the des and
2046 3des routines to generate a key of the correct parity. Update S/MIME
2047 code to use new functions and hence generate correct parity DES keys.
2048 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2049 valid (weak or incorrect parity).
2052 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2053 as looking them up. This is useful when the verified structure may contain
2054 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2055 present unless the new PKCS7_NO_CRL flag is asserted.
2058 *) Extend ASN1 oid configuration module. It now additionally accepts the
2061 shortName = some long name, 1.2.3.4
2064 *) Reimplemented the BN_CTX implementation. There is now no more static
2065 limitation on the number of variables it can handle nor the depth of the
2066 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2067 information can now expand as required, and rather than having a single
2068 static array of bignums, BN_CTX now uses a linked-list of such arrays
2069 allowing it to expand on demand whilst maintaining the usefulness of
2070 BN_CTX's "bundling".
2073 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2074 to allow all RSA operations to function using a single BN_CTX.
2077 *) Preliminary support for certificate policy evaluation and checking. This
2078 is initially intended to pass the tests outlined in "Conformance Testing
2079 of Relying Party Client Certificate Path Processing Logic" v1.07.
2082 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2083 remained unused and not that useful. A variety of other little bignum
2084 tweaks and fixes have also been made continuing on from the audit (see
2088 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2089 associated ASN1, EVP and SSL functions and old ASN1 macros.
2092 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2093 and this should never fail. So the return value from the use of
2094 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2095 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2098 *) BN_CTX_get() should return zero-valued bignums, providing the same
2099 initialised value as BN_new().
2100 [Geoff Thorpe, suggested by Ulf Möller]
2102 *) Support for inhibitAnyPolicy certificate extension.
2105 *) An audit of the BIGNUM code is underway, for which debugging code is
2106 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
2107 is considered valid when processing BIGNUMs, and causes execution to
2108 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
2109 further steps are taken to deliberately pollute unused data in BIGNUM
2110 structures to try and expose faulty code further on. For now, openssl will
2111 (in its default mode of operation) continue to tolerate the inconsistent
2112 forms that it has tolerated in the past, but authors and packagers should
2113 consider trying openssl and their own applications when compiled with
2114 these debugging symbols defined. It will help highlight potential bugs in
2115 their own code, and will improve the test coverage for OpenSSL itself. At
2116 some point, these tighter rules will become openssl's default to improve
2117 maintainability, though the assert()s and other overheads will remain only
2118 in debugging configurations. See bn.h for more details.
2119 [Geoff Thorpe, Nils Larsch, Ulf Möller]
2121 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
2122 that can only be obtained through BN_CTX_new() (which implicitly
2123 initialises it). The presence of this function only made it possible
2124 to overwrite an existing structure (and cause memory leaks).
2127 *) Because of the callback-based approach for implementing LHASH as a
2128 template type, lh_insert() adds opaque objects to hash-tables and
2129 lh_doall() or lh_doall_arg() are typically used with a destructor callback
2130 to clean up those corresponding objects before destroying the hash table
2131 (and losing the object pointers). So some over-zealous constifications in
2132 LHASH have been relaxed so that lh_insert() does not take (nor store) the
2133 objects as "const" and the lh_doall[_arg] callback wrappers are not
2134 prototyped to have "const" restrictions on the object pointers they are
2135 given (and so aren't required to cast them away any more).
2138 *) The tmdiff.h API was so ugly and minimal that our own timing utility
2139 (speed) prefers to use its own implementation. The two implementations
2140 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
2141 its object type properly exposed (MS_TM) instead of casting to/from "char
2142 *". This may still change yet if someone realises MS_TM and "ms_time_***"
2143 aren't necessarily the greatest nomenclatures - but this is what was used
2144 internally to the implementation so I've used that for now.
2147 *) Ensure that deprecated functions do not get compiled when
2148 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
2149 the self-tests were still using deprecated key-generation functions so
2150 these have been updated also.
2153 *) Reorganise PKCS#7 code to separate the digest location functionality
2154 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
2155 New function PKCS7_set_digest() to set the digest type for PKCS#7
2156 digestedData type. Add additional code to correctly generate the
2157 digestedData type and add support for this type in PKCS7 initialization
2161 *) New function PKCS7_set0_type_other() this initializes a PKCS7
2162 structure of type "other".
2165 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
2166 sure the loop does correctly stop and breaking ("division by zero")
2167 modulus operations are not performed. The (pre-generated) prime
2168 table crypto/bn/bn_prime.h was already correct, but it could not be
2169 re-generated on some platforms because of the "division by zero"
2170 situation in the script.
2171 [Ralf S. Engelschall]
2173 *) Update support for ECC-based TLS ciphersuites according to
2174 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
2175 SHA-1 now is only used for "small" curves (where the
2176 representation of a field element takes up to 24 bytes); for
2177 larger curves, the field element resulting from ECDH is directly
2178 used as premaster secret.
2179 [Douglas Stebila (Sun Microsystems Laboratories)]
2181 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
2182 curve secp160r1 to the tests.
2183 [Douglas Stebila (Sun Microsystems Laboratories)]
2185 *) Add the possibility to load symbols globally with DSO.
2186 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
2188 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
2189 control of the error stack.
2192 *) Add support for STORE in ENGINE.
2195 *) Add the STORE type. The intention is to provide a common interface
2196 to certificate and key stores, be they simple file-based stores, or
2197 HSM-type store, or LDAP stores, or...
2198 NOTE: The code is currently UNTESTED and isn't really used anywhere.
2201 *) Add a generic structure called OPENSSL_ITEM. This can be used to
2202 pass a list of arguments to any function as well as provide a way
2203 for a function to pass data back to the caller.
2206 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
2207 works like BUF_strdup() but can be used to duplicate a portion of
2208 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
2212 *) Add the function sk_find_ex() which works like sk_find(), but will
2213 return an index to an element even if an exact match couldn't be
2214 found. The index is guaranteed to point at the element where the
2215 searched-for key would be inserted to preserve sorting order.
2218 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
2219 takes an extra flags argument for optional functionality. Currently,
2220 the following flags are defined:
2222 OBJ_BSEARCH_VALUE_ON_NOMATCH
2223 This one gets OBJ_bsearch_ex() to return a pointer to the first
2224 element where the comparing function returns a negative or zero
2227 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
2228 This one gets OBJ_bsearch_ex() to return a pointer to the first
2229 element where the comparing function returns zero. This is useful
2230 if there are more than one element where the comparing function
2234 *) Make it possible to create self-signed certificates with 'openssl ca'
2235 in such a way that the self-signed certificate becomes part of the
2236 CA database and uses the same mechanisms for serial number generation
2237 as all other certificate signing. The new flag '-selfsign' enables
2238 this functionality. Adapt CA.sh and CA.pl.in.
2241 *) Add functionality to check the public key of a certificate request
2242 against a given private. This is useful to check that a certificate
2243 request can be signed by that key (self-signing).
2246 *) Make it possible to have multiple active certificates with the same
2247 subject in the CA index file. This is done only if the keyword
2248 'unique_subject' is set to 'no' in the main CA section (default
2249 if 'CA_default') of the configuration file. The value is saved
2250 with the database itself in a separate index attribute file,
2251 named like the index file with '.attr' appended to the name.
2254 *) Generate muti valued AVAs using '+' notation in config files for
2258 *) Support for nameConstraints certificate extension.
2261 *) Support for policyConstraints certificate extension.
2264 *) Support for policyMappings certificate extension.
2267 *) Make sure the default DSA_METHOD implementation only uses its
2268 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
2269 and change its own handlers to be NULL so as to remove unnecessary
2270 indirection. This lets alternative implementations fallback to the
2271 default implementation more easily.
2274 *) Support for directoryName in GeneralName related extensions
2278 *) Make it possible to link applications using Makefile.shared.
2279 Make that possible even when linking against static libraries!
2282 *) Support for single pass processing for S/MIME signing. This now
2283 means that S/MIME signing can be done from a pipe, in addition
2284 cleartext signing (multipart/signed type) is effectively streaming
2285 and the signed data does not need to be all held in memory.
2287 This is done with a new flag PKCS7_STREAM. When this flag is set
2288 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
2289 is done after the data is output (and digests calculated) in
2290 SMIME_write_PKCS7().
2293 *) Add full support for -rpath/-R, both in shared libraries and
2294 applications, at least on the platforms where it's known how
2298 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
2299 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
2300 will now compute a table of multiples of the generator that
2301 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
2302 faster (notably in the case of a single point multiplication,
2303 scalar * generator).
2304 [Nils Larsch, Bodo Moeller]
2306 *) IPv6 support for certificate extensions. The various extensions
2307 which use the IP:a.b.c.d can now take IPv6 addresses using the
2308 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
2312 *) Added an ENGINE that implements RSA by performing private key
2313 exponentiations with the GMP library. The conversions to and from
2314 GMP's mpz_t format aren't optimised nor are any montgomery forms
2315 cached, and on x86 it appears OpenSSL's own performance has caught up.
2316 However there are likely to be other architectures where GMP could
2317 provide a boost. This ENGINE is not built in by default, but it can be
2318 specified at Configure time and should be accompanied by the necessary
2319 linker additions, eg;
2320 ./config -DOPENSSL_USE_GMP -lgmp
2323 *) "openssl engine" will not display ENGINE/DSO load failure errors when
2324 testing availability of engines with "-t" - the old behaviour is
2325 produced by increasing the feature's verbosity with "-tt".
2328 *) ECDSA routines: under certain error conditions uninitialized BN objects
2329 could be freed. Solution: make sure initialization is performed early
2330 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
2334 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
2335 and DH_METHOD (eg. by ENGINE implementations) to override the normal
2336 software implementations. For DSA and DH, parameter generation can
2337 also be overriden by providing the appropriate method callbacks.
2340 *) Change the "progress" mechanism used in key-generation and
2341 primality testing to functions that take a new BN_GENCB pointer in
2342 place of callback/argument pairs. The new API functions have "_ex"
2343 postfixes and the older functions are reimplemented as wrappers for
2344 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
2345 declarations of the old functions to help (graceful) attempts to
2346 migrate to the new functions. Also, the new key-generation API
2347 functions operate on a caller-supplied key-structure and return
2348 success/failure rather than returning a key or NULL - this is to
2349 help make "keygen" another member function of RSA_METHOD etc.
2351 Example for using the new callback interface:
2353 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
2357 BN_GENCB_set(&my_cb, my_callback, my_arg);
2359 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
2360 /* For the meaning of a, b in calls to my_callback(), see the
2361 * documentation of the function that calls the callback.
2362 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
2363 * my_callback should return 1 if it wants BN_is_prime_ex()
2364 * to continue, or 0 to stop.
2369 *) Change the ZLIB compression method to be stateful, and make it
2370 available to TLS with the number defined in
2371 draft-ietf-tls-compression-04.txt.
2374 *) Add the ASN.1 structures and functions for CertificatePair, which
2375 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
2377 CertificatePair ::= SEQUENCE {
2378 forward [0] Certificate OPTIONAL,
2379 reverse [1] Certificate OPTIONAL,
2380 -- at least one of the pair shall be present -- }
2382 Also implement the PEM functions to read and write certificate
2383 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
2385 This needed to be defined, mostly for the sake of the LDAP
2386 attribute crossCertificatePair, but may prove useful elsewhere as
2390 *) Make it possible to inhibit symlinking of shared libraries in
2391 Makefile.shared, for Cygwin's sake.
2394 *) Extend the BIGNUM API by creating a function
2395 void BN_set_negative(BIGNUM *a, int neg);
2396 and a macro that behave like
2397 int BN_is_negative(const BIGNUM *a);
2399 to avoid the need to access 'a->neg' directly in applications.
2402 *) Implement fast modular reduction for pseudo-Mersenne primes
2403 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
2404 EC_GROUP_new_curve_GFp() will now automatically use this
2406 [Nils Larsch <nla@trustcenter.de>]
2408 *) Add new lock type (CRYPTO_LOCK_BN).
2411 *) Change the ENGINE framework to automatically load engines
2412 dynamically from specific directories unless they could be
2413 found to already be built in or loaded. Move all the
2414 current engines except for the cryptodev one to a new
2416 The engines in engines/ are built as shared libraries if
2417 the "shared" options was given to ./Configure or ./config.
2418 Otherwise, they are inserted in libcrypto.a.
2419 /usr/local/ssl/engines is the default directory for dynamic
2420 engines, but that can be overriden at configure time through
2421 the usual use of --prefix and/or --openssldir, and at run
2422 time with the environment variable OPENSSL_ENGINES.
2423 [Geoff Thorpe and Richard Levitte]
2425 *) Add Makefile.shared, a helper makefile to build shared
2426 libraries. Addapt Makefile.org.
2429 *) Add version info to Win32 DLLs.
2430 [Peter 'Luna' Runestig" <peter@runestig.com>]
2432 *) Add new 'medium level' PKCS#12 API. Certificates and keys
2433 can be added using this API to created arbitrary PKCS#12
2434 files while avoiding the low level API.
2436 New options to PKCS12_create(), key or cert can be NULL and
2437 will then be omitted from the output file. The encryption
2438 algorithm NIDs can be set to -1 for no encryption, the mac
2439 iteration count can be set to 0 to omit the mac.
2441 Enhance pkcs12 utility by making the -nokeys and -nocerts
2442 options work when creating a PKCS#12 file. New option -nomac
2443 to omit the mac, NONE can be set for an encryption algorithm.
2444 New code is modified to use the enhanced PKCS12_create()
2445 instead of the low level API.
2448 *) Extend ASN1 encoder to support indefinite length constructed
2449 encoding. This can output sequences tags and octet strings in
2450 this form. Modify pk7_asn1.c to support indefinite length
2451 encoding. This is experimental and needs additional code to
2452 be useful, such as an ASN1 bio and some enhanced streaming
2455 Extend template encode functionality so that tagging is passed
2456 down to the template encoder.
2459 *) Let 'openssl req' fail if an argument to '-newkey' is not
2460 recognized instead of using RSA as a default.
2463 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
2464 As these are not official, they are not included in "ALL";
2465 the "ECCdraft" ciphersuite group alias can be used to select them.
2466 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
2468 *) Add ECDH engine support.
2469 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
2471 *) Add ECDH in new directory crypto/ecdh/.
2472 [Douglas Stebila (Sun Microsystems Laboratories)]
2474 *) Let BN_rand_range() abort with an error after 100 iterations
2475 without success (which indicates a broken PRNG).
2478 *) Change BN_mod_sqrt() so that it verifies that the input value
2479 is really the square of the return value. (Previously,
2480 BN_mod_sqrt would show GIGO behaviour.)
2483 *) Add named elliptic curves over binary fields from X9.62, SECG,
2484 and WAP/WTLS; add OIDs that were still missing.
2486 [Sheueling Chang Shantz and Douglas Stebila
2487 (Sun Microsystems Laboratories)]
2489 *) Extend the EC library for elliptic curves over binary fields
2490 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
2493 EC_GF2m_simple_method
2497 EC_GROUP_new_curve_GF2m
2498 EC_GROUP_set_curve_GF2m
2499 EC_GROUP_get_curve_GF2m
2500 EC_POINT_set_affine_coordinates_GF2m
2501 EC_POINT_get_affine_coordinates_GF2m
2502 EC_POINT_set_compressed_coordinates_GF2m
2504 Point compression for binary fields is disabled by default for
2505 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
2508 As binary polynomials are represented as BIGNUMs, various members
2509 of the EC_GROUP and EC_POINT data structures can be shared
2510 between the implementations for prime fields and binary fields;
2511 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
2512 are essentially identical to their ..._GFp counterparts.
2513 (For simplicity, the '..._GFp' prefix has been dropped from
2514 various internal method names.)
2516 An internal 'field_div' method (similar to 'field_mul' and
2517 'field_sqr') has been added; this is used only for binary fields.
2519 [Sheueling Chang Shantz and Douglas Stebila
2520 (Sun Microsystems Laboratories)]
2522 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
2523 through methods ('mul', 'precompute_mult').
2525 The generic implementations (now internally called 'ec_wNAF_mul'
2526 and 'ec_wNAF_precomputed_mult') remain the default if these
2527 methods are undefined.
2529 [Sheueling Chang Shantz and Douglas Stebila
2530 (Sun Microsystems Laboratories)]
2532 *) New function EC_GROUP_get_degree, which is defined through
2533 EC_METHOD. For curves over prime fields, this returns the bit
2534 length of the modulus.
2536 [Sheueling Chang Shantz and Douglas Stebila
2537 (Sun Microsystems Laboratories)]
2539 *) New functions EC_GROUP_dup, EC_POINT_dup.
2540 (These simply call ..._new and ..._copy).
2542 [Sheueling Chang Shantz and Douglas Stebila
2543 (Sun Microsystems Laboratories)]
2545 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
2546 Polynomials are represented as BIGNUMs (where the sign bit is not
2547 used) in the following functions [macros]:
2550 BN_GF2m_sub [= BN_GF2m_add]
2551 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
2552 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
2553 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
2555 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
2556 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
2557 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
2558 BN_GF2m_cmp [= BN_ucmp]
2560 (Note that only the 'mod' functions are actually for fields GF(2^m).
2561 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
2563 For some functions, an the irreducible polynomial defining a
2564 field can be given as an 'unsigned int[]' with strictly
2565 decreasing elements giving the indices of those bits that are set;
2566 i.e., p[] represents the polynomial
2567 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
2569 p[0] > p[1] > ... > p[k] = 0.
2570 This applies to the following functions:
2575 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
2576 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
2578 BN_GF2m_mod_sqrt_arr
2579 BN_GF2m_mod_solve_quad_arr
2583 Conversion can be performed by the following functions:
2588 bntest.c has additional tests for binary polynomial arithmetic.
2590 Two implementations for BN_GF2m_mod_div() are available.
2591 The default algorithm simply uses BN_GF2m_mod_inv() and
2592 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
2593 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
2594 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
2596 [Sheueling Chang Shantz and Douglas Stebila
2597 (Sun Microsystems Laboratories)]
2599 *) Add new error code 'ERR_R_DISABLED' that can be used when some
2600 functionality is disabled at compile-time.
2601 [Douglas Stebila <douglas.stebila@sun.com>]
2603 *) Change default behaviour of 'openssl asn1parse' so that more
2604 information is visible when viewing, e.g., a certificate:
2606 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
2607 mode the content of non-printable OCTET STRINGs is output in a
2608 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
2609 avoid the appearance of a printable string.
2610 [Nils Larsch <nla@trustcenter.de>]
2612 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
2614 EC_GROUP_set_asn1_flag()
2615 EC_GROUP_get_asn1_flag()
2616 EC_GROUP_set_point_conversion_form()
2617 EC_GROUP_get_point_conversion_form()
2618 These control ASN1 encoding details:
2619 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
2620 has been set to OPENSSL_EC_NAMED_CURVE.
2621 - Points are encoded in uncompressed form by default; options for
2622 asn1_for are as for point2oct, namely
2623 POINT_CONVERSION_COMPRESSED
2624 POINT_CONVERSION_UNCOMPRESSED
2625 POINT_CONVERSION_HYBRID
2627 Also add 'seed' and 'seed_len' members to EC_GROUP with access
2630 EC_GROUP_get0_seed()
2631 EC_GROUP_get_seed_len()
2632 This is used only for ASN1 purposes (so far).
2633 [Nils Larsch <nla@trustcenter.de>]
2635 *) Add 'field_type' member to EC_METHOD, which holds the NID
2636 of the appropriate field type OID. The new function
2637 EC_METHOD_get_field_type() returns this value.
2638 [Nils Larsch <nla@trustcenter.de>]
2643 EC_POINT_point2hex()
2644 EC_POINT_hex2point()
2645 providing useful interfaces to EC_POINT_point2oct() and
2646 EC_POINT_oct2point().
2647 [Nils Larsch <nla@trustcenter.de>]
2649 *) Change internals of the EC library so that the functions
2650 EC_GROUP_set_generator()
2651 EC_GROUP_get_generator()
2652 EC_GROUP_get_order()
2653 EC_GROUP_get_cofactor()
2654 are implemented directly in crypto/ec/ec_lib.c and not dispatched
2655 to methods, which would lead to unnecessary code duplication when
2656 adding different types of curves.
2657 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
2659 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
2660 arithmetic, and such that modified wNAFs are generated
2661 (which avoid length expansion in many cases).
2664 *) Add a function EC_GROUP_check_discriminant() (defined via
2665 EC_METHOD) that verifies that the curve discriminant is non-zero.
2667 Add a function EC_GROUP_check() that makes some sanity tests
2668 on a EC_GROUP, its generator and order. This includes
2669 EC_GROUP_check_discriminant().
2670 [Nils Larsch <nla@trustcenter.de>]
2672 *) Add ECDSA in new directory crypto/ecdsa/.
2674 Add applications 'openssl ecparam' and 'openssl ecdsa'
2675 (these are based on 'openssl dsaparam' and 'openssl dsa').
2677 ECDSA support is also included in various other files across the
2678 library. Most notably,
2679 - 'openssl req' now has a '-newkey ecdsa:file' option;
2680 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
2681 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
2682 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
2683 them suitable for ECDSA where domain parameters must be
2684 extracted before the specific public key;
2685 - ECDSA engine support has been added.
2686 [Nils Larsch <nla@trustcenter.de>]
2688 *) Include some named elliptic curves, and add OIDs from X9.62,
2689 SECG, and WAP/WTLS. Each curve can be obtained from the new
2691 EC_GROUP_new_by_curve_name(),
2692 and the list of available named curves can be obtained with
2693 EC_get_builtin_curves().
2694 Also add a 'curve_name' member to EC_GROUP objects, which can be
2696 EC_GROUP_set_curve_name()
2697 EC_GROUP_get_curve_name()
2698 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
2700 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2701 was actually never needed) and in BN_mul(). The removal in BN_mul()
2702 required a small change in bn_mul_part_recursive() and the addition
2703 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2704 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2705 bn_sub_words() and bn_add_words() except they take arrays with
2709 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
2711 *) Cleanse PEM buffers before freeing them since they may contain
2713 [Benjamin Bennett <ben@psc.edu>]
2715 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2716 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2717 authentication-only ciphersuites.
2720 *) Since AES128 and AES256 share a single mask bit in the logic of
2721 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2722 kludge to work properly if AES128 is available and AES256 isn't.
2725 *) Expand security boundary to match 1.1.1 module.
2728 *) Remove redundant features: hash file source, editing of test vectors
2729 modify fipsld to use external fips_premain.c signature.
2732 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
2733 run algorithm test programs.
2736 *) Make algorithm test programs more tolerant of whitespace.
2739 *) Have SSL/TLS server implementation tolerate "mismatched" record
2740 protocol version while receiving ClientHello even if the
2741 ClientHello is fragmented. (The server can't insist on the
2742 particular protocol version it has chosen before the ServerHello
2743 message has informed the client about his choice.)
2746 *) Load error codes if they are not already present instead of using a
2747 static variable. This allows them to be cleanly unloaded and reloaded.
2750 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
2752 *) Introduce limits to prevent malicious keys being able to
2753 cause a denial of service. (CVE-2006-2940)
2754 [Steve Henson, Bodo Moeller]
2756 *) Fix ASN.1 parsing of certain invalid structures that can result
2757 in a denial of service. (CVE-2006-2937) [Steve Henson]
2759 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2760 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2762 *) Fix SSL client code which could crash if connecting to a
2763 malicious SSLv2 server. (CVE-2006-4343)
2764 [Tavis Ormandy and Will Drewry, Google Security Team]
2766 *) Change ciphersuite string processing so that an explicit
2767 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
2768 will no longer include "AES128-SHA"), and any other similar
2769 ciphersuite (same bitmap) from *other* protocol versions (so that
2770 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
2771 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
2772 changes from 0.9.8b and 0.9.8d.
2775 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
2777 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2778 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2780 *) Change the Unix randomness entropy gathering to use poll() when
2781 possible instead of select(), since the latter has some
2782 undesirable limitations.
2783 [Darryl Miles via Richard Levitte and Bodo Moeller]
2785 *) Disable rogue ciphersuites:
2787 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2788 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2789 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2791 The latter two were purportedly from
2792 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2795 Also deactive the remaining ciphersuites from
2796 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2797 unofficial, and the ID has long expired.
2800 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2801 dual-core machines) and other potential thread-safety issues.
2804 Changes between 0.9.7i and 0.9.7j [04 May 2006]
2806 *) Adapt fipsld and the build system to link against the validated FIPS
2807 module in FIPS mode.
2810 *) Fixes for VC++ 2005 build under Windows.
2813 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
2814 from a Windows bash shell such as MSYS. It is autodetected from the
2815 "config" script when run from a VC++ environment. Modify standard VC++
2816 build to use fipscanister.o from the GNU make build.
2819 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
2821 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
2822 The value now differs depending on if you build for FIPS or not.
2823 BEWARE! A program linked with a shared FIPSed libcrypto can't be
2824 safely run with a non-FIPSed libcrypto, as it may crash because of
2825 the difference induced by this change.
2828 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
2830 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2831 (part of SSL_OP_ALL). This option used to disable the
2832 countermeasure against man-in-the-middle protocol-version
2833 rollback in the SSL 2.0 server implementation, which is a bad
2834 idea. (CVE-2005-2969)
2836 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2837 for Information Security, National Institute of Advanced Industrial
2838 Science and Technology [AIST], Japan)]
2840 *) Minimal support for X9.31 signatures and PSS padding modes. This is
2841 mainly for FIPS compliance and not fully integrated at this stage.
2844 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
2845 the exponentiation using a fixed-length exponent. (Otherwise,
2846 the information leaked through timing could expose the secret key
2847 after many signatures; cf. Bleichenbacher's attack on DSA with
2851 *) Make a new fixed-window mod_exp implementation the default for
2852 RSA, DSA, and DH private-key operations so that the sequence of
2853 squares and multiplies and the memory access pattern are
2854 independent of the particular secret key. This will mitigate
2855 cache-timing and potential related attacks.
2857 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
2858 and this is automatically used by BN_mod_exp_mont() if the new flag
2859 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
2860 will use this BN flag for private exponents unless the flag
2861 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
2862 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
2864 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
2866 *) Change the client implementation for SSLv23_method() and
2867 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
2868 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
2869 (Previously, the SSL 2.0 backwards compatible Client Hello
2870 message format would be used even with SSL_OP_NO_SSLv2.)
2873 *) Add support for smime-type MIME parameter in S/MIME messages which some
2877 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
2878 a threadsafe manner. Modify rsa code to use new function and add calls
2879 to dsa and dh code (which had race conditions before).
2882 *) Include the fixed error library code in the C error file definitions
2883 instead of fixing them up at runtime. This keeps the error code
2884 structures constant.
2887 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
2889 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
2892 *) Fixes for newer kerberos headers. NB: the casts are needed because
2893 the 'length' field is signed on one version and unsigned on another
2894 with no (?) obvious way to tell the difference, without these VC++
2895 complains. Also the "definition" of FAR (blank) is no longer included
2896 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
2897 some needed definitions.
2900 *) Undo Cygwin change.
2903 *) Added support for proxy certificates according to RFC 3820.
2904 Because they may be a security thread to unaware applications,
2905 they must be explicitely allowed in run-time. See
2906 docs/HOWTO/proxy_certificates.txt for further information.
2909 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
2911 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
2912 server and client random values. Previously
2913 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
2914 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
2916 This change has negligible security impact because:
2918 1. Server and client random values still have 24 bytes of pseudo random
2921 2. Server and client random values are sent in the clear in the initial
2924 3. The master secret is derived using the premaster secret (48 bytes in
2925 size for static RSA ciphersuites) as well as client server and random
2928 The OpenSSL team would like to thank the UK NISCC for bringing this issue
2931 [Stephen Henson, reported by UK NISCC]
2933 *) Use Windows randomness collection on Cygwin.
2936 *) Fix hang in EGD/PRNGD query when communication socket is closed
2937 prematurely by EGD/PRNGD.
2938 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
2940 *) Prompt for pass phrases when appropriate for PKCS12 input format.
2943 *) Back-port of selected performance improvements from development
2944 branch, as well as improved support for PowerPC platforms.
2947 *) Add lots of checks for memory allocation failure, error codes to indicate
2948 failure and freeing up memory if a failure occurs.
2949 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
2951 *) Add new -passin argument to dgst.
2954 *) Perform some character comparisons of different types in X509_NAME_cmp:
2955 this is needed for some certificates that reencode DNs into UTF8Strings
2956 (in violation of RFC3280) and can't or wont issue name rollover
2960 *) Make an explicit check during certificate validation to see that
2961 the CA setting in each certificate on the chain is correct. As a
2962 side effect always do the following basic checks on extensions,
2963 not just when there's an associated purpose to the check:
2965 - if there is an unhandled critical extension (unless the user
2966 has chosen to ignore this fault)
2967 - if the path length has been exceeded (if one is set at all)
2968 - that certain extensions fit the associated purpose (if one has
2972 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
2974 *) Avoid a race condition when CRLs are checked in a multi threaded
2975 environment. This would happen due to the reordering of the revoked
2976 entries during signature checking and serial number lookup. Now the
2977 encoding is cached and the serial number sort performed under a lock.
2978 Add new STACK function sk_is_sorted().
2981 *) Add Delta CRL to the extension code.
2984 *) Various fixes to s3_pkt.c so alerts are sent properly.
2985 [David Holmes <d.holmes@f5.com>]
2987 *) Reduce the chances of duplicate issuer name and serial numbers (in
2988 violation of RFC3280) using the OpenSSL certificate creation utilities.
2989 This is done by creating a random 64 bit value for the initial serial
2990 number when a serial number file is created or when a self signed
2991 certificate is created using 'openssl req -x509'. The initial serial
2992 number file is created using 'openssl x509 -next_serial' in CA.pl
2993 rather than being initialized to 1.
2996 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
2998 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
2999 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3000 [Joe Orton, Steve Henson]
3002 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3004 [Joe Orton, Steve Henson]
3006 *) Make it possible to have multiple active certificates with the same
3007 subject in the CA index file. This is done only if the keyword
3008 'unique_subject' is set to 'no' in the main CA section (default
3009 if 'CA_default') of the configuration file. The value is saved
3010 with the database itself in a separate index attribute file,
3011 named like the index file with '.attr' appended to the name.
3014 *) X509 verify fixes. Disable broken certificate workarounds when
3015 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3016 keyUsage extension present. Don't accept CRLs with unhandled critical
3017 extensions: since verify currently doesn't process CRL extensions this
3018 rejects a CRL with *any* critical extensions. Add new verify error codes
3022 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
3023 A clarification of RFC2560 will require the use of OCTET STRINGs and
3024 some implementations cannot handle the current raw format. Since OpenSSL
3025 copies and compares OCSP nonces as opaque blobs without any attempt at
3026 parsing them this should not create any compatibility issues.
3029 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
3030 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
3031 this HMAC (and other) operations are several times slower than OpenSSL
3035 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
3036 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3038 *) Use the correct content when signing type "other".
3041 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
3043 *) Fix various bugs revealed by running the NISCC test suite:
3045 Stop out of bounds reads in the ASN1 code when presented with
3046 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3048 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
3050 If verify callback ignores invalid public key errors don't try to check
3051 certificate signature with the NULL public key.
3055 *) New -ignore_err option in ocsp application to stop the server
3056 exiting on the first error in a request.
3059 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3060 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3064 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3065 extra data after the compression methods not only for TLS 1.0
3066 but also for SSL 3.0 (as required by the specification).
3067 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3069 *) Change X509_certificate_type() to mark the key as exported/exportable
3070 when it's 512 *bits* long, not 512 bytes.
3073 *) Change AES_cbc_encrypt() so it outputs exact multiple of
3074 blocks during encryption.
3077 *) Various fixes to base64 BIO and non blocking I/O. On write
3078 flushes were not handled properly if the BIO retried. On read
3079 data was not being buffered properly and had various logic bugs.
3080 This also affects blocking I/O when the data being decoded is a
3084 *) Various S/MIME bugfixes and compatibility changes:
3085 output correct application/pkcs7 MIME type if
3086 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
3087 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
3088 of files as .eml work). Correctly handle very long lines in MIME
3092 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
3094 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3095 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3096 a protocol version number mismatch like a decryption error
3097 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3100 *) Turn on RSA blinding by default in the default implementation
3101 to avoid a timing attack. Applications that don't want it can call
3102 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3103 They would be ill-advised to do so in most cases.
3104 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3106 *) Change RSA blinding code so that it works when the PRNG is not
3107 seeded (in this case, the secret RSA exponent is abused as
3108 an unpredictable seed -- if it is not unpredictable, there
3109 is no point in blinding anyway). Make RSA blinding thread-safe
3110 by remembering the creator's thread ID in rsa->blinding and
3111 having all other threads use local one-time blinding factors
3112 (this requires more computation than sharing rsa->blinding, but
3113 avoids excessive locking; and if an RSA object is not shared
3114 between threads, blinding will still be very fast).
3117 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
3118 ENGINE as defaults for all supported algorithms irrespective of
3119 the 'flags' parameter. 'flags' is now honoured, so applications
3120 should make sure they are passing it correctly.
3123 *) Target "mingw" now allows native Windows code to be generated in
3124 the Cygwin environment as well as with the MinGW compiler.
3127 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
3129 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3130 via timing by performing a MAC computation even if incorrrect
3131 block cipher padding has been found. This is a countermeasure
3132 against active attacks where the attacker has to distinguish
3133 between bad padding and a MAC verification error. (CVE-2003-0078)
3135 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3136 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3137 Martin Vuagnoux (EPFL, Ilion)]
3139 *) Make the no-err option work as intended. The intention with no-err
3140 is not to have the whole error stack handling routines removed from
3141 libcrypto, it's only intended to remove all the function name and
3142 reason texts, thereby removing some of the footprint that may not
3143 be interesting if those errors aren't displayed anyway.
3145 NOTE: it's still possible for any application or module to have it's
3146 own set of error texts inserted. The routines are there, just not
3147 used by default when no-err is given.
3150 *) Add support for FreeBSD on IA64.
3151 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
3153 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
3154 Kerberos function mit_des_cbc_cksum(). Before this change,
3155 the value returned by DES_cbc_cksum() was like the one from
3156 mit_des_cbc_cksum(), except the bytes were swapped.
3157 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
3159 *) Allow an application to disable the automatic SSL chain building.
3160 Before this a rather primitive chain build was always performed in
3161 ssl3_output_cert_chain(): an application had no way to send the
3162 correct chain if the automatic operation produced an incorrect result.
3164 Now the chain builder is disabled if either:
3166 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
3168 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
3170 The reasoning behind this is that an application would not want the
3171 auto chain building to take place if extra chain certificates are
3172 present and it might also want a means of sending no additional
3173 certificates (for example the chain has two certificates and the
3177 *) Add the possibility to build without the ENGINE framework.
3178 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3180 *) Under Win32 gmtime() can return NULL: check return value in
3181 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
3184 *) DSA routines: under certain error conditions uninitialized BN objects
3185 could be freed. Solution: make sure initialization is performed early
3186 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
3187 Nils Larsch <nla@trustcenter.de> via PR#459)
3190 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
3191 checked on reconnect on the client side, therefore session resumption
3192 could still fail with a "ssl session id is different" error. This
3193 behaviour is masked when SSL_OP_ALL is used due to
3194 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
3195 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3196 followup to PR #377.
3199 *) IA-32 assembler support enhancements: unified ELF targets, support
3200 for SCO/Caldera platforms, fix for Cygwin shared build.
3203 *) Add support for FreeBSD on sparc64. As a consequence, support for
3204 FreeBSD on non-x86 processors is separate from x86 processors on
3205 the config script, much like the NetBSD support.
3206 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
3208 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
3210 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
3213 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
3214 code (06) was taken as the first octet of the session ID and the last
3215 octet was ignored consequently. As a result SSLv2 client side session
3216 caching could not have worked due to the session ID mismatch between
3218 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3222 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
3223 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
3227 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
3228 seems that in spite of existing for more than a year, many application
3229 author have done nothing to provide the necessary callbacks, which
3230 means that this particular engine will not work properly anywhere.
3231 This is a very unfortunate situation which forces us, in the name
3232 of usability, to give the hw_ncipher.c a static lock, which is part
3234 NOTE: This is for the 0.9.7 series ONLY. This hack will never
3235 appear in 0.9.8 or later. We EXPECT application authors to have
3236 dealt properly with this when 0.9.8 is released (unless we actually
3237 make such changes in the libcrypto locking code that changes will
3238 have to be made anyway).
3241 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
3242 octets have been read, EOF or an error occurs. Without this change
3243 some truncated ASN1 structures will not produce an error.
3246 *) Disable Heimdal support, since it hasn't been fully implemented.
3247 Still give the possibility to force the use of Heimdal, but with
3248 warnings and a request that patches get sent to openssl-dev.
3251 *) Add the VC-CE target, introduce the WINCE sysname, and add
3252 INSTALL.WCE and appropriate conditionals to make it build.
3253 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3255 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
3256 cygssl-x.y.z.dll, where x, y and z are the major, minor and
3257 edit numbers of the version.
3258 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3260 *) Introduce safe string copy and catenation functions
3261 (BUF_strlcpy() and BUF_strlcat()).
3262 [Ben Laurie (CHATS) and Richard Levitte]
3264 *) Avoid using fixed-size buffers for one-line DNs.
3265 [Ben Laurie (CHATS)]
3267 *) Add BUF_MEM_grow_clean() to avoid information leakage when
3268 resizing buffers containing secrets, and use where appropriate.
3269 [Ben Laurie (CHATS)]
3271 *) Avoid using fixed size buffers for configuration file location.
3272 [Ben Laurie (CHATS)]
3274 *) Avoid filename truncation for various CA files.
3275 [Ben Laurie (CHATS)]
3277 *) Use sizeof in preference to magic numbers.
3278 [Ben Laurie (CHATS)]
3280 *) Avoid filename truncation in cert requests.
3281 [Ben Laurie (CHATS)]
3283 *) Add assertions to check for (supposedly impossible) buffer
3285 [Ben Laurie (CHATS)]
3287 *) Don't cache truncated DNS entries in the local cache (this could
3288 potentially lead to a spoofing attack).
3289 [Ben Laurie (CHATS)]
3291 *) Fix various buffers to be large enough for hex/decimal
3292 representations in a platform independent manner.
3293 [Ben Laurie (CHATS)]
3295 *) Add CRYPTO_realloc_clean() to avoid information leakage when
3296 resizing buffers containing secrets, and use where appropriate.
3297 [Ben Laurie (CHATS)]
3299 *) Add BIO_indent() to avoid much slightly worrying code to do
3301 [Ben Laurie (CHATS)]
3303 *) Convert sprintf()/BIO_puts() to BIO_printf().
3304 [Ben Laurie (CHATS)]
3306 *) buffer_gets() could terminate with the buffer only half
3308 [Ben Laurie (CHATS)]
3310 *) Add assertions to prevent user-supplied crypto functions from
3311 overflowing internal buffers by having large block sizes, etc.
3312 [Ben Laurie (CHATS)]
3314 *) New OPENSSL_assert() macro (similar to assert(), but enabled
3316 [Ben Laurie (CHATS)]
3318 *) Eliminate unused copy of key in RC4.
3319 [Ben Laurie (CHATS)]
3321 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
3322 [Ben Laurie (CHATS)]
3324 *) Fix off-by-one error in EGD path.
3325 [Ben Laurie (CHATS)]
3327 *) If RANDFILE path is too long, ignore instead of truncating.
3328 [Ben Laurie (CHATS)]
3330 *) Eliminate unused and incorrectly sized X.509 structure
3332 [Ben Laurie (CHATS)]
3334 *) Eliminate unused and dangerous function knumber().
3335 [Ben Laurie (CHATS)]
3337 *) Eliminate unused and dangerous structure, KSSL_ERR.
3338 [Ben Laurie (CHATS)]
3340 *) Protect against overlong session ID context length in an encoded
3341 session object. Since these are local, this does not appear to be
3343 [Ben Laurie (CHATS)]
3345 *) Change from security patch (see 0.9.6e below) that did not affect
3346 the 0.9.6 release series:
3348 Remote buffer overflow in SSL3 protocol - an attacker could
3349 supply an oversized master key in Kerberos-enabled versions.
3351 [Ben Laurie (CHATS)]
3353 *) Change the SSL kerb5 codes to match RFC 2712.
3356 *) Make -nameopt work fully for req and add -reqopt switch.
3357 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
3359 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
3360 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
3362 *) Make sure tests can be performed even if the corresponding algorithms
3363 have been removed entirely. This was also the last step to make
3364 OpenSSL compilable with DJGPP under all reasonable conditions.
3365 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
3367 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
3368 to allow version independent disabling of normally unselected ciphers,
3369 which may be activated as a side-effect of selecting a single cipher.
3371 (E.g., cipher list string "RSA" enables ciphersuites that are left
3372 out of "ALL" because they do not provide symmetric encryption.
3373 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
3374 [Lutz Jaenicke, Bodo Moeller]
3376 *) Add appropriate support for separate platform-dependent build
3377 directories. The recommended way to make a platform-dependent
3378 build directory is the following (tested on Linux), maybe with
3381 # Place yourself outside of the OpenSSL source tree. In
3382 # this example, the environment variable OPENSSL_SOURCE
3383 # is assumed to contain the absolute OpenSSL source directory.
3384 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
3385 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
3386 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
3387 mkdir -p `dirname $F`
3388 ln -s $OPENSSL_SOURCE/$F $F
3391 To be absolutely sure not to disturb the source tree, a "make clean"
3392 is a good thing. If it isn't successfull, don't worry about it,
3393 it probably means the source directory is very clean.
3396 *) Make sure any ENGINE control commands make local copies of string
3397 pointers passed to them whenever necessary. Otherwise it is possible
3398 the caller may have overwritten (or deallocated) the original string
3399 data when a later ENGINE operation tries to use the stored values.
3400 [Götz Babin-Ebell <babinebell@trustcenter.de>]
3402 *) Improve diagnostics in file reading and command-line digests.
3403 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
3405 *) Add AES modes CFB and OFB to the object database. Correct an
3406 error in AES-CFB decryption.
3409 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
3410 allows existing EVP_CIPHER_CTX structures to be reused after
3411 calling EVP_*Final(). This behaviour is used by encryption
3412 BIOs and some applications. This has the side effect that
3413 applications must explicitly clean up cipher contexts with
3414 EVP_CIPHER_CTX_cleanup() or they will leak memory.
3417 *) Check the values of dna and dnb in bn_mul_recursive before calling
3418 bn_mul_comba (a non zero value means the a or b arrays do not contain
3419 n2 elements) and fallback to bn_mul_normal if either is not zero.
3422 *) Fix escaping of non-ASCII characters when using the -subj option
3423 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
3426 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
3427 form for "surname", serialNumber has no short form.
3428 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
3429 therefore remove "mail" short name for "internet 7".
3430 The OID for unique identifiers in X509 certificates is
3431 x500UniqueIdentifier, not uniqueIdentifier.
3432 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
3435 *) Add an "init" command to the ENGINE config module and auto initialize
3436 ENGINEs. Without any "init" command the ENGINE will be initialized
3437 after all ctrl commands have been executed on it. If init=1 the
3438 ENGINE is initailized at that point (ctrls before that point are run
3439 on the uninitialized ENGINE and after on the initialized one). If
3440 init=0 then the ENGINE will not be iniatialized at all.
3443 *) Fix the 'app_verify_callback' interface so that the user-defined
3444 argument is actually passed to the callback: In the
3445 SSL_CTX_set_cert_verify_callback() prototype, the callback
3446 declaration has been changed from
3449 int (*cb)(X509_STORE_CTX *,void *);
3450 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
3451 i=s->ctx->app_verify_callback(&ctx)
3452 has been changed into
3453 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
3455 To update applications using SSL_CTX_set_cert_verify_callback(),
3456 a dummy argument can be added to their callback functions.
3457 [D. K. Smetters <smetters@parc.xerox.com>]
3459 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
3460 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
3462 *) Add and OPENSSL_LOAD_CONF define which will cause
3463 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
3464 This allows older applications to transparently support certain
3465 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
3466 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
3467 load the config file and OPENSSL_add_all_algorithms_conf() which will
3468 always load it have also been added.
3471 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
3472 Adjust NIDs and EVP layer.
3473 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3475 *) Config modules support in openssl utility.
3477 Most commands now load modules from the config file,
3478 though in a few (such as version) this isn't done
3479 because it couldn't be used for anything.
3481 In the case of ca and req the config file used is
3482 the same as the utility itself: that is the -config
3483 command line option can be used to specify an
3487 *) Move default behaviour from OPENSSL_config(). If appname is NULL
3488 use "openssl_conf" if filename is NULL use default openssl config file.
3491 *) Add an argument to OPENSSL_config() to allow the use of an alternative
3492 config section name. Add a new flag to tolerate a missing config file
3493 and move code to CONF_modules_load_file().
3496 *) Support for crypto accelerator cards from Accelerated Encryption
3497 Processing, www.aep.ie. (Use engine 'aep')
3498 The support was copied from 0.9.6c [engine] and adapted/corrected
3499 to work with the new engine framework.
3500 [AEP Inc. and Richard Levitte]
3502 *) Support for SureWare crypto accelerator cards from Baltimore
3503 Technologies. (Use engine 'sureware')
3504 The support was copied from 0.9.6c [engine] and adapted
3505 to work with the new engine framework.
3508 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
3509 make the newer ENGINE framework commands for the CHIL engine work.
3510 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
3512 *) Make it possible to produce shared libraries on ReliantUNIX.
3513 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
3515 *) Add the configuration target debug-linux-ppro.
3516 Make 'openssl rsa' use the general key loading routines
3517 implemented in apps.c, and make those routines able to
3518 handle the key format FORMAT_NETSCAPE and the variant
3520 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3522 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3523 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3525 *) Add -keyform to rsautl, and document -engine.
3526 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
3528 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
3529 BIO_R_NO_SUCH_FILE error code rather than the generic
3530 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
3533 *) Add new functions
3535 ERR_peek_last_error_line
3536 ERR_peek_last_error_line_data.
3537 These are similar to
3540 ERR_peek_error_line_data,
3541 but report on the latest error recorded rather than the first one
3542 still in the error queue.
3543 [Ben Laurie, Bodo Moeller]
3545 *) default_algorithms option in ENGINE config module. This allows things
3547 default_algorithms = ALL
3548 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
3551 *) Prelminary ENGINE config module.
3554 *) New experimental application configuration code.
3557 *) Change the AES code to follow the same name structure as all other
3558 symmetric ciphers, and behave the same way. Move everything to
3559 the directory crypto/aes, thereby obsoleting crypto/rijndael.
3560 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3562 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
3563 [Ben Laurie and Theo de Raadt]
3565 *) Add option to output public keys in req command.
3566 [Massimiliano Pala madwolf@openca.org]
3568 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
3569 (up to about 10% better than before for P-192 and P-224).
3572 *) New functions/macros
3574 SSL_CTX_set_msg_callback(ctx, cb)
3575 SSL_CTX_set_msg_callback_arg(ctx, arg)
3576 SSL_set_msg_callback(ssl, cb)
3577 SSL_set_msg_callback_arg(ssl, arg)
3579 to request calling a callback function
3581 void cb(int write_p, int version, int content_type,
3582 const void *buf, size_t len, SSL *ssl, void *arg)
3584 whenever a protocol message has been completely received
3585 (write_p == 0) or sent (write_p == 1). Here 'version' is the
3586 protocol version according to which the SSL library interprets
3587 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
3588 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
3589 the content type as defined in the SSL 3.0/TLS 1.0 protocol
3590 specification (change_cipher_spec(20), alert(21), handshake(22)).
3591 'buf' and 'len' point to the actual message, 'ssl' to the
3592 SSL object, and 'arg' is the application-defined value set by
3593 SSL[_CTX]_set_msg_callback_arg().
3595 'openssl s_client' and 'openssl s_server' have new '-msg' options
3596 to enable a callback that displays all protocol messages.
3599 *) Change the shared library support so shared libraries are built as
3600 soon as the corresponding static library is finished, and thereby get
3601 openssl and the test programs linked against the shared library.
3602 This still only happens when the keyword "shard" has been given to
3603 the configuration scripts.
3605 NOTE: shared library support is still an experimental thing, and
3606 backward binary compatibility is still not guaranteed.
3607 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
3609 *) Add support for Subject Information Access extension.
3610 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3612 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
3613 additional bytes when new memory had to be allocated, not just
3614 when reusing an existing buffer.
3617 *) New command line and configuration option 'utf8' for the req command.
3618 This allows field values to be specified as UTF8 strings.
3621 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
3622 runs for the former and machine-readable output for the latter.
3625 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
3626 of the e-mail address in the DN (i.e., it will go into a certificate
3627 extension only). The new configuration file option 'email_in_dn = no'
3628 has the same effect.
3629 [Massimiliano Pala madwolf@openca.org]
3631 *) Change all functions with names starting with des_ to be starting
3632 with DES_ instead. Add wrappers that are compatible with libdes,
3633 but are named _ossl_old_des_*. Finally, add macros that map the
3634 des_* symbols to the corresponding _ossl_old_des_* if libdes
3635 compatibility is desired. If OpenSSL 0.9.6c compatibility is
3636 desired, the des_* symbols will be mapped to DES_*, with one
3639 Since we provide two compatibility mappings, the user needs to
3640 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
3641 compatibility is desired. The default (i.e., when that macro
3642 isn't defined) is OpenSSL 0.9.6c compatibility.
3644 There are also macros that enable and disable the support of old
3645 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
3646 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
3647 are defined, the default will apply: to support the old des routines.
3649 In either case, one must include openssl/des.h to get the correct
3650 definitions. Do not try to just include openssl/des_old.h, that
3653 NOTE: This is a major break of an old API into a new one. Software
3654 authors are encouraged to switch to the DES_ style functions. Some
3655 time in the future, des_old.h and the libdes compatibility functions
3656 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
3657 default), and then completely removed.
3660 *) Test for certificates which contain unsupported critical extensions.
3661 If such a certificate is found during a verify operation it is
3662 rejected by default: this behaviour can be overridden by either
3663 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
3664 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
3665 X509_supported_extension() has also been added which returns 1 if a
3666 particular extension is supported.
3669 *) Modify the behaviour of EVP cipher functions in similar way to digests
3670 to retain compatibility with existing code.
3673 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
3674 compatibility with existing code. In particular the 'ctx' parameter does
3675 not have to be to be initialized before the call to EVP_DigestInit() and
3676 it is tidied up after a call to EVP_DigestFinal(). New function
3677 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
3678 EVP_MD_CTX_copy() changed to not require the destination to be
3679 initialized valid and new function EVP_MD_CTX_copy_ex() added which
3680 requires the destination to be valid.
3682 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
3683 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
3686 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
3687 so that complete 'Handshake' protocol structures are kept in memory
3688 instead of overwriting 'msg_type' and 'length' with 'body' data.
3691 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
3692 [Massimo Santin via Richard Levitte]
3694 *) Major restructuring to the underlying ENGINE code. This includes
3695 reduction of linker bloat, separation of pure "ENGINE" manipulation
3696 (initialisation, etc) from functionality dealing with implementations
3697 of specific crypto iterfaces. This change also introduces integrated
3698 support for symmetric ciphers and digest implementations - so ENGINEs
3699 can now accelerate these by providing EVP_CIPHER and EVP_MD
3700 implementations of their own. This is detailed in crypto/engine/README
3701 as it couldn't be adequately described here. However, there are a few
3702 API changes worth noting - some RSA, DSA, DH, and RAND functions that
3703 were changed in the original introduction of ENGINE code have now
3704 reverted back - the hooking from this code to ENGINE is now a good
3705 deal more passive and at run-time, operations deal directly with
3706 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
3707 dereferencing through an ENGINE pointer any more. Also, the ENGINE
3708 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
3709 they were not being used by the framework as there is no concept of a
3710 BIGNUM_METHOD and they could not be generalised to the new
3711 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
3712 ENGINE_cpy() has been removed as it cannot be consistently defined in
3716 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
3719 *) Change mkdef.pl to sort symbols that get the same entry number,
3720 and make sure the automatically generated functions ERR_load_*
3721 become part of libeay.num as well.
3724 *) New function SSL_renegotiate_pending(). This returns true once
3725 renegotiation has been requested (either SSL_renegotiate() call
3726 or HelloRequest/ClientHello receveived from the peer) and becomes
3727 false once a handshake has been completed.
3728 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
3729 sends a HelloRequest, but does not ensure that a handshake takes
3730 place. SSL_renegotiate_pending() is useful for checking if the
3731 client has followed the request.)
3734 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
3735 By default, clients may request session resumption even during
3736 renegotiation (if session ID contexts permit); with this option,
3737 session resumption is possible only in the first handshake.
3739 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
3740 more bits available for options that should not be part of
3741 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
3744 *) Add some demos for certificate and certificate request creation.
3747 *) Make maximum certificate chain size accepted from the peer application
3748 settable (SSL*_get/set_max_cert_list()), as proposed by
3749 "Douglas E. Engert" <deengert@anl.gov>.
3752 *) Add support for shared libraries for Unixware-7
3753 (Boyd Lynn Gerber <gerberb@zenez.com>).
3756 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
3757 be done prior to destruction. Use this to unload error strings from
3758 ENGINEs that load their own error strings. NB: This adds two new API
3759 functions to "get" and "set" this destroy handler in an ENGINE.
3762 *) Alter all existing ENGINE implementations (except "openssl" and
3763 "openbsd") to dynamically instantiate their own error strings. This
3764 makes them more flexible to be built both as statically-linked ENGINEs
3765 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
3766 Also, add stub code to each that makes building them as self-contained
3767 shared-libraries easier (see README.ENGINE).
3770 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
3771 implementations into applications that are completely implemented in
3772 self-contained shared-libraries. The "dynamic" ENGINE exposes control
3773 commands that can be used to configure what shared-library to load and
3774 to control aspects of the way it is handled. Also, made an update to
3775 the README.ENGINE file that brings its information up-to-date and
3776 provides some information and instructions on the "dynamic" ENGINE
3777 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
3780 *) Make it possible to unload ranges of ERR strings with a new
3781 "ERR_unload_strings" function.
3784 *) Add a copy() function to EVP_MD.
3787 *) Make EVP_MD routines take a context pointer instead of just the
3788 md_data void pointer.
3791 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
3792 that the digest can only process a single chunk of data
3793 (typically because it is provided by a piece of
3794 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
3795 is only going to provide a single chunk of data, and hence the
3796 framework needn't accumulate the data for oneshot drivers.
3799 *) As with "ERR", make it possible to replace the underlying "ex_data"
3800 functions. This change also alters the storage and management of global
3801 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
3802 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
3803 index counters. The API functions that use this state have been changed
3804 to take a "class_index" rather than pointers to the class's local STACK
3805 and counter, and there is now an API function to dynamically create new
3806 classes. This centralisation allows us to (a) plug a lot of the
3807 thread-safety problems that existed, and (b) makes it possible to clean
3808 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
3809 such data would previously have always leaked in application code and
3810 workarounds were in place to make the memory debugging turn a blind eye
3811 to it. Application code that doesn't use this new function will still
3812 leak as before, but their memory debugging output will announce it now
3813 rather than letting it slide.
3815 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
3816 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
3817 has a return value to indicate success or failure.
3820 *) Make it possible to replace the underlying "ERR" functions such that the
3821 global state (2 LHASH tables and 2 locks) is only used by the "default"
3822 implementation. This change also adds two functions to "get" and "set"
3823 the implementation prior to it being automatically set the first time
3824 any other ERR function takes place. Ie. an application can call "get",
3825 pass the return value to a module it has just loaded, and that module
3826 can call its own "set" function using that value. This means the
3827 module's "ERR" operations will use (and modify) the error state in the
3828 application and not in its own statically linked copy of OpenSSL code.
3831 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
3832 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
3833 the operation, and provides a more encapsulated way for external code
3834 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
3835 to use these functions rather than manually incrementing the counts.
3837 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
3840 *) Add EVP test program.
3843 *) Add symmetric cipher support to ENGINE. Expect the API to change!
3846 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
3847 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
3848 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
3849 These allow a CRL to be built without having to access X509_CRL fields
3850 directly. Modify 'ca' application to use new functions.
3853 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
3854 bug workarounds. Rollback attack detection is a security feature.
3855 The problem will only arise on OpenSSL servers when TLSv1 is not
3856 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
3857 Software authors not wanting to support TLSv1 will have special reasons
3858 for their choice and can explicitly enable this option.
3859 [Bodo Moeller, Lutz Jaenicke]
3861 *) Rationalise EVP so it can be extended: don't include a union of
3862 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
3863 (similar to those existing for EVP_CIPHER_CTX).
3868 EVP_MD_CTX_init(&md); /* new function call */
3869 EVP_DigestInit(&md, EVP_sha1());
3870 EVP_DigestUpdate(&md, in, len);
3871 EVP_DigestFinal(&md, out, NULL);
3872 EVP_MD_CTX_cleanup(&md); /* new function call */
3876 *) Make DES key schedule conform to the usual scheme, as well as
3877 correcting its structure. This means that calls to DES functions
3878 now have to pass a pointer to a des_key_schedule instead of a
3879 plain des_key_schedule (which was actually always a pointer
3882 des_key_schedule ks;
3884 des_set_key_checked(..., &ks);
3885 des_ncbc_encrypt(..., &ks, ...);
3887 (Note that a later change renames 'des_...' into 'DES_...'.)
3890 *) Initial reduction of linker bloat: the use of some functions, such as
3891 PEM causes large amounts of unused functions to be linked in due to
3892 poor organisation. For example pem_all.c contains every PEM function
3893 which has a knock on effect of linking in large amounts of (unused)
3894 ASN1 code. Grouping together similar functions and splitting unrelated
3895 functions prevents this.
3898 *) Cleanup of EVP macros.
3901 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
3902 correct _ecb suffix.
3905 *) Add initial OCSP responder support to ocsp application. The
3906 revocation information is handled using the text based index
3907 use by the ca application. The responder can either handle
3908 requests generated internally, supplied in files (for example
3909 via a CGI script) or using an internal minimal server.
3912 *) Add configuration choices to get zlib compression for TLS.
3915 *) Changes to Kerberos SSL for RFC 2712 compliance:
3916 1. Implemented real KerberosWrapper, instead of just using
3917 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
3918 2. Implemented optional authenticator field of KerberosWrapper.
3920 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
3921 and authenticator structs; see crypto/krb5/.
3923 Generalized Kerberos calls to support multiple Kerberos libraries.
3924 [Vern Staats <staatsvr@asc.hpc.mil>,
3925 Jeffrey Altman <jaltman@columbia.edu>
3926 via Richard Levitte]
3928 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
3929 already does with RSA. testdsa.h now has 'priv_key/pub_key'
3930 values for each of the key sizes rather than having just
3931 parameters (and 'speed' generating keys each time).
3934 *) Speed up EVP routines.
3937 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
3938 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
3939 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
3940 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
3942 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
3943 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
3944 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
3947 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
3949 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
3952 *) Added the OS2-EMX target.
3953 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
3955 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
3956 to support NCONF routines in extension code. New function CONF_set_nconf()
3957 to allow functions which take an NCONF to also handle the old LHASH
3958 structure: this means that the old CONF compatible routines can be
3959 retained (in particular wrt extensions) without having to duplicate the
3960 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
3963 *) Enhance the general user interface with mechanisms for inner control
3964 and with possibilities to have yes/no kind of prompts.
3967 *) Change all calls to low level digest routines in the library and
3968 applications to use EVP. Add missing calls to HMAC_cleanup() and
3969 don't assume HMAC_CTX can be copied using memcpy().
3970 [Verdon Walker <VWalker@novell.com>, Steve Henson]
3972 *) Add the possibility to control engines through control names but with
3973 arbitrary arguments instead of just a string.
3974 Change the key loaders to take a UI_METHOD instead of a callback
3975 function pointer. NOTE: this breaks binary compatibility with earlier
3976 versions of OpenSSL [engine].
3977 Adapt the nCipher code for these new conditions and add a card insertion
3981 *) Enhance the general user interface with mechanisms to better support
3982 dialog box interfaces, application-defined prompts, the possibility
3983 to use defaults (for example default passwords from somewhere else)
3984 and interrupts/cancellations.
3987 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
3988 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
3991 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
3992 tidy up some unnecessarily weird code in 'sk_new()').
3993 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
3995 *) Change the key loading routines for ENGINEs to use the same kind
3996 callback (pem_password_cb) as all other routines that need this
4000 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4001 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4002 than this minimum value is recommended.
4005 *) New random seeder for OpenVMS, using the system process statistics
4006 that are easily reachable.
4009 *) Windows apparently can't transparently handle global
4010 variables defined in DLLs. Initialisations such as:
4012 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4014 wont compile. This is used by the any applications that need to
4015 declare their own ASN1 modules. This was fixed by adding the option
4016 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4017 needed for static libraries under Win32.
4020 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
4021 setting of purpose and trust fields. New X509_STORE trust and
4022 purpose functions and tidy up setting in other SSL functions.
4025 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
4026 structure. These are inherited by X509_STORE_CTX when it is
4027 initialised. This allows various defaults to be set in the
4028 X509_STORE structure (such as flags for CRL checking and custom
4029 purpose or trust settings) for functions which only use X509_STORE_CTX
4030 internally such as S/MIME.
4032 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
4033 trust settings if they are not set in X509_STORE. This allows X509_STORE
4034 purposes and trust (in S/MIME for example) to override any set by default.
4036 Add command line options for CRL checking to smime, s_client and s_server
4040 *) Initial CRL based revocation checking. If the CRL checking flag(s)
4041 are set then the CRL is looked up in the X509_STORE structure and
4042 its validity and signature checked, then if the certificate is found
4043 in the CRL the verify fails with a revoked error.
4045 Various new CRL related callbacks added to X509_STORE_CTX structure.
4047 Command line options added to 'verify' application to support this.
4049 This needs some additional work, such as being able to handle multiple
4050 CRLs with different times, extension based lookup (rather than just
4051 by subject name) and ultimately more complete V2 CRL extension
4055 *) Add a general user interface API (crypto/ui/). This is designed
4056 to replace things like des_read_password and friends (backward
4057 compatibility functions using this new API are provided).
4058 The purpose is to remove prompting functions from the DES code
4059 section as well as provide for prompting through dialog boxes in
4060 a window system and the like.
4063 *) Add "ex_data" support to ENGINE so implementations can add state at a
4064 per-structure level rather than having to store it globally.
4067 *) Make it possible for ENGINE structures to be copied when retrieved by
4068 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
4069 This causes the "original" ENGINE structure to act like a template,
4070 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
4071 operational state can be localised to each ENGINE structure, despite the
4072 fact they all share the same "methods". New ENGINE structures returned in
4073 this case have no functional references and the return value is the single
4074 structural reference. This matches the single structural reference returned
4075 by ENGINE_by_id() normally, when it is incremented on the pre-existing
4079 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
4080 needs to match any other type at all we need to manually clear the
4084 *) Changes to the "openssl engine" utility to include;
4085 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
4086 about an ENGINE's available control commands.
4087 - executing control commands from command line arguments using the
4088 '-pre' and '-post' switches. '-post' is only used if '-t' is
4089 specified and the ENGINE is successfully initialised. The syntax for
4090 the individual commands are colon-separated, for example;
4091 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
4094 *) New dynamic control command support for ENGINEs. ENGINEs can now
4095 declare their own commands (numbers), names (strings), descriptions,
4096 and input types for run-time discovery by calling applications. A
4097 subset of these commands are implicitly classed as "executable"
4098 depending on their input type, and only these can be invoked through
4099 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
4100 can be based on user input, config files, etc). The distinction is
4101 that "executable" commands cannot return anything other than a boolean
4102 result and can only support numeric or string input, whereas some
4103 discoverable commands may only be for direct use through
4104 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
4105 pointers, or other custom uses. The "executable" commands are to
4106 support parameterisations of ENGINE behaviour that can be
4107 unambiguously defined by ENGINEs and used consistently across any
4108 OpenSSL-based application. Commands have been added to all the
4109 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
4110 control over shared-library paths without source code alterations.
4113 *) Changed all ENGINE implementations to dynamically allocate their
4114 ENGINEs rather than declaring them statically. Apart from this being
4115 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
4116 this also allows the implementations to compile without using the
4117 internal engine_int.h header.
4120 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
4121 'const' value. Any code that should be able to modify a RAND_METHOD
4122 should already have non-const pointers to it (ie. they should only
4123 modify their own ones).
4126 *) Made a variety of little tweaks to the ENGINE code.
4127 - "atalla" and "ubsec" string definitions were moved from header files
4128 to C code. "nuron" string definitions were placed in variables
4129 rather than hard-coded - allowing parameterisation of these values
4130 later on via ctrl() commands.
4131 - Removed unused "#if 0"'d code.
4132 - Fixed engine list iteration code so it uses ENGINE_free() to release
4133 structural references.
4134 - Constified the RAND_METHOD element of ENGINE structures.
4135 - Constified various get/set functions as appropriate and added
4136 missing functions (including a catch-all ENGINE_cpy that duplicates
4137 all ENGINE values onto a new ENGINE except reference counts/state).
4138 - Removed NULL parameter checks in get/set functions. Setting a method
4139 or function to NULL is a way of cancelling out a previously set
4140 value. Passing a NULL ENGINE parameter is just plain stupid anyway
4141 and doesn't justify the extra error symbols and code.
4142 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
4143 flags from engine_int.h to engine.h.
4144 - Changed prototypes for ENGINE handler functions (init(), finish(),
4145 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
4148 *) Implement binary inversion algorithm for BN_mod_inverse in addition
4149 to the algorithm using long division. The binary algorithm can be
4150 used only if the modulus is odd. On 32-bit systems, it is faster
4151 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
4152 roughly 5-15% for 256-bit moduli), so we use it only for moduli
4153 up to 450 bits. In 64-bit environments, the binary algorithm
4154 appears to be advantageous for much longer moduli; here we use it
4155 for moduli up to 2048 bits.
4158 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
4159 could not support the combine flag in choice fields.
4162 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
4163 extensions from a certificate request to the certificate.
4166 *) Allow multiple 'certopt' and 'nameopt' options to be separated
4167 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
4168 file: this allows the display of the certificate about to be
4169 signed to be customised, to allow certain fields to be included
4170 or excluded and extension details. The old system didn't display
4171 multicharacter strings properly, omitted fields not in the policy
4172 and couldn't display additional details such as extensions.
4175 *) Function EC_POINTs_mul for multiple scalar multiplication
4176 of an arbitrary number of elliptic curve points
4177 \sum scalars[i]*points[i],
4178 optionally including the generator defined for the EC_GROUP:
4179 scalar*generator + \sum scalars[i]*points[i].
4181 EC_POINT_mul is a simple wrapper function for the typical case
4182 that the point list has just one item (besides the optional
4186 *) First EC_METHODs for curves over GF(p):
4188 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
4189 operations and provides various method functions that can also
4190 operate with faster implementations of modular arithmetic.
4192 EC_GFp_mont_method() reuses most functions that are part of
4193 EC_GFp_simple_method, but uses Montgomery arithmetic.
4195 [Bodo Moeller; point addition and point doubling
4196 implementation directly derived from source code provided by
4197 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
4199 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
4200 crypto/ec/ec_lib.c):
4202 Curves are EC_GROUP objects (with an optional group generator)
4203 based on EC_METHODs that are built into the library.
4205 Points are EC_POINT objects based on EC_GROUP objects.
4207 Most of the framework would be able to handle curves over arbitrary
4208 finite fields, but as there are no obvious types for fields other
4209 than GF(p), some functions are limited to that for now.
4212 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
4213 that the file contains a complete HTTP response.
4216 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
4217 change the def and num file printf format specifier from "%-40sXXX"
4218 to "%-39s XXX". The latter will always guarantee a space after the
4219 field while the former will cause them to run together if the field
4220 is 40 of more characters long.
4223 *) Constify the cipher and digest 'method' functions and structures
4224 and modify related functions to take constant EVP_MD and EVP_CIPHER
4228 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
4229 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
4232 *) Modify EVP_Digest*() routines so they now return values. Although the
4233 internal software routines can never fail additional hardware versions
4237 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
4239 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
4240 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
4243 ERR_R_NESTED_ASN1_ERROR
4245 ERR_R_MISSING_ASN1_EOS
4246 were 4 .. 9, conflicting with
4247 ERR_LIB_RSA (= ERR_R_RSA_LIB)
4249 ERR_LIB_PEM (= ERR_R_PEM_LIB).
4250 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
4252 Add new error code 'ERR_R_INTERNAL_ERROR'.
4255 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
4259 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
4260 sets the subject name for a new request or supersedes the
4261 subject name in a given request. Formats that can be parsed are
4262 'CN=Some Name, OU=myOU, C=IT'
4264 'CN=Some Name/OU=myOU/C=IT'.
4266 Add options '-batch' and '-verbose' to 'openssl req'.
4267 [Massimiliano Pala <madwolf@hackmasters.net>]
4269 *) Introduce the possibility to access global variables through
4270 functions on platform were that's the best way to handle exporting
4271 global variables in shared libraries. To enable this functionality,
4272 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
4273 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
4274 is normally done by Configure or something similar).
4276 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
4277 in the source file (foo.c) like this:
4279 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
4280 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
4282 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
4283 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
4285 OPENSSL_DECLARE_GLOBAL(int,foo);
4286 #define foo OPENSSL_GLOBAL_REF(foo)
4287 OPENSSL_DECLARE_GLOBAL(double,bar);
4288 #define bar OPENSSL_GLOBAL_REF(bar)
4290 The #defines are very important, and therefore so is including the
4291 header file everywhere where the defined globals are used.
4293 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
4294 of ASN.1 items, but that structure is a bit different.
4296 The largest change is in util/mkdef.pl which has been enhanced with
4297 better and easier to understand logic to choose which symbols should
4298 go into the Windows .def files as well as a number of fixes and code
4299 cleanup (among others, algorithm keywords are now sorted
4300 lexicographically to avoid constant rewrites).
4303 *) In BN_div() keep a copy of the sign of 'num' before writing the
4304 result to 'rm' because if rm==num the value will be overwritten
4305 and produce the wrong result if 'num' is negative: this caused
4306 problems with BN_mod() and BN_nnmod().
4309 *) Function OCSP_request_verify(). This checks the signature on an
4310 OCSP request and verifies the signer certificate. The signer
4311 certificate is just checked for a generic purpose and OCSP request
4315 *) Add OCSP_check_validity() function to check the validity of OCSP
4316 responses. OCSP responses are prepared in real time and may only
4317 be a few seconds old. Simply checking that the current time lies
4318 between thisUpdate and nextUpdate max reject otherwise valid responses
4319 caused by either OCSP responder or client clock inaccuracy. Instead
4320 we allow thisUpdate and nextUpdate to fall within a certain period of
4321 the current time. The age of the response can also optionally be
4322 checked. Two new options -validity_period and -status_age added to
4326 *) If signature or public key algorithm is unrecognized print out its
4327 OID rather that just UNKNOWN.
4330 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
4331 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
4332 ID to be generated from the issuer certificate alone which can then be
4333 passed to OCSP_id_issuer_cmp().
4336 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
4337 ASN1 modules to export functions returning ASN1_ITEM pointers
4338 instead of the ASN1_ITEM structures themselves. This adds several
4339 new macros which allow the underlying ASN1 function/structure to
4340 be accessed transparently. As a result code should not use ASN1_ITEM
4341 references directly (such as &X509_it) but instead use the relevant
4342 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
4343 use of the new ASN1 code on platforms where exporting structures
4344 is problematical (for example in shared libraries) but exporting
4345 functions returning pointers to structures is not.
4348 *) Add support for overriding the generation of SSL/TLS session IDs.
4349 These callbacks can be registered either in an SSL_CTX or per SSL.
4350 The purpose of this is to allow applications to control, if they wish,
4351 the arbitrary values chosen for use as session IDs, particularly as it
4352 can be useful for session caching in multiple-server environments. A
4353 command-line switch for testing this (and any client code that wishes
4354 to use such a feature) has been added to "s_server".
4355 [Geoff Thorpe, Lutz Jaenicke]
4357 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
4358 of the form '#if defined(...) || defined(...) || ...' and
4359 '#if !defined(...) && !defined(...) && ...'. This also avoids
4360 the growing number of special cases it was previously handling.
4363 *) Make all configuration macros available for application by making
4364 sure they are available in opensslconf.h, by giving them names starting
4365 with "OPENSSL_" to avoid conflicts with other packages and by making
4366 sure e_os2.h will cover all platform-specific cases together with
4368 Additionally, it is now possible to define configuration/platform-
4369 specific names (called "system identities"). In the C code, these
4370 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
4371 macro with the name beginning with "OPENSSL_SYS_", which is determined
4372 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
4376 *) New option -set_serial to 'req' and 'x509' this allows the serial
4377 number to use to be specified on the command line. Previously self
4378 signed certificates were hard coded with serial number 0 and the
4379 CA options of 'x509' had to use a serial number in a file which was
4383 *) New options to 'ca' utility to support V2 CRL entry extensions.
4384 Currently CRL reason, invalidity date and hold instruction are
4385 supported. Add new CRL extensions to V3 code and some new objects.
4388 *) New function EVP_CIPHER_CTX_set_padding() this is used to
4389 disable standard block padding (aka PKCS#5 padding) in the EVP
4390 API, which was previously mandatory. This means that the data is
4391 not padded in any way and so the total length much be a multiple
4392 of the block size, otherwise an error occurs.
4395 *) Initial (incomplete) OCSP SSL support.
4398 *) New function OCSP_parse_url(). This splits up a URL into its host,
4399 port and path components: primarily to parse OCSP URLs. New -url
4400 option to ocsp utility.
4403 *) New nonce behavior. The return value of OCSP_check_nonce() now
4404 reflects the various checks performed. Applications can decide
4405 whether to tolerate certain situations such as an absent nonce
4406 in a response when one was present in a request: the ocsp application
4407 just prints out a warning. New function OCSP_add1_basic_nonce()
4408 this is to allow responders to include a nonce in a response even if
4409 the request is nonce-less.
4412 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
4413 skipped when using openssl x509 multiple times on a single input file,
4414 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
4417 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
4418 set string type: to handle setting ASN1_TIME structures. Fix ca
4419 utility to correctly initialize revocation date of CRLs.
4422 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
4423 the clients preferred ciphersuites and rather use its own preferences.
4424 Should help to work around M$ SGC (Server Gated Cryptography) bug in
4425 Internet Explorer by ensuring unchanged hash method during stepup.
4426 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
4429 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
4430 to aes and add a new 'exist' option to print out symbols that don't
4434 *) Additional options to ocsp utility to allow flags to be set and
4435 additional certificates supplied.
4438 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
4439 OCSP client a number of certificate to only verify the response
4443 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
4444 handle the new API. Currently only ECB, CBC modes supported. Add new
4447 Add TLS AES ciphersuites as described in RFC3268, "Advanced
4448 Encryption Standard (AES) Ciphersuites for Transport Layer
4449 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
4450 not enabled by default and were not part of the "ALL" ciphersuite
4451 alias because they were not yet official; they could be
4452 explicitly requested by specifying the "AESdraft" ciphersuite
4453 group alias. In the final release of OpenSSL 0.9.7, the group
4454 alias is called "AES" and is part of "ALL".)
4455 [Ben Laurie, Steve Henson, Bodo Moeller]
4457 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
4458 request to response.
4461 *) Functions for OCSP responders. OCSP_request_onereq_count(),
4462 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
4463 extract information from a certificate request. OCSP_response_create()
4464 creates a response and optionally adds a basic response structure.
4465 OCSP_basic_add1_status() adds a complete single response to a basic
4466 response and returns the OCSP_SINGLERESP structure just added (to allow
4467 extensions to be included for example). OCSP_basic_add1_cert() adds a
4468 certificate to a basic response and OCSP_basic_sign() signs a basic
4469 response with various flags. New helper functions ASN1_TIME_check()
4470 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
4471 (converts ASN1_TIME to GeneralizedTime).
4474 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
4475 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
4476 structure from a certificate. X509_pubkey_digest() digests the public_key
4477 contents: this is used in various key identifiers.
4480 *) Make sk_sort() tolerate a NULL argument.
4481 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
4483 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
4484 passed by the function are trusted implicitly. If any of them signed the
4485 response then it is assumed to be valid and is not verified.
4488 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
4489 to data. This was previously part of the PKCS7 ASN1 code. This
4490 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
4491 [Steve Henson, reported by Kenneth R. Robinette
4492 <support@securenetterm.com>]
4494 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
4495 routines: without these tracing memory leaks is very painful.
4496 Fix leaks in PKCS12 and PKCS7 routines.
4499 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
4500 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
4501 effectively meant GeneralizedTime would never be used. Now it
4502 is initialised to -1 but X509_time_adj() now has to check the value
4503 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
4504 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
4505 [Steve Henson, reported by Kenneth R. Robinette
4506 <support@securenetterm.com>]
4508 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
4509 result in a zero length in the ASN1_INTEGER structure which was
4510 not consistent with the structure when d2i_ASN1_INTEGER() was used
4511 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
4512 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
4513 where it did not print out a minus for negative ASN1_INTEGER.
4516 *) Add summary printout to ocsp utility. The various functions which
4517 convert status values to strings have been renamed to:
4518 OCSP_response_status_str(), OCSP_cert_status_str() and
4519 OCSP_crl_reason_str() and are no longer static. New options
4520 to verify nonce values and to disable verification. OCSP response
4521 printout format cleaned up.
4524 *) Add additional OCSP certificate checks. These are those specified
4525 in RFC2560. This consists of two separate checks: the CA of the
4526 certificate being checked must either be the OCSP signer certificate
4527 or the issuer of the OCSP signer certificate. In the latter case the
4528 OCSP signer certificate must contain the OCSP signing extended key
4529 usage. This check is performed by attempting to match the OCSP
4530 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
4531 in the OCSP_CERTID structures of the response.
4534 *) Initial OCSP certificate verification added to OCSP_basic_verify()
4535 and related routines. This uses the standard OpenSSL certificate
4536 verify routines to perform initial checks (just CA validity) and
4537 to obtain the certificate chain. Then additional checks will be
4538 performed on the chain. Currently the root CA is checked to see
4539 if it is explicitly trusted for OCSP signing. This is used to set
4540 a root CA as a global signing root: that is any certificate that
4541 chains to that CA is an acceptable OCSP signing certificate.
4544 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
4545 extensions from a separate configuration file.
4546 As when reading extensions from the main configuration file,
4547 the '-extensions ...' option may be used for specifying the
4549 [Massimiliano Pala <madwolf@comune.modena.it>]
4551 *) New OCSP utility. Allows OCSP requests to be generated or
4552 read. The request can be sent to a responder and the output
4553 parsed, outputed or printed in text form. Not complete yet:
4554 still needs to check the OCSP response validity.
4557 *) New subcommands for 'openssl ca':
4558 'openssl ca -status <serial>' prints the status of the cert with
4559 the given serial number (according to the index file).
4560 'openssl ca -updatedb' updates the expiry status of certificates
4562 [Massimiliano Pala <madwolf@comune.modena.it>]
4564 *) New '-newreq-nodes' command option to CA.pl. This is like
4565 '-newreq', but calls 'openssl req' with the '-nodes' option
4566 so that the resulting key is not encrypted.
4567 [Damien Miller <djm@mindrot.org>]
4569 *) New configuration for the GNU Hurd.
4570 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
4572 *) Initial code to implement OCSP basic response verify. This
4573 is currently incomplete. Currently just finds the signer's
4574 certificate and verifies the signature on the response.
4577 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
4578 value of OPENSSLDIR. This is available via the new '-d' option
4579 to 'openssl version', and is also included in 'openssl version -a'.
4582 *) Allowing defining memory allocation callbacks that will be given
4583 file name and line number information in additional arguments
4584 (a const char* and an int). The basic functionality remains, as
4585 well as the original possibility to just replace malloc(),
4586 realloc() and free() by functions that do not know about these
4587 additional arguments. To register and find out the current
4588 settings for extended allocation functions, the following
4589 functions are provided:
4591 CRYPTO_set_mem_ex_functions
4592 CRYPTO_set_locked_mem_ex_functions
4593 CRYPTO_get_mem_ex_functions
4594 CRYPTO_get_locked_mem_ex_functions
4596 These work the same way as CRYPTO_set_mem_functions and friends.
4597 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
4598 extended allocation function is enabled.
4599 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
4600 a conventional allocation function is enabled.
4601 [Richard Levitte, Bodo Moeller]
4603 *) Finish off removing the remaining LHASH function pointer casts.
4604 There should no longer be any prototype-casting required when using
4605 the LHASH abstraction, and any casts that remain are "bugs". See
4606 the callback types and macros at the head of lhash.h for details
4607 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
4610 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
4611 If /dev/[u]random devices are not available or do not return enough
4612 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
4614 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
4615 /etc/entropy will be queried once each in this sequence, quering stops
4616 when enough entropy was collected without querying more sockets.
4619 *) Change the Unix RAND_poll() variant to be able to poll several
4620 random devices, as specified by DEVRANDOM, until a sufficient amount
4621 of data has been collected. We spend at most 10 ms on each file
4622 (select timeout) and read in non-blocking mode. DEVRANDOM now
4623 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
4624 (previously it was just the string "/dev/urandom"), so on typical
4625 platforms the 10 ms delay will never occur.
4626 Also separate out the Unix variant to its own file, rand_unix.c.
4627 For VMS, there's a currently-empty rand_vms.c.
4630 *) Move OCSP client related routines to ocsp_cl.c. These
4631 provide utility functions which an application needing
4632 to issue a request to an OCSP responder and analyse the
4633 response will typically need: as opposed to those which an
4634 OCSP responder itself would need which will be added later.
4636 OCSP_request_sign() signs an OCSP request with an API similar
4637 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
4638 response. OCSP_response_get1_basic() extracts basic response
4639 from response. OCSP_resp_find_status(): finds and extracts status
4640 information from an OCSP_CERTID structure (which will be created
4641 when the request structure is built). These are built from lower
4642 level functions which work on OCSP_SINGLERESP structures but
4643 wont normally be used unless the application wishes to examine
4644 extensions in the OCSP response for example.
4646 Replace nonce routines with a pair of functions.
4647 OCSP_request_add1_nonce() adds a nonce value and optionally
4648 generates a random value. OCSP_check_nonce() checks the
4649 validity of the nonce in an OCSP response.
4652 *) Change function OCSP_request_add() to OCSP_request_add0_id().
4653 This doesn't copy the supplied OCSP_CERTID and avoids the
4654 need to free up the newly created id. Change return type
4655 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
4656 This can then be used to add extensions to the request.
4657 Deleted OCSP_request_new(), since most of its functionality
4658 is now in OCSP_REQUEST_new() (and the case insensitive name
4659 clash) apart from the ability to set the request name which
4660 will be added elsewhere.
4663 *) Update OCSP API. Remove obsolete extensions argument from
4664 various functions. Extensions are now handled using the new
4665 OCSP extension code. New simple OCSP HTTP function which
4666 can be used to send requests and parse the response.
4669 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
4670 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
4671 uses the special reorder version of SET OF to sort the attributes
4672 and reorder them to match the encoded order. This resolves a long
4673 standing problem: a verify on a PKCS7 structure just after signing
4674 it used to fail because the attribute order did not match the
4675 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
4676 it uses the received order. This is necessary to tolerate some broken
4677 software that does not order SET OF. This is handled by encoding
4678 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
4679 to produce the required SET OF.
4682 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
4683 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
4684 files to get correct declarations of the ASN.1 item variables.
4687 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
4688 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
4689 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
4690 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
4691 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
4692 ASN1_ITEM and no wrapper functions.
4695 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
4696 replace the old function pointer based I/O routines. Change most of
4697 the *_d2i_bio() and *_d2i_fp() functions to use these.
4700 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
4701 lines, recognice more "algorithms" that can be deselected, and make
4702 it complain about algorithm deselection that isn't recognised.
4705 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
4706 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
4707 to use new functions. Add NO_ASN1_OLD which can be set to remove
4708 some old style ASN1 functions: this can be used to determine if old
4709 code will still work when these eventually go away.
4712 *) New extension functions for OCSP structures, these follow the
4713 same conventions as certificates and CRLs.
4716 *) New function X509V3_add1_i2d(). This automatically encodes and
4717 adds an extension. Its behaviour can be customised with various
4718 flags to append, replace or delete. Various wrappers added for
4719 certifcates and CRLs.
4722 *) Fix to avoid calling the underlying ASN1 print routine when
4723 an extension cannot be parsed. Correct a typo in the
4724 OCSP_SERVICELOC extension. Tidy up print OCSP format.
4727 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
4728 entries for variables.
4731 *) Add functionality to apps/openssl.c for detecting locking
4732 problems: As the program is single-threaded, all we have
4733 to do is register a locking callback using an array for
4734 storing which locks are currently held by the program.
4737 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
4738 SSL_get_ex_data_X509_STORE_idx(), which is used in
4739 ssl_verify_cert_chain() and thus can be called at any time
4740 during TLS/SSL handshakes so that thread-safety is essential.
4741 Unfortunately, the ex_data design is not at all suited
4742 for multi-threaded use, so it probably should be abolished.
4745 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
4746 [Broadcom, tweaked and integrated by Geoff Thorpe]
4748 *) Move common extension printing code to new function
4749 X509V3_print_extensions(). Reorganise OCSP print routines and
4750 implement some needed OCSP ASN1 functions. Add OCSP extensions.
4753 *) New function X509_signature_print() to remove duplication in some
4757 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
4758 set (this was treated exactly the same as SET OF previously). This
4759 is used to reorder the STACK representing the structure to match the
4760 encoding. This will be used to get round a problem where a PKCS7
4761 structure which was signed could not be verified because the STACK
4762 order did not reflect the encoded order.
4765 *) Reimplement the OCSP ASN1 module using the new code.
4768 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
4769 for its ASN1 operations. The old style function pointers still exist
4770 for now but they will eventually go away.
4773 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
4774 completely replaces the old ASN1 functionality with a table driven
4775 encoder and decoder which interprets an ASN1_ITEM structure describing
4776 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
4777 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
4778 has also been converted to the new form.
4781 *) Change BN_mod_exp_recp so that negative moduli are tolerated
4782 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
4783 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
4784 for negative moduli.
4787 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
4788 of not touching the result's sign bit.
4791 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
4795 *) Changed the LHASH code to use prototypes for callbacks, and created
4796 macros to declare and implement thin (optionally static) functions
4797 that provide type-safety and avoid function pointer casting for the
4798 type-specific callbacks.
4801 *) Added Kerberos Cipher Suites to be used with TLS, as written in
4803 [Veers Staats <staatsvr@asc.hpc.mil>,
4804 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
4806 *) Reformat the FAQ so the different questions and answers can be divided
4807 in sections depending on the subject.
4810 *) Have the zlib compression code load ZLIB.DLL dynamically under
4814 *) New function BN_mod_sqrt for computing square roots modulo a prime
4815 (using the probabilistic Tonelli-Shanks algorithm unless
4816 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
4817 be handled deterministically).
4818 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
4820 *) Make BN_mod_inverse faster by explicitly handling small quotients
4821 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
4822 512 bits], about 30% for larger ones [1024 or 2048 bits].)
4825 *) New function BN_kronecker.
4828 *) Fix BN_gcd so that it works on negative inputs; the result is
4829 positive unless both parameters are zero.
4830 Previously something reasonably close to an infinite loop was
4831 possible because numbers could be growing instead of shrinking
4832 in the implementation of Euclid's algorithm.
4835 *) Fix BN_is_word() and BN_is_one() macros to take into account the
4836 sign of the number in question.
4838 Fix BN_is_word(a,w) to work correctly for w == 0.
4840 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
4841 because its test if the absolute value of 'a' equals 'w'.
4842 Note that BN_abs_is_word does *not* handle w == 0 reliably;
4843 it exists mostly for use in the implementations of BN_is_zero(),
4844 BN_is_one(), and BN_is_word().
4847 *) New function BN_swap.
4850 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
4851 the exponentiation functions are more likely to produce reasonable
4852 results on negative inputs.
4855 *) Change BN_mod_mul so that the result is always non-negative.
4856 Previously, it could be negative if one of the factors was negative;
4857 I don't think anyone really wanted that behaviour.
4860 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
4861 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
4862 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
4863 and add new functions:
4872 BN_mod_lshift1_quick
4876 These functions always generate non-negative results.
4878 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
4879 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
4881 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
4882 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
4883 be reduced modulo m.
4884 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
4887 The following entry accidentily appeared in the CHANGES file
4888 distributed with OpenSSL 0.9.7. The modifications described in
4889 it do *not* apply to OpenSSL 0.9.7.
4891 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4892 was actually never needed) and in BN_mul(). The removal in BN_mul()
4893 required a small change in bn_mul_part_recursive() and the addition
4894 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4895 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4896 bn_sub_words() and bn_add_words() except they take arrays with
4901 *) In 'openssl passwd', verify passwords read from the terminal
4902 unless the '-salt' option is used (which usually means that
4903 verification would just waste user's time since the resulting
4904 hash is going to be compared with some given password hash)
4905 or the new '-noverify' option is used.
4907 This is an incompatible change, but it does not affect
4908 non-interactive use of 'openssl passwd' (passwords on the command
4909 line, '-stdin' option, '-in ...' option) and thus should not
4913 *) Remove all references to RSAref, since there's no more need for it.
4916 *) Make DSO load along a path given through an environment variable
4917 (SHLIB_PATH) with shl_load().
4920 *) Constify the ENGINE code as a result of BIGNUM constification.
4921 Also constify the RSA code and most things related to it. In a
4922 few places, most notable in the depth of the ASN.1 code, ugly
4923 casts back to non-const were required (to be solved at a later
4927 *) Make it so the openssl application has all engines loaded by default.
4930 *) Constify the BIGNUM routines a little more.
4933 *) Add the following functions:
4935 ENGINE_load_cswift()
4937 ENGINE_load_atalla()
4939 ENGINE_load_builtin_engines()
4941 That way, an application can itself choose if external engines that
4942 are built-in in OpenSSL shall ever be used or not. The benefit is
4943 that applications won't have to be linked with libdl or other dso
4944 libraries unless it's really needed.
4946 Changed 'openssl engine' to load all engines on demand.
4947 Changed the engine header files to avoid the duplication of some
4948 declarations (they differed!).
4951 *) 'openssl engine' can now list capabilities.
4954 *) Better error reporting in 'openssl engine'.
4957 *) Never call load_dh_param(NULL) in s_server.
4960 *) Add engine application. It can currently list engines by name and
4961 identity, and test if they are actually available.
4964 *) Improve RPM specification file by forcing symbolic linking and making
4965 sure the installed documentation is also owned by root.root.
4966 [Damien Miller <djm@mindrot.org>]
4968 *) Give the OpenSSL applications more possibilities to make use of
4969 keys (public as well as private) handled by engines.
4972 *) Add OCSP code that comes from CertCo.
4975 *) Add VMS support for the Rijndael code.
4978 *) Added untested support for Nuron crypto accelerator.
4981 *) Add support for external cryptographic devices. This code was
4982 previously distributed separately as the "engine" branch.
4983 [Geoff Thorpe, Richard Levitte]
4985 *) Rework the filename-translation in the DSO code. It is now possible to
4986 have far greater control over how a "name" is turned into a filename
4987 depending on the operating environment and any oddities about the
4988 different shared library filenames on each system.
4991 *) Support threads on FreeBSD-elf in Configure.
4994 *) Fix for SHA1 assembly problem with MASM: it produces
4995 warnings about corrupt line number information when assembling
4996 with debugging information. This is caused by the overlapping
4998 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5001 NCONF_get_number() has no error checking at all. As a replacement,
5002 NCONF_get_number_e() is defined (_e for "error checking") and is
5003 promoted strongly. The old NCONF_get_number is kept around for
5004 binary backward compatibility.
5005 Make it possible for methods to load from something other than a BIO,
5006 by providing a function pointer that is given a name instead of a BIO.
5007 For example, this could be used to load configuration data from an
5011 *) Fix for non blocking accept BIOs. Added new I/O special reason
5012 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5013 with non blocking I/O was not possible because no retry code was
5014 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5018 *) Added the beginnings of Rijndael support.
5021 *) Fix for bug in DirectoryString mask setting. Add support for
5022 X509_NAME_print_ex() in 'req' and X509_print_ex() function
5023 to allow certificate printing to more controllable, additional
5024 'certopt' option to 'x509' to allow new printing options to be
5028 *) Clean old EAY MD5 hack from e_os.h.
5031 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5033 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5034 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5035 [Joe Orton, Steve Henson]
5037 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5039 *) Fix additional bug revealed by the NISCC test suite:
5041 Stop bug triggering large recursion when presented with
5042 certain ASN.1 tags (CVE-2003-0851)
5045 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5047 *) Fix various bugs revealed by running the NISCC test suite:
5049 Stop out of bounds reads in the ASN1 code when presented with
5050 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5052 If verify callback ignores invalid public key errors don't try to check
5053 certificate signature with the NULL public key.
5057 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5058 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5062 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5063 extra data after the compression methods not only for TLS 1.0
5064 but also for SSL 3.0 (as required by the specification).
5065 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5067 *) Change X509_certificate_type() to mark the key as exported/exportable
5068 when it's 512 *bits* long, not 512 bytes.
5071 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5073 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5074 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5075 a protocol version number mismatch like a decryption error
5076 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5079 *) Turn on RSA blinding by default in the default implementation
5080 to avoid a timing attack. Applications that don't want it can call
5081 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5082 They would be ill-advised to do so in most cases.
5083 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5085 *) Change RSA blinding code so that it works when the PRNG is not
5086 seeded (in this case, the secret RSA exponent is abused as
5087 an unpredictable seed -- if it is not unpredictable, there
5088 is no point in blinding anyway). Make RSA blinding thread-safe
5089 by remembering the creator's thread ID in rsa->blinding and
5090 having all other threads use local one-time blinding factors
5091 (this requires more computation than sharing rsa->blinding, but
5092 avoids excessive locking; and if an RSA object is not shared
5093 between threads, blinding will still be very fast).
5096 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5098 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5099 via timing by performing a MAC computation even if incorrrect
5100 block cipher padding has been found. This is a countermeasure
5101 against active attacks where the attacker has to distinguish
5102 between bad padding and a MAC verification error. (CVE-2003-0078)
5104 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5105 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5106 Martin Vuagnoux (EPFL, Ilion)]
5108 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
5110 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
5111 memory from it's contents. This is done with a counter that will
5112 place alternating values in each byte. This can be used to solve
5113 two issues: 1) the removal of calls to memset() by highly optimizing
5114 compilers, and 2) cleansing with other values than 0, since those can
5115 be read through on certain media, for example a swap space on disk.
5118 *) Bugfix: client side session caching did not work with external caching,
5119 because the session->cipher setting was not restored when reloading
5120 from the external cache. This problem was masked, when
5121 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
5122 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
5125 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
5126 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
5127 [Zeev Lieber <zeev-l@yahoo.com>]
5129 *) Undo an undocumented change introduced in 0.9.6e which caused
5130 repeated calls to OpenSSL_add_all_ciphers() and
5131 OpenSSL_add_all_digests() to be ignored, even after calling
5135 *) Change the default configuration reader to deal with last line not
5136 being properly terminated.
5139 *) Change X509_NAME_cmp() so it applies the special rules on handling
5140 DN values that are of type PrintableString, as well as RDNs of type
5141 emailAddress where the value has the type ia5String.
5142 [stefank@valicert.com via Richard Levitte]
5144 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
5145 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
5146 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
5147 the bitwise-OR of the two for use by the majority of applications
5148 wanting this behaviour, and update the docs. The documented
5149 behaviour and actual behaviour were inconsistent and had been
5150 changing anyway, so this is more a bug-fix than a behavioural
5152 [Geoff Thorpe, diagnosed by Nadav Har'El]
5154 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
5155 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
5158 *) Fix initialization code race conditions in
5159 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
5160 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
5161 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
5162 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
5163 ssl2_get_cipher_by_char(),
5164 ssl3_get_cipher_by_char().
5165 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
5167 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
5168 the cached sessions are flushed, as the remove_cb() might use ex_data
5169 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
5170 (see [openssl.org #212]).
5171 [Geoff Thorpe, Lutz Jaenicke]
5173 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
5174 length, instead of the encoding length to d2i_ASN1_OBJECT.
5177 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
5179 *) [In 0.9.6g-engine release:]
5180 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
5181 [Lynn Gazis <lgazis@rainbow.com>]
5183 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
5185 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
5186 and get fix the header length calculation.
5187 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
5188 Alon Kantor <alonk@checkpoint.com> (and others),
5191 *) Use proper error handling instead of 'assertions' in buffer
5192 overflow checks added in 0.9.6e. This prevents DoS (the
5193 assertions could call abort()).
5194 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
5196 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
5198 *) Add various sanity checks to asn1_get_length() to reject
5199 the ASN1 length bytes if they exceed sizeof(long), will appear
5200 negative or the content length exceeds the length of the
5202 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5204 *) Fix cipher selection routines: ciphers without encryption had no flags
5205 for the cipher strength set and where therefore not handled correctly
5206 by the selection routines (PR #130).
5209 *) Fix EVP_dsa_sha macro.
5213 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
5214 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
5215 that was added in OpenSSL 0.9.6d.
5217 As the countermeasure turned out to be incompatible with some
5218 broken SSL implementations, the new option is part of SSL_OP_ALL.
5219 SSL_OP_ALL is usually employed when compatibility with weird SSL
5220 implementations is desired (e.g. '-bugs' option to 's_client' and
5221 's_server'), so the new option is automatically set in many
5225 *) Changes in security patch:
5227 Changes marked "(CHATS)" were sponsored by the Defense Advanced
5228 Research Projects Agency (DARPA) and Air Force Research Laboratory,
5229 Air Force Materiel Command, USAF, under agreement number
5232 *) Add various sanity checks to asn1_get_length() to reject
5233 the ASN1 length bytes if they exceed sizeof(long), will appear
5234 negative or the content length exceeds the length of the
5235 supplied buffer. (CVE-2002-0659)
5236 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5238 *) Assertions for various potential buffer overflows, not known to
5240 [Ben Laurie (CHATS)]
5242 *) Various temporary buffers to hold ASCII versions of integers were
5243 too small for 64 bit platforms. (CVE-2002-0655)
5244 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
5246 *) Remote buffer overflow in SSL3 protocol - an attacker could
5247 supply an oversized session ID to a client. (CVE-2002-0656)
5248 [Ben Laurie (CHATS)]
5250 *) Remote buffer overflow in SSL2 protocol - an attacker could
5251 supply an oversized client master key. (CVE-2002-0656)
5252 [Ben Laurie (CHATS)]
5254 Changes between 0.9.6c and 0.9.6d [9 May 2002]
5256 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
5257 encoded as NULL) with id-dsa-with-sha1.
5258 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
5260 *) Check various X509_...() return values in apps/req.c.
5261 [Nils Larsch <nla@trustcenter.de>]
5263 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
5264 an end-of-file condition would erronously be flagged, when the CRLF
5265 was just at the end of a processed block. The bug was discovered when
5266 processing data through a buffering memory BIO handing the data to a
5267 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
5268 <ptsekov@syntrex.com> and Nedelcho Stanev.
5271 *) Implement a countermeasure against a vulnerability recently found
5272 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
5273 before application data chunks to avoid the use of known IVs
5274 with data potentially chosen by the attacker.
5277 *) Fix length checks in ssl3_get_client_hello().
5280 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
5281 to prevent ssl3_read_internal() from incorrectly assuming that
5282 ssl3_read_bytes() found application data while handshake
5283 processing was enabled when in fact s->s3->in_read_app_data was
5284 merely automatically cleared during the initial handshake.
5285 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
5287 *) Fix object definitions for Private and Enterprise: they were not
5288 recognized in their shortname (=lowercase) representation. Extend
5289 obj_dat.pl to issue an error when using undefined keywords instead
5290 of silently ignoring the problem (Svenning Sorensen
5291 <sss@sss.dnsalias.net>).
5294 *) Fix DH_generate_parameters() so that it works for 'non-standard'
5295 generators, i.e. generators other than 2 and 5. (Previously, the
5296 code did not properly initialise the 'add' and 'rem' values to
5297 BN_generate_prime().)
5299 In the new general case, we do not insist that 'generator' is
5300 actually a primitive root: This requirement is rather pointless;
5301 a generator of the order-q subgroup is just as good, if not
5305 *) Map new X509 verification errors to alerts. Discovered and submitted by
5306 Tom Wu <tom@arcot.com>.
5309 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
5310 returning non-zero before the data has been completely received
5311 when using non-blocking I/O.
5312 [Bodo Moeller; problem pointed out by John Hughes]
5314 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
5315 [Ben Laurie, Lutz Jaenicke]
5317 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
5318 Yoram Zahavi <YoramZ@gilian.com>).
5321 *) Add information about CygWin 1.3 and on, and preserve proper
5322 configuration for the versions before that.
5323 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5325 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
5326 check whether we deal with a copy of a session and do not delete from
5327 the cache in this case. Problem reported by "Izhar Shoshani Levi"
5328 <izhar@checkpoint.com>.
5331 *) Do not store session data into the internal session cache, if it
5332 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
5333 flag is set). Proposed by Aslam <aslam@funk.com>.
5336 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
5340 *) [In 0.9.6d-engine release:]
5341 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5342 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5344 *) Add the configuration target linux-s390x.
5345 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
5347 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
5348 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
5349 variable as an indication that a ClientHello message has been
5350 received. As the flag value will be lost between multiple
5351 invocations of ssl3_accept when using non-blocking I/O, the
5352 function may not be aware that a handshake has actually taken
5353 place, thus preventing a new session from being added to the
5356 To avoid this problem, we now set s->new_session to 2 instead of
5357 using a local variable.
5358 [Lutz Jaenicke, Bodo Moeller]
5360 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
5361 if the SSL_R_LENGTH_MISMATCH error is detected.
5362 [Geoff Thorpe, Bodo Moeller]
5364 *) New 'shared_ldflag' column in Configure platform table.
5367 *) Fix EVP_CIPHER_mode macro.
5368 ["Dan S. Camper" <dan@bti.net>]
5370 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
5371 type, we must throw them away by setting rr->length to 0.
5372 [D P Chang <dpc@qualys.com>]
5374 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
5376 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
5377 <Dominikus.Scherkl@biodata.com>. (The previous implementation
5378 worked incorrectly for those cases where range = 10..._2 and
5379 3*range is two bits longer than range.)
5382 *) Only add signing time to PKCS7 structures if it is not already
5386 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
5387 OBJ_ld_ce should be OBJ_id_ce.
5388 Also some ip-pda OIDs in crypto/objects/objects.txt were
5389 incorrect (cf. RFC 3039).
5390 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
5392 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
5393 returns early because it has nothing to do.
5394 [Andy Schneider <andy.schneider@bjss.co.uk>]
5396 *) [In 0.9.6c-engine release:]
5397 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
5398 [Andy Schneider <andy.schneider@bjss.co.uk>]
5400 *) [In 0.9.6c-engine release:]
5401 Add support for Cryptographic Appliance's keyserver technology.
5402 (Use engine 'keyclient')
5403 [Cryptographic Appliances and Geoff Thorpe]
5405 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
5406 is called via tools/c89.sh because arguments have to be
5407 rearranged (all '-L' options must appear before the first object
5409 [Richard Shapiro <rshapiro@abinitio.com>]
5411 *) [In 0.9.6c-engine release:]
5412 Add support for Broadcom crypto accelerator cards, backported
5414 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
5416 *) [In 0.9.6c-engine release:]
5417 Add support for SureWare crypto accelerator cards from
5418 Baltimore Technologies. (Use engine 'sureware')
5419 [Baltimore Technologies and Mark Cox]
5421 *) [In 0.9.6c-engine release:]
5422 Add support for crypto accelerator cards from Accelerated
5423 Encryption Processing, www.aep.ie. (Use engine 'aep')
5424 [AEP Inc. and Mark Cox]
5426 *) Add a configuration entry for gcc on UnixWare.
5427 [Gary Benson <gbenson@redhat.com>]
5429 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
5430 messages are stored in a single piece (fixed-length part and
5431 variable-length part combined) and fix various bugs found on the way.
5434 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
5435 instead. BIO_gethostbyname() does not know what timeouts are
5436 appropriate, so entries would stay in cache even when they have
5438 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
5440 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
5441 faced with a pathologically small ClientHello fragment that does
5442 not contain client_version: Instead of aborting with an error,
5443 simply choose the highest available protocol version (i.e.,
5444 TLS 1.0 unless it is disabled). In practice, ClientHello
5445 messages are never sent like this, but this change gives us
5446 strictly correct behaviour at least for TLS.
5449 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
5450 never resets s->method to s->ctx->method when called from within
5451 one of the SSL handshake functions.
5452 [Bodo Moeller; problem pointed out by Niko Baric]
5454 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
5455 (sent using the client's version number) if client_version is
5456 smaller than the protocol version in use. Also change
5457 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
5458 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
5459 the client will at least see that alert.
5462 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
5466 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
5467 client receives HelloRequest while in a handshake.
5468 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
5470 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
5471 should end in 'break', not 'goto end' which circuments various
5472 cleanups done in state SSL_ST_OK. But session related stuff
5473 must be disabled for SSL_ST_OK in the case that we just sent a
5476 Also avoid some overhead by not calling ssl_init_wbio_buffer()
5477 before just sending a HelloRequest.
5478 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
5480 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
5481 reveal whether illegal block cipher padding was found or a MAC
5482 verification error occured. (Neither SSLerr() codes nor alerts
5483 are directly visible to potential attackers, but the information
5484 may leak via logfiles.)
5486 Similar changes are not required for the SSL 2.0 implementation
5487 because the number of padding bytes is sent in clear for SSL 2.0,
5488 and the extra bytes are just ignored. However ssl/s2_pkt.c
5489 failed to verify that the purported number of padding bytes is in
5493 *) Add OpenUNIX-8 support including shared libraries
5494 (Boyd Lynn Gerber <gerberb@zenez.com>).
5497 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
5498 'wristwatch attack' using huge encoding parameters (cf.
5499 James H. Manger's CRYPTO 2001 paper). Note that the
5500 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
5501 encoding parameters and hence was not vulnerable.
5504 *) BN_sqr() bug fix.
5505 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
5507 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
5508 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
5509 followed by modular reduction.
5510 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
5512 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
5513 equivalent based on BN_pseudo_rand() instead of BN_rand().
5516 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
5517 This function was broken, as the check for a new client hello message
5518 to handle SGC did not allow these large messages.
5519 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
5522 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
5525 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
5526 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
5529 *) Rework the configuration and shared library support for Tru64 Unix.
5530 The configuration part makes use of modern compiler features and
5531 still retains old compiler behavior for those that run older versions
5532 of the OS. The shared library support part includes a variant that
5533 uses the RPATH feature, and is available through the special
5534 configuration target "alpha-cc-rpath", which will never be selected
5536 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
5538 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
5539 with the same message size as in ssl3_get_certificate_request().
5540 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
5541 messages might inadvertently be reject as too long.
5542 [Petr Lampa <lampa@fee.vutbr.cz>]
5544 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
5547 *) Modified SSL library such that the verify_callback that has been set
5548 specificly for an SSL object with SSL_set_verify() is actually being
5549 used. Before the change, a verify_callback set with this function was
5550 ignored and the verify_callback() set in the SSL_CTX at the time of
5551 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
5552 to allow the necessary settings.
5555 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
5556 explicitly to NULL, as at least on Solaris 8 this seems not always to be
5557 done automatically (in contradiction to the requirements of the C
5558 standard). This made problems when used from OpenSSH.
5561 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
5562 dh->length and always used
5564 BN_rand_range(priv_key, dh->p).
5566 BN_rand_range() is not necessary for Diffie-Hellman, and this
5567 specific range makes Diffie-Hellman unnecessarily inefficient if
5568 dh->length (recommended exponent length) is much smaller than the
5569 length of dh->p. We could use BN_rand_range() if the order of
5570 the subgroup was stored in the DH structure, but we only have
5575 BN_rand(priv_key, l, ...)
5577 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
5583 RSA_eay_public_encrypt
5584 RSA_eay_private_decrypt
5585 RSA_eay_private_encrypt (signing)
5586 RSA_eay_public_decrypt (signature verification)
5588 (default implementations for RSA_public_encrypt,
5589 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
5590 always reject numbers >= n.
5593 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
5594 to synchronize access to 'locking_thread'. This is necessary on
5595 systems where access to 'locking_thread' (an 'unsigned long'
5596 variable) is not atomic.
5599 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
5600 *before* setting the 'crypto_lock_rand' flag. The previous code had
5601 a race condition if 0 is a valid thread ID.
5602 [Travis Vitek <vitek@roguewave.com>]
5604 *) Add support for shared libraries under Irix.
5605 [Albert Chin-A-Young <china@thewrittenword.com>]
5607 *) Add configuration option to build on Linux on both big-endian and
5609 [Ralf Baechle <ralf@uni-koblenz.de>]
5611 *) Add the possibility to create shared libraries on HP-UX.
5614 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
5616 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
5617 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
5618 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
5619 PRNG state recovery was possible based on the output of
5620 one PRNG request appropriately sized to gain knowledge on
5621 'md' followed by enough consecutive 1-byte PRNG requests
5622 to traverse all of 'state'.
5624 1. When updating 'md_local' (the current thread's copy of 'md')
5625 during PRNG output generation, hash all of the previous
5626 'md_local' value, not just the half used for PRNG output.
5628 2. Make the number of bytes from 'state' included into the hash
5629 independent from the number of PRNG bytes requested.
5631 The first measure alone would be sufficient to avoid
5632 Markku-Juhani's attack. (Actually it had never occurred
5633 to me that the half of 'md_local' used for chaining was the
5634 half from which PRNG output bytes were taken -- I had always
5635 assumed that the secret half would be used.) The second
5636 measure makes sure that additional data from 'state' is never
5637 mixed into 'md_local' in small portions; this heuristically
5638 further strengthens the PRNG.
5641 *) Fix crypto/bn/asm/mips3.s.
5644 *) When only the key is given to "enc", the IV is undefined. Print out
5645 an error message in this case.
5648 *) Handle special case when X509_NAME is empty in X509 printing routines.
5651 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
5652 positive and less than q.
5655 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
5656 used: it isn't thread safe and the add_lock_callback should handle
5658 [Paul Rose <Paul.Rose@bridge.com>]
5660 *) Verify that incoming data obeys the block size in
5661 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
5665 [Ulf Möller, Bodo Möller]
5667 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
5668 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
5669 when fixing the server behaviour for backwards-compatible 'client
5670 hello' messages. (Note that the attack is impractical against
5671 SSL 3.0 and TLS 1.0 anyway because length and version checking
5672 means that the probability of guessing a valid ciphertext is
5673 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
5676 Before 0.9.5, the countermeasure (hide the error by generating a
5677 random 'decryption result') did not work properly because
5678 ERR_clear_error() was missing, meaning that SSL_get_error() would
5679 detect the supposedly ignored error.
5681 Both problems are now fixed.
5684 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
5685 (previously it was 1024).
5688 *) Fix for compatibility mode trust settings: ignore trust settings
5689 unless some valid trust or reject settings are present.
5692 *) Fix for blowfish EVP: its a variable length cipher.
5695 *) Fix various bugs related to DSA S/MIME verification. Handle missing
5696 parameters in DSA public key structures and return an error in the
5697 DSA routines if parameters are absent.
5700 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
5701 in the current directory if neither $RANDFILE nor $HOME was set.
5702 RAND_file_name() in 0.9.6a returned NULL in this case. This has
5703 caused some confusion to Windows users who haven't defined $HOME.
5704 Thus RAND_file_name() is changed again: e_os.h can define a
5705 DEFAULT_HOME, which will be used if $HOME is not set.
5706 For Windows, we use "C:"; on other platforms, we still require
5707 environment variables.
5709 *) Move 'if (!initialized) RAND_poll()' into regions protected by
5710 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
5711 having multiple threads call RAND_poll() concurrently.
5714 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
5715 combination of a flag and a thread ID variable.
5716 Otherwise while one thread is in ssleay_rand_bytes (which sets the
5717 flag), *other* threads can enter ssleay_add_bytes without obeying
5718 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
5719 that they do not hold after the first thread unsets add_do_not_lock).
5722 *) Change bctest again: '-x' expressions are not available in all
5726 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
5728 *) Fix a couple of memory leaks in PKCS7_dataDecode()
5729 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
5731 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
5732 the default extension for executables, if any. Also, make the perl
5733 scripts that use symlink() to test if it really exists and use "cp"
5734 if it doesn't. All this made OpenSSL compilable and installable in
5738 *) Fix for asn1_GetSequence() for indefinite length constructed data.
5739 If SEQUENCE is length is indefinite just set c->slen to the total
5740 amount of data available.
5741 [Steve Henson, reported by shige@FreeBSD.org]
5742 [This change does not apply to 0.9.7.]
5744 *) Change bctest to avoid here-documents inside command substitution
5745 (workaround for FreeBSD /bin/sh bug).
5746 For compatibility with Ultrix, avoid shell functions (introduced
5747 in the bctest version that searches along $PATH).
5750 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
5751 with des_encrypt() defined on some operating systems, like Solaris
5755 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
5756 On the Importance of Eliminating Errors in Cryptographic
5757 Computations, J. Cryptology 14 (2001) 2, 101-119,
5758 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
5761 *) MIPS assembler BIGNUM division bug fix.
5764 *) Disabled incorrect Alpha assembler code.
5767 *) Fix PKCS#7 decode routines so they correctly update the length
5768 after reading an EOC for the EXPLICIT tag.
5770 [This change does not apply to 0.9.7.]
5772 *) Fix bug in PKCS#12 key generation routines. This was triggered
5773 if a 3DES key was generated with a 0 initial byte. Include
5774 PKCS12_BROKEN_KEYGEN compilation option to retain the old
5775 (but broken) behaviour.
5778 *) Enhance bctest to search for a working bc along $PATH and print
5780 [Tim Rice <tim@multitalents.net> via Richard Levitte]
5782 *) Fix memory leaks in err.c: free err_data string if necessary;
5783 don't write to the wrong index in ERR_set_error_data.
5786 *) Implement ssl23_peek (analogous to ssl23_read), which previously
5790 *) Replace rdtsc with _emit statements for VC++ version 5.
5791 [Jeremy Cooper <jeremy@baymoo.org>]
5793 *) Make it possible to reuse SSLv2 sessions.
5796 *) In copy_email() check for >= 0 as a return value for
5797 X509_NAME_get_index_by_NID() since 0 is a valid index.
5798 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
5800 *) Avoid coredump with unsupported or invalid public keys by checking if
5801 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
5802 PKCS7_verify() fails with non detached data.
5805 *) Don't use getenv in library functions when run as setuid/setgid.
5806 New function OPENSSL_issetugid().
5809 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
5810 due to incorrect handling of multi-threading:
5812 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
5814 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
5816 3. Count how many times MemCheck_off() has been called so that
5817 nested use can be treated correctly. This also avoids
5818 inband-signalling in the previous code (which relied on the
5819 assumption that thread ID 0 is impossible).
5822 *) Add "-rand" option also to s_client and s_server.
5825 *) Fix CPU detection on Irix 6.x.
5826 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
5827 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
5829 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
5832 [This change does not apply to 0.9.7.]
5834 *) Use the cached encoding of an X509_NAME structure rather than
5835 copying it. This is apparently the reason for the libsafe "errors"
5836 but the code is actually correct.
5839 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
5840 Bleichenbacher's DSA attack.
5841 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
5842 to be set and top=0 forces the highest bit to be set; top=-1 is new
5843 and leaves the highest bit random.
5844 [Ulf Moeller, Bodo Moeller]
5846 *) In the NCONF_...-based implementations for CONF_... queries
5847 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
5848 a temporary CONF structure with the data component set to NULL
5849 (which gives segmentation faults in lh_retrieve).
5850 Instead, use NULL for the CONF pointer in CONF_get_string and
5851 CONF_get_number (which may use environment variables) and directly
5852 return NULL from CONF_get_section.
5855 *) Fix potential buffer overrun for EBCDIC.
5858 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
5859 keyUsage if basicConstraints absent for a CA.
5862 *) Make SMIME_write_PKCS7() write mail header values with a format that
5863 is more generally accepted (no spaces before the semicolon), since
5864 some programs can't parse those values properly otherwise. Also make
5865 sure BIO's that break lines after each write do not create invalid
5869 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
5870 macros previously used would not encode an empty SEQUENCE OF
5871 and break the signature.
5873 [This change does not apply to 0.9.7.]
5875 *) Zero the premaster secret after deriving the master secret in
5879 *) Add some EVP_add_digest_alias registrations (as found in
5880 OpenSSL_add_all_digests()) to SSL_library_init()
5881 aka OpenSSL_add_ssl_algorithms(). This provides improved
5882 compatibility with peers using X.509 certificates
5883 with unconventional AlgorithmIdentifier OIDs.
5886 *) Fix for Irix with NO_ASM.
5887 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
5889 *) ./config script fixes.
5890 [Ulf Moeller, Richard Levitte]
5892 *) Fix 'openssl passwd -1'.
5895 *) Change PKCS12_key_gen_asc() so it can cope with non null
5896 terminated strings whose length is passed in the passlen
5897 parameter, for example from PEM callbacks. This was done
5898 by adding an extra length parameter to asc2uni().
5899 [Steve Henson, reported by <oddissey@samsung.co.kr>]
5901 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
5902 call failed, free the DSA structure.
5905 *) Fix to uni2asc() to cope with zero length Unicode strings.
5906 These are present in some PKCS#12 files.
5909 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
5910 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
5911 when writing a 32767 byte record.
5912 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
5914 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
5915 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
5917 (RSA objects have a reference count access to which is protected
5918 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
5919 so they are meant to be shared between threads.)
5920 [Bodo Moeller, Geoff Thorpe; original patch submitted by
5921 "Reddie, Steven" <Steven.Reddie@ca.com>]
5923 *) Fix a deadlock in CRYPTO_mem_leaks().
5926 *) Use better test patterns in bntest.
5929 *) rand_win.c fix for Borland C.
5932 *) BN_rshift bugfix for n == 0.
5935 *) Add a 'bctest' script that checks for some known 'bc' bugs
5936 so that 'make test' does not abort just because 'bc' is broken.
5939 *) Store verify_result within SSL_SESSION also for client side to
5940 avoid potential security hole. (Re-used sessions on the client side
5941 always resulted in verify_result==X509_V_OK, not using the original
5942 result of the server certificate verification.)
5945 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
5946 SSL3_RT_APPLICATION_DATA, return 0.
5947 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
5951 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
5952 releases, have been re-implemented by renaming the previous
5953 implementations of ssl2_read and ssl3_read to ssl2_read_internal
5954 and ssl3_read_internal, respectively, and adding 'peek' parameters
5955 to them. The new ssl[23]_{read,peek} functions are calls to
5956 ssl[23]_read_internal with the 'peek' flag set appropriately.
5957 A 'peek' parameter has also been added to ssl3_read_bytes, which
5958 does the actual work for ssl3_read_internal.
5961 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
5962 the method-specific "init()" handler. Also clean up ex_data after
5963 calling the method-specific "finish()" handler. Previously, this was
5964 happening the other way round.
5967 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
5968 The previous value, 12, was not always sufficient for BN_mod_exp().
5971 *) Make sure that shared libraries get the internal name engine with
5972 the full version number and not just 0. This should mark the
5973 shared libraries as not backward compatible. Of course, this should
5974 be changed again when we can guarantee backward binary compatibility.
5977 *) Fix typo in get_cert_by_subject() in by_dir.c
5978 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
5980 *) Rework the system to generate shared libraries:
5982 - Make note of the expected extension for the shared libraries and
5983 if there is a need for symbolic links from for example libcrypto.so.0
5984 to libcrypto.so.0.9.7. There is extended info in Configure for
5987 - Make as few rebuilds of the shared libraries as possible.
5989 - Still avoid linking the OpenSSL programs with the shared libraries.
5991 - When installing, install the shared libraries separately from the
5995 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
5997 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
5998 and not in SSL_clear because the latter is also used by the
5999 accept/connect functions; previously, the settings made by
6000 SSL_set_read_ahead would be lost during the handshake.
6001 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6003 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6004 Previously, it would create entries for disableed algorithms no
6008 *) Added several new manual pages for SSL_* function.
6011 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6013 *) In ssl23_get_client_hello, generate an error message when faced
6014 with an initial SSL 3.0/TLS record that is too small to contain the
6015 first two bytes of the ClientHello message, i.e. client_version.
6016 (Note that this is a pathologic case that probably has never happened
6017 in real life.) The previous approach was to use the version number
6018 from the record header as a substitute; but our protocol choice
6019 should not depend on that one because it is not authenticated
6020 by the Finished messages.
6023 *) More robust randomness gathering functions for Windows.
6024 [Jeffrey Altman <jaltman@columbia.edu>]
6026 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
6027 not set then we don't setup the error code for issuer check errors
6028 to avoid possibly overwriting other errors which the callback does
6029 handle. If an application does set the flag then we assume it knows
6030 what it is doing and can handle the new informational codes
6034 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
6035 a general "ANY" type, as such it should be able to decode anything
6036 including tagged types. However it didn't check the class so it would
6037 wrongly interpret tagged types in the same way as their universal
6038 counterpart and unknown types were just rejected. Changed so that the
6039 tagged and unknown types are handled in the same way as a SEQUENCE:
6040 that is the encoding is stored intact. There is also a new type
6041 "V_ASN1_OTHER" which is used when the class is not universal, in this
6042 case we have no idea what the actual type is so we just lump them all
6046 *) On VMS, stdout may very well lead to a file that is written to
6047 in a record-oriented fashion. That means that every write() will
6048 write a separate record, which will be read separately by the
6049 programs trying to read from it. This can be very confusing.
6051 The solution is to put a BIO filter in the way that will buffer
6052 text until a linefeed is reached, and then write everything a
6053 line at a time, so every record written will be an actual line,
6054 not chunks of lines and not (usually doesn't happen, but I've
6055 seen it once) several lines in one record. BIO_f_linebuffer() is
6058 Currently, it's a VMS-only method, because that's where it has
6059 been tested well enough.
6062 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
6063 it can return incorrect results.
6064 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
6065 but it was in 0.9.6-beta[12].)
6068 *) Disable the check for content being present when verifying detached
6069 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
6070 include zero length content when signing messages.
6073 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
6074 BIO_ctrl (for BIO pairs).
6077 *) Add DSO method for VMS.
6080 *) Bug fix: Montgomery multiplication could produce results with the
6084 *) Add RPM specification openssl.spec and modify it to build three
6085 packages. The default package contains applications, application
6086 documentation and run-time libraries. The devel package contains
6087 include files, static libraries and function documentation. The
6088 doc package contains the contents of the doc directory. The original
6089 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
6092 *) Add a large number of documentation files for many SSL routines.
6093 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6095 *) Add a configuration entry for Sony News 4.
6096 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
6098 *) Don't set the two most significant bits to one when generating a
6099 random number < q in the DSA library.
6102 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
6103 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
6104 the underlying transport is blocking) if a handshake took place.
6105 (The default behaviour is needed by applications such as s_client
6106 and s_server that use select() to determine when to use SSL_read;
6107 but for applications that know in advance when to expect data, it
6108 just makes things more complicated.)
6111 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
6115 *) Add a few more EBCDIC conditionals that make `req' and `x509'
6116 work better on such systems.
6117 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6119 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
6120 Update PKCS12_parse() so it copies the friendlyName and the
6121 keyid to the certificates aux info.
6124 *) Fix bug in PKCS7_verify() which caused an infinite loop
6125 if there was more than one signature.
6126 [Sven Uszpelkat <su@celocom.de>]
6128 *) Major change in util/mkdef.pl to include extra information
6129 about each symbol, as well as presentig variables as well
6130 as functions. This change means that there's n more need
6131 to rebuild the .num files when some algorithms are excluded.
6134 *) Allow the verify time to be set by an application,
6135 rather than always using the current time.
6138 *) Phase 2 verify code reorganisation. The certificate
6139 verify code now looks up an issuer certificate by a
6140 number of criteria: subject name, authority key id
6141 and key usage. It also verifies self signed certificates
6142 by the same criteria. The main comparison function is
6143 X509_check_issued() which performs these checks.
6145 Lot of changes were necessary in order to support this
6146 without completely rewriting the lookup code.
6148 Authority and subject key identifier are now cached.
6150 The LHASH 'certs' is X509_STORE has now been replaced
6151 by a STACK_OF(X509_OBJECT). This is mainly because an
6152 LHASH can't store or retrieve multiple objects with
6153 the same hash value.
6155 As a result various functions (which were all internal
6156 use only) have changed to handle the new X509_STORE
6157 structure. This will break anything that messed round
6158 with X509_STORE internally.
6160 The functions X509_STORE_add_cert() now checks for an
6161 exact match, rather than just subject name.
6163 The X509_STORE API doesn't directly support the retrieval
6164 of multiple certificates matching a given criteria, however
6165 this can be worked round by performing a lookup first
6166 (which will fill the cache with candidate certificates)
6167 and then examining the cache for matches. This is probably
6168 the best we can do without throwing out X509_LOOKUP
6169 entirely (maybe later...).
6171 The X509_VERIFY_CTX structure has been enhanced considerably.
6173 All certificate lookup operations now go via a get_issuer()
6174 callback. Although this currently uses an X509_STORE it
6175 can be replaced by custom lookups. This is a simple way
6176 to bypass the X509_STORE hackery necessary to make this
6177 work and makes it possible to use more efficient techniques
6178 in future. A very simple version which uses a simple
6179 STACK for its trusted certificate store is also provided
6180 using X509_STORE_CTX_trusted_stack().
6182 The verify_cb() and verify() callbacks now have equivalents
6183 in the X509_STORE_CTX structure.
6185 X509_STORE_CTX also has a 'flags' field which can be used
6186 to customise the verify behaviour.
6189 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
6190 excludes S/MIME capabilities.
6193 *) When a certificate request is read in keep a copy of the
6194 original encoding of the signed data and use it when outputing
6195 again. Signatures then use the original encoding rather than
6196 a decoded, encoded version which may cause problems if the
6197 request is improperly encoded.
6200 *) For consistency with other BIO_puts implementations, call
6201 buffer_write(b, ...) directly in buffer_puts instead of calling
6204 In BIO_puts, increment b->num_write as in BIO_write.
6205 [Peter.Sylvester@EdelWeb.fr]
6207 *) Fix BN_mul_word for the case where the word is 0. (We have to use
6208 BN_zero, we may not return a BIGNUM with an array consisting of
6212 *) Avoid calling abort() from within the library when problems are
6213 detected, except if preprocessor symbols have been defined
6214 (such as REF_CHECK, BN_DEBUG etc.).
6217 *) New openssl application 'rsautl'. This utility can be
6218 used for low level RSA operations. DER public key
6219 BIO/fp routines also added.
6222 *) New Configure entry and patches for compiling on QNX 4.
6223 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
6225 *) A demo state-machine implementation was sponsored by
6226 Nuron (http://www.nuron.com/) and is now available in
6227 demos/state_machine.
6230 *) New options added to the 'dgst' utility for signature
6231 generation and verification.
6234 *) Unrecognized PKCS#7 content types are now handled via a
6235 catch all ASN1_TYPE structure. This allows unsupported
6236 types to be stored as a "blob" and an application can
6237 encode and decode it manually.
6240 *) Fix various signed/unsigned issues to make a_strex.c
6242 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
6244 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
6245 length if passed a buffer. ASN1_INTEGER_to_BN failed
6246 if passed a NULL BN and its argument was negative.
6247 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
6249 *) Modification to PKCS#7 encoding routines to output definite
6250 length encoding. Since currently the whole structures are in
6251 memory there's not real point in using indefinite length
6252 constructed encoding. However if OpenSSL is compiled with
6253 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
6256 *) Added BIO_vprintf() and BIO_vsnprintf().
6259 *) Added more prefixes to parse for in the the strings written
6260 through a logging bio, to cover all the levels that are available
6261 through syslog. The prefixes are now:
6263 PANIC, EMERG, EMR => LOG_EMERG
6264 ALERT, ALR => LOG_ALERT
6265 CRIT, CRI => LOG_CRIT
6266 ERROR, ERR => LOG_ERR
6267 WARNING, WARN, WAR => LOG_WARNING
6268 NOTICE, NOTE, NOT => LOG_NOTICE
6269 INFO, INF => LOG_INFO
6270 DEBUG, DBG => LOG_DEBUG
6272 and as before, if none of those prefixes are present at the
6273 beginning of the string, LOG_ERR is chosen.
6275 On Win32, the LOG_* levels are mapped according to this:
6277 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
6278 LOG_WARNING => EVENTLOG_WARNING_TYPE
6279 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
6283 *) Made it possible to reconfigure with just the configuration
6284 argument "reconf" or "reconfigure". The command line arguments
6285 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
6286 and are retrieved from there when reconfiguring.
6290 [Assar Westerlund <assar@sics.se>, Richard Levitte]
6292 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
6295 *) The obj_dat.pl script was messing up the sorting of object
6296 names. The reason was that it compared the quoted version
6297 of strings as a result "OCSP" > "OCSP Signing" because
6298 " > SPACE. Changed script to store unquoted versions of
6299 names and add quotes on output. It was also omitting some
6300 names from the lookup table if they were given a default
6301 value (that is if SN is missing it is given the same
6302 value as LN and vice versa), these are now added on the
6303 grounds that if an object has a name we should be able to
6304 look it up. Finally added warning output when duplicate
6305 short or long names are found.
6308 *) Changes needed for Tandem NSK.
6309 [Scott Uroff <scott@xypro.com>]
6311 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
6312 RSA_padding_check_SSLv23(), special padding was never detected
6313 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
6314 version rollback attacks was not effective.
6316 In s23_clnt.c, don't use special rollback-attack detection padding
6317 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
6318 client; similarly, in s23_srvr.c, don't do the rollback check if
6319 SSL 2.0 is the only protocol enabled in the server.
6322 *) Make it possible to get hexdumps of unprintable data with 'openssl
6323 asn1parse'. By implication, the functions ASN1_parse_dump() and
6324 BIO_dump_indent() are added.
6327 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
6328 these print out strings and name structures based on various
6329 flags including RFC2253 support and proper handling of
6330 multibyte characters. Added options to the 'x509' utility
6331 to allow the various flags to be set.
6334 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
6335 Also change the functions X509_cmp_current_time() and
6336 X509_gmtime_adj() work with an ASN1_TIME structure,
6337 this will enable certificates using GeneralizedTime in validity
6338 dates to be checked.
6341 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
6342 negative public key encodings) on by default,
6343 NO_NEG_PUBKEY_BUG can be set to disable it.
6346 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
6347 content octets. An i2c_ASN1_OBJECT is unnecessary because
6348 the encoding can be trivially obtained from the structure.
6351 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
6352 not read locks (CRYPTO_r_[un]lock).
6355 *) A first attempt at creating official support for shared
6356 libraries through configuration. I've kept it so the
6357 default is static libraries only, and the OpenSSL programs
6358 are always statically linked for now, but there are
6359 preparations for dynamic linking in place.
6360 This has been tested on Linux and Tru64.
6363 *) Randomness polling function for Win9x, as described in:
6364 Peter Gutmann, Software Generation of Practically Strong
6368 *) Fix so PRNG is seeded in req if using an already existing
6372 *) New options to smime application. -inform and -outform
6373 allow alternative formats for the S/MIME message including
6374 PEM and DER. The -content option allows the content to be
6375 specified separately. This should allow things like Netscape
6376 form signing output easier to verify.
6379 *) Fix the ASN1 encoding of tags using the 'long form'.
6382 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
6383 STRING types. These convert content octets to and from the
6384 underlying type. The actual tag and length octets are
6385 already assumed to have been read in and checked. These
6386 are needed because all other string types have virtually
6387 identical handling apart from the tag. By having versions
6388 of the ASN1 functions that just operate on content octets
6389 IMPLICIT tagging can be handled properly. It also allows
6390 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
6391 and ASN1_INTEGER are identical apart from the tag.
6394 *) Change the handling of OID objects as follows:
6396 - New object identifiers are inserted in objects.txt, following
6397 the syntax given in objects.README.
6398 - objects.pl is used to process obj_mac.num and create a new
6400 - obj_dat.pl is used to create a new obj_dat.h, using the data in
6403 This is currently kind of a hack, and the perl code in objects.pl
6404 isn't very elegant, but it works as I intended. The simplest way
6405 to check that it worked correctly is to look in obj_dat.h and
6406 check the array nid_objs and make sure the objects haven't moved
6407 around (this is important!). Additions are OK, as well as
6408 consistent name changes.
6411 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
6414 *) Addition of the command line parameter '-rand file' to 'openssl req'.
6415 The given file adds to whatever has already been seeded into the
6416 random pool through the RANDFILE configuration file option or
6417 environment variable, or the default random state file.
6420 *) mkstack.pl now sorts each macro group into lexical order.
6421 Previously the output order depended on the order the files
6422 appeared in the directory, resulting in needless rewriting
6426 *) Patches to make OpenSSL compile under Win32 again. Mostly
6427 work arounds for the VC++ problem that it treats func() as
6428 func(void). Also stripped out the parts of mkdef.pl that
6429 added extra typesafe functions: these no longer exist.
6432 *) Reorganisation of the stack code. The macros are now all
6433 collected in safestack.h . Each macro is defined in terms of
6434 a "stack macro" of the form SKM_<name>(type, a, b). The
6435 DEBUG_SAFESTACK is now handled in terms of function casts,
6436 this has the advantage of retaining type safety without the
6437 use of additional functions. If DEBUG_SAFESTACK is not defined
6438 then the non typesafe macros are used instead. Also modified the
6439 mkstack.pl script to handle the new form. Needs testing to see
6440 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
6441 the default if no major problems. Similar behaviour for ASN1_SET_OF
6442 and PKCS12_STACK_OF.
6445 *) When some versions of IIS use the 'NET' form of private key the
6446 key derivation algorithm is different. Normally MD5(password) is
6447 used as a 128 bit RC4 key. In the modified case
6448 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
6449 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
6450 as the old Netscape_RSA functions except they have an additional
6451 'sgckey' parameter which uses the modified algorithm. Also added
6452 an -sgckey command line option to the rsa utility. Thanks to
6453 Adrian Peck <bertie@ncipher.com> for posting details of the modified
6454 algorithm to openssl-dev.
6457 *) The evp_local.h macros were using 'c.##kname' which resulted in
6458 invalid expansion on some systems (SCO 5.0.5 for example).
6459 Corrected to 'c.kname'.
6460 [Phillip Porch <root@theporch.com>]
6462 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
6463 a STACK of email addresses from a certificate or request, these look
6464 in the subject name and the subject alternative name extensions and
6465 omit any duplicate addresses.
6468 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
6469 This makes DSA verification about 2 % faster.
6472 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
6473 (meaning that now 2^5 values will be precomputed, which is only 4 KB
6474 plus overhead for 1024 bit moduli).
6475 This makes exponentiations about 0.5 % faster for 1024 bit
6476 exponents (as measured by "openssl speed rsa2048").
6479 *) Rename memory handling macros to avoid conflicts with other
6481 Malloc => OPENSSL_malloc
6482 Malloc_locked => OPENSSL_malloc_locked
6483 Realloc => OPENSSL_realloc
6484 Free => OPENSSL_free
6487 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
6488 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
6491 *) CygWin32 support.
6492 [John Jarvie <jjarvie@newsguy.com>]
6494 *) The type-safe stack code has been rejigged. It is now only compiled
6495 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
6496 by default all type-specific stack functions are "#define"d back to
6497 standard stack functions. This results in more streamlined output
6498 but retains the type-safety checking possibilities of the original
6502 *) The STACK code has been cleaned up, and certain type declarations
6503 that didn't make a lot of sense have been brought in line. This has
6504 also involved a cleanup of sorts in safestack.h to more correctly
6505 map type-safe stack functions onto their plain stack counterparts.
6506 This work has also resulted in a variety of "const"ifications of
6507 lots of the code, especially "_cmp" operations which should normally
6508 be prototyped with "const" parameters anyway.
6511 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
6512 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
6513 (The PRNG state consists of two parts, the large pool 'state' and 'md',
6514 where all of 'md' is used each time the PRNG is used, but 'state'
6515 is used only indexed by a cyclic counter. As entropy may not be
6516 well distributed from the beginning, 'md' is important as a
6517 chaining variable. However, the output function chains only half
6518 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
6519 all of 'md', and seeding with STATE_SIZE dummy bytes will result
6520 in all of 'state' being rewritten, with the new values depending
6521 on virtually all of 'md'. This overcomes the 80 bit limitation.)
6524 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
6525 the handshake is continued after ssl_verify_cert_chain();
6526 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
6527 can lead to 'unexplainable' connection aborts later.
6528 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
6530 *) Major EVP API cipher revision.
6531 Add hooks for extra EVP features. This allows various cipher
6532 parameters to be set in the EVP interface. Support added for variable
6533 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
6534 setting of RC2 and RC5 parameters.
6536 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
6539 Remove lots of duplicated code from the EVP library. For example *every*
6540 cipher init() function handles the 'iv' in the same way according to the
6541 cipher mode. They also all do nothing if the 'key' parameter is NULL and
6542 for CFB and OFB modes they zero ctx->num.
6544 New functionality allows removal of S/MIME code RC2 hack.
6546 Most of the routines have the same form and so can be declared in terms
6549 By shifting this to the top level EVP_CipherInit() it can be removed from
6550 all individual ciphers. If the cipher wants to handle IVs or keys
6551 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
6554 Change lots of functions like EVP_EncryptUpdate() to now return a
6555 value: although software versions of the algorithms cannot fail
6556 any installed hardware versions can.
6559 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
6560 this option is set, tolerate broken clients that send the negotiated
6561 protocol version number instead of the requested protocol version
6565 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
6566 i.e. non-zero for export ciphersuites, zero otherwise.
6567 Previous versions had this flag inverted, inconsistent with
6568 rsa_tmp_cb (..._TMP_RSA_CB).
6569 [Bodo Moeller; problem reported by Amit Chopra]
6571 *) Add missing DSA library text string. Work around for some IIS
6572 key files with invalid SEQUENCE encoding.
6575 *) Add a document (doc/standards.txt) that list all kinds of standards
6576 and so on that are implemented in OpenSSL.
6579 *) Enhance c_rehash script. Old version would mishandle certificates
6580 with the same subject name hash and wouldn't handle CRLs at all.
6581 Added -fingerprint option to crl utility, to support new c_rehash
6585 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
6588 *) Fix for SSL server purpose checking. Server checking was
6589 rejecting certificates which had extended key usage present
6590 but no ssl client purpose.
6591 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
6593 *) Make PKCS#12 code work with no password. The PKCS#12 spec
6594 is a little unclear about how a blank password is handled.
6595 Since the password in encoded as a BMPString with terminating
6596 double NULL a zero length password would end up as just the
6597 double NULL. However no password at all is different and is
6598 handled differently in the PKCS#12 key generation code. NS
6599 treats a blank password as zero length. MSIE treats it as no
6600 password on export: but it will try both on import. We now do
6601 the same: PKCS12_parse() tries zero length and no password if
6602 the password is set to "" or NULL (NULL is now a valid password:
6603 it wasn't before) as does the pkcs12 application.
6606 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
6607 perror when PEM_read_bio_X509_REQ fails, the error message must
6608 be obtained from the error queue.
6611 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
6612 it in ERR_remove_state if appropriate, and change ERR_get_state
6613 accordingly to avoid race conditions (this is necessary because
6614 thread_hash is no longer constant once set).
6617 *) Bugfix for linux-elf makefile.one.
6620 *) RSA_get_default_method() will now cause a default
6621 RSA_METHOD to be chosen if one doesn't exist already.
6622 Previously this was only set during a call to RSA_new()
6623 or RSA_new_method(NULL) meaning it was possible for
6624 RSA_get_default_method() to return NULL.
6627 *) Added native name translation to the existing DSO code
6628 that will convert (if the flag to do so is set) filenames
6629 that are sufficiently small and have no path information
6630 into a canonical native form. Eg. "blah" converted to
6631 "libblah.so" or "blah.dll" etc.
6634 *) New function ERR_error_string_n(e, buf, len) which is like
6635 ERR_error_string(e, buf), but writes at most 'len' bytes
6636 including the 0 terminator. For ERR_error_string_n, 'buf'
6638 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
6640 *) CONF library reworked to become more general. A new CONF
6641 configuration file reader "class" is implemented as well as a
6642 new functions (NCONF_*, for "New CONF") to handle it. The now
6643 old CONF_* functions are still there, but are reimplemented to
6644 work in terms of the new functions. Also, a set of functions
6645 to handle the internal storage of the configuration data is
6646 provided to make it easier to write new configuration file
6647 reader "classes" (I can definitely see something reading a
6648 configuration file in XML format, for example), called _CONF_*,
6649 or "the configuration storage API"...
6651 The new configuration file reading functions are:
6653 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
6654 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
6656 NCONF_default, NCONF_WIN32
6658 NCONF_dump_fp, NCONF_dump_bio
6660 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
6661 NCONF_new creates a new CONF object. This works in the same way
6662 as other interfaces in OpenSSL, like the BIO interface.
6663 NCONF_dump_* dump the internal storage of the configuration file,
6664 which is useful for debugging. All other functions take the same
6665 arguments as the old CONF_* functions wth the exception of the
6666 first that must be a `CONF *' instead of a `LHASH *'.
6668 To make it easer to use the new classes with the old CONF_* functions,
6669 the function CONF_set_default_method is provided.
6672 *) Add '-tls1' option to 'openssl ciphers', which was already
6673 mentioned in the documentation but had not been implemented.
6674 (This option is not yet really useful because even the additional
6675 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
6678 *) Initial DSO code added into libcrypto for letting OpenSSL (and
6679 OpenSSL-based applications) load shared libraries and bind to
6680 them in a portable way.
6681 [Geoff Thorpe, with contributions from Richard Levitte]
6683 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
6685 *) Make sure _lrotl and _lrotr are only used with MSVC.
6687 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
6688 (the default implementation of RAND_status).
6690 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
6691 to '-clrext' (= clear extensions), as intended and documented.
6692 [Bodo Moeller; inconsistency pointed out by Michael Attili
6695 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
6696 was larger than the MD block size.
6697 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
6699 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
6700 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
6701 using the passed key: if the passed key was a private key the result
6702 of X509_print(), for example, would be to print out all the private key
6706 *) des_quad_cksum() byte order bug fix.
6707 [Ulf Möller, using the problem description in krb4-0.9.7, where
6708 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
6710 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
6712 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
6714 *) For easily testing in shell scripts whether some command
6715 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
6716 returns with exit code 0 iff no command of the given name is available.
6717 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
6718 the output goes to stdout and nothing is printed to stderr.
6719 Additional arguments are always ignored.
6721 Since for each cipher there is a command of the same name,
6722 the 'no-cipher' compilation switches can be tested this way.
6724 ('openssl no-XXX' is not able to detect pseudo-commands such
6725 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
6728 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
6731 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
6732 is set; it will be thrown away anyway because each handshake creates
6734 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
6735 to parameters -- in previous versions (since OpenSSL 0.9.3) the
6736 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
6737 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
6740 *) New s_client option -ign_eof: EOF at stdin is ignored, and
6741 'Q' and 'R' lose their special meanings (quit/renegotiate).
6742 This is part of what -quiet does; unlike -quiet, -ign_eof
6743 does not suppress any output.
6746 *) Add compatibility options to the purpose and trust code. The
6747 purpose X509_PURPOSE_ANY is "any purpose" which automatically
6748 accepts a certificate or CA, this was the previous behaviour,
6749 with all the associated security issues.
6751 X509_TRUST_COMPAT is the old trust behaviour: only and
6752 automatically trust self signed roots in certificate store. A
6753 new trust setting X509_TRUST_DEFAULT is used to specify that
6754 a purpose has no associated trust setting and it should instead
6755 use the value in the default purpose.
6758 *) Fix the PKCS#8 DSA private key code so it decodes keys again
6759 and fix a memory leak.
6762 *) In util/mkerr.pl (which implements 'make errors'), preserve
6763 reason strings from the previous version of the .c file, as
6764 the default to have only downcase letters (and digits) in
6765 automatically generated reasons codes is not always appropriate.
6768 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
6769 using strerror. Previously, ERR_reason_error_string() returned
6770 library names as reason strings for SYSerr; but SYSerr is a special
6771 case where small numbers are errno values, not library numbers.
6774 *) Add '-dsaparam' option to 'openssl dhparam' application. This
6775 converts DSA parameters into DH parameters. (When creating parameters,
6776 DSA_generate_parameters is used.)
6779 *) Include 'length' (recommended exponent length) in C code generated
6780 by 'openssl dhparam -C'.
6783 *) The second argument to set_label in perlasm was already being used
6784 so couldn't be used as a "file scope" flag. Moved to third argument
6788 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
6789 instead of RAND_bytes for encryption IVs and salts.
6792 *) Include RAND_status() into RAND_METHOD instead of implementing
6793 it only for md_rand.c Otherwise replacing the PRNG by calling
6794 RAND_set_rand_method would be impossible.
6797 *) Don't let DSA_generate_key() enter an infinite loop if the random
6798 number generation fails.
6801 *) New 'rand' application for creating pseudo-random output.
6804 *) Added configuration support for Linux/IA64
6805 [Rolf Haberrecker <rolf@suse.de>]
6807 *) Assembler module support for Mingw32.
6810 *) Shared library support for HPUX (in shlib/).
6811 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
6813 *) Shared library support for Solaris gcc.
6814 [Lutz Behnke <behnke@trustcenter.de>]
6816 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
6818 *) PKCS7_encrypt() was adding text MIME headers twice because they
6819 were added manually and by SMIME_crlf_copy().
6822 *) In bntest.c don't call BN_rand with zero bits argument.
6823 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
6825 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
6826 case was implemented. This caused BN_div_recp() to fail occasionally.
6829 *) Add an optional second argument to the set_label() in the perl
6830 assembly language builder. If this argument exists and is set
6831 to 1 it signals that the assembler should use a symbol whose
6832 scope is the entire file, not just the current function. This
6833 is needed with MASM which uses the format label:: for this scope.
6834 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
6836 *) Change the ASN1 types so they are typedefs by default. Before
6837 almost all types were #define'd to ASN1_STRING which was causing
6838 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
6842 *) Change names of new functions to the new get1/get0 naming
6843 convention: After 'get1', the caller owns a reference count
6844 and has to call ..._free; 'get0' returns a pointer to some
6845 data structure without incrementing reference counters.
6846 (Some of the existing 'get' functions increment a reference
6847 counter, some don't.)
6848 Similarly, 'set1' and 'add1' functions increase reference
6849 counters or duplicate objects.
6852 *) Allow for the possibility of temp RSA key generation failure:
6853 the code used to assume it always worked and crashed on failure.
6856 *) Fix potential buffer overrun problem in BIO_printf().
6857 [Ulf Möller, using public domain code by Patrick Powell; problem
6858 pointed out by David Sacerdote <das33@cornell.edu>]
6860 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
6861 RAND_egd() and RAND_status(). In the command line application,
6862 the EGD socket can be specified like a seed file using RANDFILE
6866 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
6867 Some CAs (e.g. Verisign) distribute certificates in this form.
6870 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
6871 list to exclude them. This means that no special compilation option
6872 is needed to use anonymous DH: it just needs to be included in the
6876 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
6877 EVP_MD_type. The old functionality is available in a new macro called
6878 EVP_MD_md(). Change code that uses it and update docs.
6881 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
6882 where the 'void *' argument is replaced by a function pointer argument.
6883 Previously 'void *' was abused to point to functions, which works on
6884 many platforms, but is not correct. As these functions are usually
6885 called by macros defined in OpenSSL header files, most source code
6886 should work without changes.
6889 *) <openssl/opensslconf.h> (which is created by Configure) now contains
6890 sections with information on -D... compiler switches used for
6891 compiling the library so that applications can see them. To enable
6892 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
6893 must be defined. E.g.,
6894 #define OPENSSL_ALGORITHM_DEFINES
6895 #include <openssl/opensslconf.h>
6896 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
6897 [Richard Levitte, Ulf and Bodo Möller]
6899 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
6903 *) Change the 'other' type in certificate aux info to a STACK_OF
6904 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
6905 the required ASN1 format: arbitrary types determined by an OID.
6908 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
6909 argument to 'req'. This is not because the function is newer or
6910 better than others it just uses the work 'NEW' in the certificate
6911 request header lines. Some software needs this.
6914 *) Reorganise password command line arguments: now passwords can be
6915 obtained from various sources. Delete the PEM_cb function and make
6916 it the default behaviour: i.e. if the callback is NULL and the
6917 usrdata argument is not NULL interpret it as a null terminated pass
6918 phrase. If usrdata and the callback are NULL then the pass phrase
6919 is prompted for as usual.
6922 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
6923 the support is automatically enabled. The resulting binaries will
6924 autodetect the card and use it if present.
6925 [Ben Laurie and Compaq Inc.]
6927 *) Work around for Netscape hang bug. This sends certificate request
6928 and server done in one record. Since this is perfectly legal in the
6929 SSL/TLS protocol it isn't a "bug" option and is on by default. See
6930 the bugs/SSLv3 entry for more info.
6933 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
6936 *) Add -rand argument to smime and pkcs12 applications and read/write
6940 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
6943 *) Add command line password options to the remaining applications.
6946 *) Bug fix for BN_div_recp() for numerators with an even number of
6950 *) More tests in bntest.c, and changed test_bn output.
6953 *) ./config recognizes MacOS X now.
6956 *) Bug fix for BN_div() when the first words of num and divsor are
6957 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
6960 *) Add support for various broken PKCS#8 formats, and command line
6961 options to produce them.
6964 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
6965 get temporary BIGNUMs from a BN_CTX.
6968 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
6972 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
6973 include a #define from the old name to the new. The original intent
6974 was that statically linked binaries could for example just call
6975 SSLeay_add_all_ciphers() to just add ciphers to the table and not
6976 link with digests. This never worked becayse SSLeay_add_all_digests()
6977 and SSLeay_add_all_ciphers() were in the same source file so calling
6978 one would link with the other. They are now in separate source files.
6981 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
6984 *) Use a less unusual form of the Miller-Rabin primality test (it used
6985 a binary algorithm for exponentiation integrated into the Miller-Rabin
6986 loop, our standard modexp algorithms are faster).
6989 *) Support for the EBCDIC character set completed.
6990 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
6992 *) Source code cleanups: use const where appropriate, eliminate casts,
6993 use void * instead of char * in lhash.
6996 *) Bugfix: ssl3_send_server_key_exchange was not restartable
6997 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
6998 this the server could overwrite ephemeral keys that the client
7002 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7003 using 50 iterations of the Rabin-Miller test.
7005 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7006 iterations of the Rabin-Miller test as required by the appendix
7007 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7008 As BN_is_prime_fasttest includes trial division, DSA parameter
7009 generation becomes much faster.
7011 This implies a change for the callback functions in DSA_is_prime
7012 and DSA_generate_parameters: The callback function is called once
7013 for each positive witness in the Rabin-Miller test, not just
7014 occasionally in the inner loop; and the parameters to the
7015 callback function now provide an iteration count for the outer
7016 loop rather than for the current invocation of the inner loop.
7017 DSA_generate_parameters additionally can call the callback
7018 function with an 'iteration count' of -1, meaning that a
7019 candidate has passed the trial division test (when q is generated
7020 from an application-provided seed, trial division is skipped).
7023 *) New function BN_is_prime_fasttest that optionally does trial
7024 division before starting the Rabin-Miller test and has
7025 an additional BN_CTX * argument (whereas BN_is_prime always
7026 has to allocate at least one BN_CTX).
7027 'callback(1, -1, cb_arg)' is called when a number has passed the
7028 trial division stage.
7031 *) Fix for bug in CRL encoding. The validity dates weren't being handled
7035 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
7038 *) New function BN_pseudo_rand().
7041 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
7042 bignum version of BN_from_montgomery() with the working code from
7043 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
7047 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
7048 made it impossible to use the same SSL_SESSION data structure in
7049 SSL2 clients in multiple threads.
7052 *) The return value of RAND_load_file() no longer counts bytes obtained
7053 by stat(). RAND_load_file(..., -1) is new and uses the complete file
7054 to seed the PRNG (previously an explicit byte count was required).
7055 [Ulf Möller, Bodo Möller]
7057 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
7058 used (char *) instead of (void *) and had casts all over the place.
7061 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
7064 *) Retain source code compatibility for BN_prime_checks macro:
7065 BN_is_prime(..., BN_prime_checks, ...) now uses
7066 BN_prime_checks_for_size to determine the appropriate number of
7067 Rabin-Miller iterations.
7070 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
7071 DH_CHECK_P_NOT_SAFE_PRIME.
7072 (Check if this is true? OpenPGP calls them "strong".)
7075 *) Merge the functionality of "dh" and "gendh" programs into a new program
7076 "dhparam". The old programs are retained for now but will handle DH keys
7077 (instead of parameters) in future.
7080 *) Make the ciphers, s_server and s_client programs check the return values
7081 when a new cipher list is set.
7084 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
7085 ciphers. Before when the 56bit ciphers were enabled the sorting was
7088 The syntax for the cipher sorting has been extended to support sorting by
7089 cipher-strength (using the strength_bits hard coded in the tables).
7090 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
7092 Fix a bug in the cipher-command parser: when supplying a cipher command
7093 string with an "undefined" symbol (neither command nor alphanumeric
7094 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
7095 an error is flagged.
7097 Due to the strength-sorting extension, the code of the
7098 ssl_create_cipher_list() function was completely rearranged. I hope that
7099 the readability was also increased :-)
7100 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7102 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
7103 for the first serial number and places 2 in the serial number file. This
7104 avoids problems when the root CA is created with serial number zero and
7105 the first user certificate has the same issuer name and serial number
7109 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
7110 the new code. Add documentation for this stuff.
7113 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
7114 X509_*() to X509at_*() on the grounds that they don't handle X509
7115 structures and behave in an analagous way to the X509v3 functions:
7116 they shouldn't be called directly but wrapper functions should be used
7119 So we also now have some wrapper functions that call the X509at functions
7120 when passed certificate requests. (TO DO: similar things can be done with
7121 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
7122 things. Some of these need some d2i or i2d and print functionality
7123 because they handle more complex structures.)
7126 *) Add missing #ifndefs that caused missing symbols when building libssl
7127 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
7128 NO_RSA in ssl/s2*.c.
7129 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
7131 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
7132 has a return value which indicates the quality of the random data
7133 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
7134 error queue. New function RAND_pseudo_bytes() generates output that is
7135 guaranteed to be unique but not unpredictable. RAND_add is like
7136 RAND_seed, but takes an extra argument for an entropy estimate
7137 (RAND_seed always assumes full entropy).
7140 *) Do more iterations of Rabin-Miller probable prime test (specifically,
7141 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
7142 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
7143 in crypto/bn/bn_prime.c for the complete table). This guarantees a
7144 false-positive rate of at most 2^-80 for random input.
7147 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
7150 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
7151 in the 0.9.5 release), this returns the chain
7152 from an X509_CTX structure with a dup of the stack and all
7153 the X509 reference counts upped: so the stack will exist
7154 after X509_CTX_cleanup() has been called. Modify pkcs12.c
7157 Also make SSL_SESSION_print() print out the verify return
7161 *) Add manpage for the pkcs12 command. Also change the default
7162 behaviour so MAC iteration counts are used unless the new
7163 -nomaciter option is used. This improves file security and
7164 only older versions of MSIE (4.0 for example) need it.
7167 *) Honor the no-xxx Configure options when creating .DEF files.
7170 *) Add PKCS#10 attributes to field table: challengePassword,
7171 unstructuredName and unstructuredAddress. These are taken from
7172 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
7173 international characters are used.
7175 More changes to X509_ATTRIBUTE code: allow the setting of types
7176 based on strings. Remove the 'loc' parameter when adding
7177 attributes because these will be a SET OF encoding which is sorted
7181 *) Initial changes to the 'req' utility to allow request generation
7182 automation. This will allow an application to just generate a template
7183 file containing all the field values and have req construct the
7186 Initial support for X509_ATTRIBUTE handling. Stacks of these are
7187 used all over the place including certificate requests and PKCS#7
7188 structures. They are currently handled manually where necessary with
7189 some primitive wrappers for PKCS#7. The new functions behave in a
7190 manner analogous to the X509 extension functions: they allow
7191 attributes to be looked up by NID and added.
7193 Later something similar to the X509V3 code would be desirable to
7194 automatically handle the encoding, decoding and printing of the
7195 more complex types. The string types like challengePassword can
7196 be handled by the string table functions.
7198 Also modified the multi byte string table handling. Now there is
7199 a 'global mask' which masks out certain types. The table itself
7200 can use the flag STABLE_NO_MASK to ignore the mask setting: this
7201 is useful when for example there is only one permissible type
7202 (as in countryName) and using the mask might result in no valid
7206 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
7207 SSL_get_peer_finished to allow applications to obtain the latest
7208 Finished messages sent to the peer or expected from the peer,
7209 respectively. (SSL_get_peer_finished is usually the Finished message
7210 actually received from the peer, otherwise the protocol will be aborted.)
7212 As the Finished message are message digests of the complete handshake
7213 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
7214 be used for external authentication procedures when the authentication
7215 provided by SSL/TLS is not desired or is not enough.
7218 *) Enhanced support for Alpha Linux is added. Now ./config checks if
7219 the host supports BWX extension and if Compaq C is present on the
7220 $PATH. Just exploiting of the BWX extension results in 20-30%
7221 performance kick for some algorithms, e.g. DES and RC4 to mention
7222 a couple. Compaq C in turn generates ~20% faster code for MD5 and
7226 *) Add support for MS "fast SGC". This is arguably a violation of the
7227 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
7228 weak crypto and after checking the certificate is SGC a second one
7229 with strong crypto. MS SGC stops the first handshake after receiving
7230 the server certificate message and sends a second client hello. Since
7231 a server will typically do all the time consuming operations before
7232 expecting any further messages from the client (server key exchange
7233 is the most expensive) there is little difference between the two.
7235 To get OpenSSL to support MS SGC we have to permit a second client
7236 hello message after we have sent server done. In addition we have to
7237 reset the MAC if we do get this second client hello.
7240 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
7241 if a DER encoded private key is RSA or DSA traditional format. Changed
7242 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
7243 format DER encoded private key. Newer code should use PKCS#8 format which
7244 has the key type encoded in the ASN1 structure. Added DER private key
7245 support to pkcs8 application.
7248 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
7249 ciphersuites has been selected (as required by the SSL 3/TLS 1
7250 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
7251 is set, we interpret this as a request to violate the specification
7252 (the worst that can happen is a handshake failure, and 'correct'
7253 behaviour would result in a handshake failure anyway).
7256 *) In SSL_CTX_add_session, take into account that there might be multiple
7257 SSL_SESSION structures with the same session ID (e.g. when two threads
7258 concurrently obtain them from an external cache).
7259 The internal cache can handle only one SSL_SESSION with a given ID,
7260 so if there's a conflict, we now throw out the old one to achieve
7264 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
7265 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
7266 some routines that use cipher OIDs: some ciphers do not have OIDs
7267 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
7271 *) Simplify the trust setting structure and code. Now we just have
7272 two sequences of OIDs for trusted and rejected settings. These will
7273 typically have values the same as the extended key usage extension
7274 and any application specific purposes.
7276 The trust checking code now has a default behaviour: it will just
7277 check for an object with the same NID as the passed id. Functions can
7278 be provided to override either the default behaviour or the behaviour
7279 for a given id. SSL client, server and email already have functions
7280 in place for compatibility: they check the NID and also return "trusted"
7281 if the certificate is self signed.
7284 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
7285 traditional format into an EVP_PKEY structure.
7288 *) Add a password callback function PEM_cb() which either prompts for
7289 a password if usr_data is NULL or otherwise assumes it is a null
7290 terminated password. Allow passwords to be passed on command line
7291 environment or config files in a few more utilities.
7294 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
7295 keys. Add some short names for PKCS#8 PBE algorithms and allow them
7296 to be specified on the command line for the pkcs8 and pkcs12 utilities.
7297 Update documentation.
7300 *) Support for ASN1 "NULL" type. This could be handled before by using
7301 ASN1_TYPE but there wasn't any function that would try to read a NULL
7302 and produce an error if it couldn't. For compatibility we also have
7303 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
7304 don't allocate anything because they don't need to.
7307 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
7309 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
7311 *) Rebuild of the memory allocation routines used by OpenSSL code and
7312 possibly others as well. The purpose is to make an interface that
7313 provide hooks so anyone can build a separate set of allocation and
7314 deallocation routines to be used by OpenSSL, for example memory
7315 pool implementations, or something else, which was previously hard
7316 since Malloc(), Realloc() and Free() were defined as macros having
7317 the values malloc, realloc and free, respectively (except for Win32
7318 compilations). The same is provided for memory debugging code.
7319 OpenSSL already comes with functionality to find memory leaks, but
7320 this gives people a chance to debug other memory problems.
7322 With these changes, a new set of functions and macros have appeared:
7324 CRYPTO_set_mem_debug_functions() [F]
7325 CRYPTO_get_mem_debug_functions() [F]
7326 CRYPTO_dbg_set_options() [F]
7327 CRYPTO_dbg_get_options() [F]
7328 CRYPTO_malloc_debug_init() [M]
7330 The memory debug functions are NULL by default, unless the library
7331 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
7332 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
7333 gives the standard debugging functions that come with OpenSSL) or
7334 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
7335 provided by the library user) must be used. When the standard
7336 debugging functions are used, CRYPTO_dbg_set_options can be used to
7337 request additional information:
7338 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
7339 the CRYPTO_MDEBUG_xxx macro when compiling the library.
7341 Also, things like CRYPTO_set_mem_functions will always give the
7342 expected result (the new set of functions is used for allocation
7343 and deallocation) at all times, regardless of platform and compiler
7346 To finish it up, some functions that were never use in any other
7347 way than through macros have a new API and new semantic:
7350 CRYPTO_dbg_realloc()
7353 All macros of value have retained their old syntax.
7354 [Richard Levitte and Bodo Moeller]
7356 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
7357 ordering of SMIMECapabilities wasn't in "strength order" and there
7358 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
7362 *) Some ASN1 types with illegal zero length encoding (INTEGER,
7363 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
7364 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
7366 *) Merge in my S/MIME library for OpenSSL. This provides a simple
7367 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
7368 functionality to handle multipart/signed properly) and a utility
7369 called 'smime' to call all this stuff. This is based on code I
7370 originally wrote for Celo who have kindly allowed it to be
7371 included in OpenSSL.
7374 *) Add variants des_set_key_checked and des_set_key_unchecked of
7375 des_set_key (aka des_key_sched). Global variable des_check_key
7376 decides which of these is called by des_set_key; this way
7377 des_check_key behaves as it always did, but applications and
7378 the library itself, which was buggy for des_check_key == 1,
7379 have a cleaner way to pick the version they need.
7382 *) New function PKCS12_newpass() which changes the password of a
7386 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
7387 dynamic mix. In both cases the ids can be used as an index into the
7388 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
7389 functions so they accept a list of the field values and the
7390 application doesn't need to directly manipulate the X509_TRUST
7394 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
7398 *) Modify the way the V3 extension code looks up extensions. This now
7399 works in a similar way to the object code: we have some "standard"
7400 extensions in a static table which is searched with OBJ_bsearch()
7401 and the application can add dynamic ones if needed. The file
7402 crypto/x509v3/ext_dat.h now has the info: this file needs to be
7403 updated whenever a new extension is added to the core code and kept
7404 in ext_nid order. There is a simple program 'tabtest.c' which checks
7405 this. New extensions are not added too often so this file can readily
7406 be maintained manually.
7408 There are two big advantages in doing things this way. The extensions
7409 can be looked up immediately and no longer need to be "added" using
7410 X509V3_add_standard_extensions(): this function now does nothing.
7411 [Side note: I get *lots* of email saying the extension code doesn't
7412 work because people forget to call this function]
7413 Also no dynamic allocation is done unless new extensions are added:
7414 so if we don't add custom extensions there is no need to call
7415 X509V3_EXT_cleanup().
7418 *) Modify enc utility's salting as follows: make salting the default. Add a
7419 magic header, so unsalted files fail gracefully instead of just decrypting
7420 to garbage. This is because not salting is a big security hole, so people
7421 should be discouraged from doing it.
7424 *) Fixes and enhancements to the 'x509' utility. It allowed a message
7425 digest to be passed on the command line but it only used this
7426 parameter when signing a certificate. Modified so all relevant
7427 operations are affected by the digest parameter including the
7428 -fingerprint and -x509toreq options. Also -x509toreq choked if a
7429 DSA key was used because it didn't fix the digest.
7432 *) Initial certificate chain verify code. Currently tests the untrusted
7433 certificates for consistency with the verify purpose (which is set
7434 when the X509_STORE_CTX structure is set up) and checks the pathlength.
7436 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
7437 this is because it will reject chains with invalid extensions whereas
7438 every previous version of OpenSSL and SSLeay made no checks at all.
7440 Trust code: checks the root CA for the relevant trust settings. Trust
7441 settings have an initial value consistent with the verify purpose: e.g.
7442 if the verify purpose is for SSL client use it expects the CA to be
7443 trusted for SSL client use. However the default value can be changed to
7444 permit custom trust settings: one example of this would be to only trust
7445 certificates from a specific "secure" set of CAs.
7447 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
7448 which should be used for version portability: especially since the
7449 verify structure is likely to change more often now.
7451 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
7452 to set them. If not set then assume SSL clients will verify SSL servers
7455 Two new options to the verify program: -untrusted allows a set of
7456 untrusted certificates to be passed in and -purpose which sets the
7457 intended purpose of the certificate. If a purpose is set then the
7458 new chain verify code is used to check extension consistency.
7461 *) Support for the authority information access extension.
7464 *) Modify RSA and DSA PEM read routines to transparently handle
7465 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
7466 public keys in a format compatible with certificate
7467 SubjectPublicKeyInfo structures. Unfortunately there were already
7468 functions called *_PublicKey_* which used various odd formats so
7469 these are retained for compatibility: however the DSA variants were
7470 never in a public release so they have been deleted. Changed dsa/rsa
7471 utilities to handle the new format: note no releases ever handled public
7472 keys so we should be OK.
7474 The primary motivation for this change is to avoid the same fiasco
7475 that dogs private keys: there are several incompatible private key
7476 formats some of which are standard and some OpenSSL specific and
7477 require various evil hacks to allow partial transparent handling and
7478 even then it doesn't work with DER formats. Given the option anything
7479 other than PKCS#8 should be dumped: but the other formats have to
7480 stay in the name of compatibility.
7482 With public keys and the benefit of hindsight one standard format
7483 is used which works with EVP_PKEY, RSA or DSA structures: though
7484 it clearly returns an error if you try to read the wrong kind of key.
7486 Added a -pubkey option to the 'x509' utility to output the public key.
7487 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
7488 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
7489 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
7490 that do the same as the EVP_PKEY_assign_*() except they up the
7491 reference count of the added key (they don't "swallow" the
7495 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
7496 CRLs would fail if the file contained no certificates or no CRLs:
7497 added a new function to read in both types and return the number
7498 read: this means that if none are read it will be an error. The
7499 DER versions of the certificate and CRL reader would always fail
7500 because it isn't possible to mix certificates and CRLs in DER format
7501 without choking one or the other routine. Changed this to just read
7502 a certificate: this is the best we can do. Also modified the code
7503 in apps/verify.c to take notice of return codes: it was previously
7504 attempting to read in certificates from NULL pointers and ignoring
7505 any errors: this is one reason why the cert and CRL reader seemed
7506 to work. It doesn't check return codes from the default certificate
7507 routines: these may well fail if the certificates aren't installed.
7510 *) Code to support otherName option in GeneralName.
7513 *) First update to verify code. Change the verify utility
7514 so it warns if it is passed a self signed certificate:
7515 for consistency with the normal behaviour. X509_verify
7516 has been modified to it will now verify a self signed
7517 certificate if *exactly* the same certificate appears
7518 in the store: it was previously impossible to trust a
7519 single self signed certificate. This means that:
7520 openssl verify ss.pem
7521 now gives a warning about a self signed certificate but
7522 openssl verify -CAfile ss.pem ss.pem
7526 *) For servers, store verify_result in SSL_SESSION data structure
7527 (and add it to external session representation).
7528 This is needed when client certificate verifications fails,
7529 but an application-provided verification callback (set by
7530 SSL_CTX_set_cert_verify_callback) allows accepting the session
7531 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
7532 but returns 1): When the session is reused, we have to set
7533 ssl->verify_result to the appropriate error code to avoid
7535 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
7537 *) Fix a bug in the new PKCS#7 code: it didn't consider the
7538 case in PKCS7_dataInit() where the signed PKCS7 structure
7539 didn't contain any existing data because it was being created.
7540 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
7542 *) Add a salt to the key derivation routines in enc.c. This
7543 forms the first 8 bytes of the encrypted file. Also add a
7544 -S option to allow a salt to be input on the command line.
7547 *) New function X509_cmp(). Oddly enough there wasn't a function
7548 to compare two certificates. We do this by working out the SHA1
7549 hash and comparing that. X509_cmp() will be needed by the trust
7553 *) SSL_get1_session() is like SSL_get_session(), but increments
7554 the reference count in the SSL_SESSION returned.
7555 [Geoff Thorpe <geoff@eu.c2.net>]
7557 *) Fix for 'req': it was adding a null to request attributes.
7558 Also change the X509_LOOKUP and X509_INFO code to handle
7559 certificate auxiliary information.
7562 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
7566 *) Add the possibility to add extra information to the memory leak
7567 detecting output, to form tracebacks, showing from where each
7568 allocation was originated: CRYPTO_push_info("constant string") adds
7569 the string plus current file name and line number to a per-thread
7570 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
7571 is like calling CYRPTO_pop_info() until the stack is empty.
7572 Also updated memory leak detection code to be multi-thread-safe.
7575 *) Add options -text and -noout to pkcs7 utility and delete the
7576 encryption options which never did anything. Update docs.
7579 *) Add options to some of the utilities to allow the pass phrase
7580 to be included on either the command line (not recommended on
7581 OSes like Unix) or read from the environment. Update the
7582 manpages and fix a few bugs.
7585 *) Add a few manpages for some of the openssl commands.
7588 *) Fix the -revoke option in ca. It was freeing up memory twice,
7589 leaking and not finding already revoked certificates.
7592 *) Extensive changes to support certificate auxiliary information.
7593 This involves the use of X509_CERT_AUX structure and X509_AUX
7594 functions. An X509_AUX function such as PEM_read_X509_AUX()
7595 can still read in a certificate file in the usual way but it
7596 will also read in any additional "auxiliary information". By
7597 doing things this way a fair degree of compatibility can be
7598 retained: existing certificates can have this information added
7599 using the new 'x509' options.
7601 Current auxiliary information includes an "alias" and some trust
7602 settings. The trust settings will ultimately be used in enhanced
7603 certificate chain verification routines: currently a certificate
7604 can only be trusted if it is self signed and then it is trusted
7608 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
7609 The problem was that one of the replacement routines had not been working
7610 since SSLeay releases. For now the offending routine has been replaced
7611 with non-optimised assembler. Even so, this now gives around 95%
7612 performance improvement for 1024 bit RSA signs.
7615 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
7616 handling. Most clients have the effective key size in bits equal to
7617 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
7618 A few however don't do this and instead use the size of the decrypted key
7619 to determine the RC2 key length and the AlgorithmIdentifier to determine
7620 the effective key length. In this case the effective key length can still
7621 be 40 bits but the key length can be 168 bits for example. This is fixed
7622 by manually forcing an RC2 key into the EVP_PKEY structure because the
7623 EVP code can't currently handle unusual RC2 key sizes: it always assumes
7624 the key length and effective key length are equal.
7627 *) Add a bunch of functions that should simplify the creation of
7628 X509_NAME structures. Now you should be able to do:
7629 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
7630 and have it automatically work out the correct field type and fill in
7631 the structures. The more adventurous can try:
7632 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
7633 and it will (hopefully) work out the correct multibyte encoding.
7636 *) Change the 'req' utility to use the new field handling and multibyte
7637 copy routines. Before the DN field creation was handled in an ad hoc
7638 way in req, ca, and x509 which was rather broken and didn't support
7639 BMPStrings or UTF8Strings. Since some software doesn't implement
7640 BMPStrings or UTF8Strings yet, they can be enabled using the config file
7641 using the dirstring_type option. See the new comment in the default
7642 openssl.cnf for more info.
7645 *) Make crypto/rand/md_rand.c more robust:
7646 - Assure unique random numbers after fork().
7647 - Make sure that concurrent threads access the global counter and
7648 md serializably so that we never lose entropy in them
7649 or use exactly the same state in multiple threads.
7650 Access to the large state is not always serializable because
7651 the additional locking could be a performance killer, and
7652 md should be large enough anyway.
7655 *) New file apps/app_rand.c with commonly needed functionality
7656 for handling the random seed file.
7658 Use the random seed file in some applications that previously did not:
7660 dsaparam -genkey (which also ignored its '-rand' option),
7663 x509 (when signing).
7664 Except on systems with /dev/urandom, it is crucial to have a random
7665 seed file at least for key creation, DSA signing, and for DH exchanges;
7666 for RSA signatures we could do without one.
7668 gendh and gendsa (unlike genrsa) used to read only the first byte
7669 of each file listed in the '-rand' option. The function as previously
7670 found in genrsa is now in app_rand.c and is used by all programs
7671 that support '-rand'.
7674 *) In RAND_write_file, use mode 0600 for creating files;
7675 don't just chmod when it may be too late.
7678 *) Report an error from X509_STORE_load_locations
7679 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
7682 *) New function ASN1_mbstring_copy() this copies a string in either
7683 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
7684 into an ASN1_STRING type. A mask of permissible types is passed
7685 and it chooses the "minimal" type to use or an error if not type
7689 *) Add function equivalents to the various macros in asn1.h. The old
7690 macros are retained with an M_ prefix. Code inside the library can
7691 use the M_ macros. External code (including the openssl utility)
7692 should *NOT* in order to be "shared library friendly".
7695 *) Add various functions that can check a certificate's extensions
7696 to see if it usable for various purposes such as SSL client,
7697 server or S/MIME and CAs of these types. This is currently
7698 VERY EXPERIMENTAL but will ultimately be used for certificate chain
7699 verification. Also added a -purpose flag to x509 utility to
7700 print out all the purposes.
7703 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
7707 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
7708 for, obtain and decode and extension and obtain its critical flag.
7709 This allows all the necessary extension code to be handled in a
7710 single function call.
7713 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
7714 platforms. See crypto/rc4/rc4_enc.c for further details.
7717 *) New -noout option to asn1parse. This causes no output to be produced
7718 its main use is when combined with -strparse and -out to extract data
7719 from a file (which may not be in ASN.1 format).
7722 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
7723 when producing the local key id.
7724 [Richard Levitte <levitte@stacken.kth.se>]
7726 *) New option -dhparam in s_server. This allows a DH parameter file to be
7727 stated explicitly. If it is not stated then it tries the first server
7728 certificate file. The previous behaviour hard coded the filename
7732 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
7733 a public key to be input or output. For example:
7734 openssl rsa -in key.pem -pubout -out pubkey.pem
7735 Also added necessary DSA public key functions to handle this.
7738 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
7739 in the message. This was handled by allowing
7740 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
7741 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
7743 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
7744 to the end of the strings whereas this didn't. This would cause problems
7745 if strings read with d2i_ASN1_bytes() were later modified.
7746 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
7748 *) Fix for base64 decode bug. When a base64 bio reads only one line of
7749 data and it contains EOF it will end up returning an error. This is
7750 caused by input 46 bytes long. The cause is due to the way base64
7751 BIOs find the start of base64 encoded data. They do this by trying a
7752 trial decode on each line until they find one that works. When they
7753 do a flag is set and it starts again knowing it can pass all the
7754 data directly through the decoder. Unfortunately it doesn't reset
7755 the context it uses. This means that if EOF is reached an attempt
7756 is made to pass two EOFs through the context and this causes the
7757 resulting error. This can also cause other problems as well. As is
7758 usual with these problems it takes *ages* to find and the fix is
7759 trivial: move one line.
7760 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
7762 *) Ugly workaround to get s_client and s_server working under Windows. The
7763 old code wouldn't work because it needed to select() on sockets and the
7764 tty (for keypresses and to see if data could be written). Win32 only
7765 supports select() on sockets so we select() with a 1s timeout on the
7766 sockets and then see if any characters are waiting to be read, if none
7767 are present then we retry, we also assume we can always write data to
7768 the tty. This isn't nice because the code then blocks until we've
7769 received a complete line of data and it is effectively polling the
7770 keyboard at 1s intervals: however it's quite a bit better than not
7771 working at all :-) A dedicated Windows application might handle this
7772 with an event loop for example.
7775 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
7776 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
7777 will be called when RSA_sign() and RSA_verify() are used. This is useful
7778 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
7779 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
7780 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
7781 This necessitated the support of an extra signature type NID_md5_sha1
7782 for SSL signatures and modifications to the SSL library to use it instead
7783 of calling RSA_public_decrypt() and RSA_private_encrypt().
7786 *) Add new -verify -CAfile and -CApath options to the crl program, these
7787 will lookup a CRL issuers certificate and verify the signature in a
7788 similar way to the verify program. Tidy up the crl program so it
7789 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
7790 less strict. It will now permit CRL extensions even if it is not
7791 a V2 CRL: this will allow it to tolerate some broken CRLs.
7794 *) Initialize all non-automatic variables each time one of the openssl
7795 sub-programs is started (this is necessary as they may be started
7796 multiple times from the "OpenSSL>" prompt).
7797 [Lennart Bang, Bodo Moeller]
7799 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
7800 removing all other RSA functionality (this is what NO_RSA does). This
7801 is so (for example) those in the US can disable those operations covered
7802 by the RSA patent while allowing storage and parsing of RSA keys and RSA
7806 *) Non-copying interface to BIO pairs.
7807 (still largely untested)
7810 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
7811 ASCII string. This was handled independently in various places before.
7814 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
7815 UTF8 strings a character at a time.
7818 *) Use client_version from client hello to select the protocol
7819 (s23_srvr.c) and for RSA client key exchange verification
7820 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
7823 *) Add various utility functions to handle SPKACs, these were previously
7824 handled by poking round in the structure internals. Added new function
7825 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
7826 print, verify and generate SPKACs. Based on an original idea from
7827 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
7830 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
7833 *) Allow the config file extension section to be overwritten on the
7834 command line. Based on an original idea from Massimiliano Pala
7835 <madwolf@comune.modena.it>. The new option is called -extensions
7836 and can be applied to ca, req and x509. Also -reqexts to override
7837 the request extensions in req and -crlexts to override the crl extensions
7841 *) Add new feature to the SPKAC handling in ca. Now you can include
7842 the same field multiple times by preceding it by "XXXX." for example:
7845 this is the same syntax as used in the req config file.
7848 *) Allow certificate extensions to be added to certificate requests. These
7849 are specified in a 'req_extensions' option of the req section of the
7850 config file. They can be printed out with the -text option to req but
7851 are otherwise ignored at present.
7854 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
7855 data read consists of only the final block it would not decrypted because
7856 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
7857 A misplaced 'break' also meant the decrypted final block might not be
7858 copied until the next read.
7861 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
7862 a few extra parameters to the DH structure: these will be useful if
7863 for example we want the value of 'q' or implement X9.42 DH.
7866 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
7867 provides hooks that allow the default DSA functions or functions on a
7868 "per key" basis to be replaced. This allows hardware acceleration and
7869 hardware key storage to be handled without major modification to the
7870 library. Also added low level modexp hooks and CRYPTO_EX structure and
7871 associated functions.
7874 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
7875 as "read only": it can't be written to and the buffer it points to will
7876 not be freed. Reading from a read only BIO is much more efficient than
7877 a normal memory BIO. This was added because there are several times when
7878 an area of memory needs to be read from a BIO. The previous method was
7879 to create a memory BIO and write the data to it, this results in two
7880 copies of the data and an O(n^2) reading algorithm. There is a new
7881 function BIO_new_mem_buf() which creates a read only memory BIO from
7882 an area of memory. Also modified the PKCS#7 routines to use read only
7886 *) Bugfix: ssl23_get_client_hello did not work properly when called in
7887 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
7888 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
7889 but a retry condition occured while trying to read the rest.
7892 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
7893 NID_pkcs7_encrypted by default: this was wrong since this should almost
7894 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
7895 the encrypted data type: this is a more sensible place to put it and it
7896 allows the PKCS#12 code to be tidied up that duplicated this
7900 *) Changed obj_dat.pl script so it takes its input and output files on
7901 the command line. This should avoid shell escape redirection problems
7905 *) Initial support for certificate extension requests, these are included
7906 in things like Xenroll certificate requests. Included functions to allow
7907 extensions to be obtained and added.
7910 *) -crlf option to s_client and s_server for sending newlines as
7911 CRLF (as required by many protocols).
7914 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
7916 *) Install libRSAglue.a when OpenSSL is built with RSAref.
7917 [Ralf S. Engelschall]
7919 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
7920 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
7922 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
7926 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
7927 DH parameters/keys (q is lost during that conversion, but the resulting
7928 DH parameters contain its length).
7930 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
7931 much faster than DH_generate_parameters (which creates parameters
7932 where p = 2*q + 1), and also the smaller q makes DH computations
7933 much more efficient (160-bit exponentiation instead of 1024-bit
7934 exponentiation); so this provides a convenient way to support DHE
7935 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
7936 utter importance to use
7937 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
7939 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
7940 when such DH parameters are used, because otherwise small subgroup
7941 attacks may become possible!
7944 *) Avoid memory leak in i2d_DHparams.
7947 *) Allow the -k option to be used more than once in the enc program:
7948 this allows the same encrypted message to be read by multiple recipients.
7951 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
7952 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
7953 it will always use the numerical form of the OID, even if it has a short
7957 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
7958 method only got called if p,q,dmp1,dmq1,iqmp components were present,
7959 otherwise bn_mod_exp was called. In the case of hardware keys for example
7960 no private key components need be present and it might store extra data
7961 in the RSA structure, which cannot be accessed from bn_mod_exp.
7962 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
7963 private key operations.
7966 *) Added support for SPARC Linux.
7969 *) pem_password_cb function type incompatibly changed from
7970 typedef int pem_password_cb(char *buf, int size, int rwflag);
7972 ....(char *buf, int size, int rwflag, void *userdata);
7973 so that applications can pass data to their callbacks:
7974 The PEM[_ASN1]_{read,write}... functions and macros now take an
7975 additional void * argument, which is just handed through whenever
7976 the password callback is called.
7977 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
7979 New function SSL_CTX_set_default_passwd_cb_userdata.
7981 Compatibility note: As many C implementations push function arguments
7982 onto the stack in reverse order, the new library version is likely to
7983 interoperate with programs that have been compiled with the old
7984 pem_password_cb definition (PEM_whatever takes some data that
7985 happens to be on the stack as its last argument, and the callback
7986 just ignores this garbage); but there is no guarantee whatsoever that
7989 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
7990 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
7991 problems not only on Windows, but also on some Unix platforms.
7992 To avoid problematic command lines, these definitions are now in an
7993 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
7994 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
7997 *) MIPS III/IV assembler module is reimplemented.
8000 *) More DES library cleanups: remove references to srand/rand and
8001 delete an unused file.
8004 *) Add support for the the free Netwide assembler (NASM) under Win32,
8005 since not many people have MASM (ml) and it can be hard to obtain.
8006 This is currently experimental but it seems to work OK and pass all
8007 the tests. Check out INSTALL.W32 for info.
8010 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8011 without temporary keys kept an extra copy of the server key,
8012 and connections with temporary keys did not free everything in case
8016 *) New function RSA_check_key and new openssl rsa option -check
8017 for verifying the consistency of RSA keys.
8018 [Ulf Moeller, Bodo Moeller]
8020 *) Various changes to make Win32 compile work:
8021 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
8022 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
8023 comparison" warnings.
8024 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
8027 *) Add a debugging option to PKCS#5 v2 key generation function: when
8028 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
8029 derived keys are printed to stderr.
8032 *) Copy the flags in ASN1_STRING_dup().
8033 [Roman E. Pavlov <pre@mo.msk.ru>]
8035 *) The x509 application mishandled signing requests containing DSA
8036 keys when the signing key was also DSA and the parameters didn't match.
8038 It was supposed to omit the parameters when they matched the signing key:
8039 the verifying software was then supposed to automatically use the CA's
8040 parameters if they were absent from the end user certificate.
8042 Omitting parameters is no longer recommended. The test was also
8043 the wrong way round! This was probably due to unusual behaviour in
8044 EVP_cmp_parameters() which returns 1 if the parameters match.
8045 This meant that parameters were omitted when they *didn't* match and
8046 the certificate was useless. Certificates signed with 'ca' didn't have
8048 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
8050 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
8051 The interface is as follows:
8052 Applications can use
8053 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
8054 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
8055 "off" is now the default.
8056 The library internally uses
8057 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
8058 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
8059 to disable memory-checking temporarily.
8061 Some inconsistent states that previously were possible (and were
8062 even the default) are now avoided.
8064 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
8065 with each memory chunk allocated; this is occasionally more helpful
8066 than just having a counter.
8068 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
8070 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
8074 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
8075 which largely parallels "options", but is for changing API behaviour,
8076 whereas "options" are about protocol behaviour.
8077 Initial "mode" flags are:
8079 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
8080 a single record has been written.
8081 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
8082 retries use the same buffer location.
8083 (But all of the contents must be
8087 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
8090 *) Fix problems with no-hmac etc.
8091 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
8093 *) New functions RSA_get_default_method(), RSA_set_method() and
8094 RSA_get_method(). These allows replacement of RSA_METHODs without having
8095 to mess around with the internals of an RSA structure.
8098 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
8099 Also really enable memory leak checks in openssl.c and in some
8101 [Chad C. Mulligan, Bodo Moeller]
8103 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
8104 up the length of negative integers. This has now been simplified to just
8105 store the length when it is first determined and use it later, rather
8106 than trying to keep track of where data is copied and updating it to
8108 [Steve Henson, reported by Brien Wheeler
8109 <bwheeler@authentica-security.com>]
8111 *) Add a new function PKCS7_signatureVerify. This allows the verification
8112 of a PKCS#7 signature but with the signing certificate passed to the
8113 function itself. This contrasts with PKCS7_dataVerify which assumes the
8114 certificate is present in the PKCS#7 structure. This isn't always the
8115 case: certificates can be omitted from a PKCS#7 structure and be
8116 distributed by "out of band" means (such as a certificate database).
8119 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
8120 function prototypes in pem.h, also change util/mkdef.pl to add the
8121 necessary function names.
8124 *) mk1mf.pl (used by Windows builds) did not properly read the
8125 options set by Configure in the top level Makefile, and Configure
8126 was not even able to write more than one option correctly.
8127 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
8130 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
8131 file to be loaded from a BIO or FILE pointer. The BIO version will
8132 for example allow memory BIOs to contain config info.
8135 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
8136 Whoever hopes to achieve shared-library compatibility across versions
8137 must use this, not the compile-time macro.
8138 (Exercise 0.9.4: Which is the minimum library version required by
8140 Note: All this applies only to multi-threaded programs, others don't
8144 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
8145 through a BIO pair triggered the default case, i.e.
8146 SSLerr(...,SSL_R_UNKNOWN_STATE).
8149 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
8150 can use the SSL library even if none of the specific BIOs is
8154 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
8155 for the encoded length.
8156 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
8158 *) Add initial documentation of the X509V3 functions.
8161 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
8162 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
8163 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
8164 secure PKCS#8 private key format with a high iteration count.
8167 *) Fix determination of Perl interpreter: A perl or perl5
8168 _directory_ in $PATH was also accepted as the interpreter.
8169 [Ralf S. Engelschall]
8171 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
8172 wrong with it but it was very old and did things like calling
8173 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
8177 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
8178 to use the new extension code.
8181 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
8182 with macros. This should make it easier to change their form, add extra
8183 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
8187 *) Add to configuration table a new entry that can specify an alternative
8188 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
8189 according to Mark Crispin <MRC@Panda.COM>.
8193 *) DES CBC did not update the IV. Weird.
8196 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
8197 Changing the behaviour of the former might break existing programs --
8198 where IV updating is needed, des_ncbc_encrypt can be used.
8201 *) When bntest is run from "make test" it drives bc to check its
8202 calculations, as well as internally checking them. If an internal check
8203 fails, it needs to cause bc to give a non-zero result or make test carries
8204 on without noticing the failure. Fixed.
8207 *) DES library cleanups.
8210 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
8211 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
8212 ciphers. NOTE: although the key derivation function has been verified
8213 against some published test vectors it has not been extensively tested
8214 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
8218 *) Instead of "mkdir -p", which is not fully portable, use new
8219 Perl script "util/mkdir-p.pl".
8222 *) Rewrite the way password based encryption (PBE) is handled. It used to
8223 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
8224 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
8225 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
8226 the 'parameter' field of the AlgorithmIdentifier is passed to the
8227 underlying key generation function so it must do its own ASN1 parsing.
8228 This has also changed the EVP_PBE_CipherInit() function which now has a
8229 'parameter' argument instead of literal salt and iteration count values
8230 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
8233 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
8234 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
8235 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
8236 KEY" because this clashed with PKCS#8 unencrypted string. Since this
8237 value was just used as a "magic string" and not used directly its
8238 value doesn't matter.
8241 *) Introduce some semblance of const correctness to BN. Shame C doesn't
8245 *) "linux-sparc64" configuration (ultrapenguin).
8246 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
8247 "linux-sparc" configuration.
8248 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
8250 *) config now generates no-xxx options for missing ciphers.
8253 *) Support the EBCDIC character set (work in progress).
8254 File ebcdic.c not yet included because it has a different license.
8255 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8257 *) Support BS2000/OSD-POSIX.
8258 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8260 *) Make callbacks for key generation use void * instead of char *.
8263 *) Make S/MIME samples compile (not yet tested).
8266 *) Additional typesafe stacks.
8269 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
8273 Changes between 0.9.3 and 0.9.3a [29 May 1999]
8275 *) New configuration variant "sco5-gcc".
8277 *) Updated some demos.
8278 [Sean O Riordain, Wade Scholine]
8280 *) Add missing BIO_free at exit of pkcs12 application.
8283 *) Fix memory leak in conf.c.
8286 *) Updates for Win32 to assembler version of MD5.
8289 *) Set #! path to perl in apps/der_chop to where we found it
8290 instead of using a fixed path.
8293 *) SHA library changes for irix64-mips4-cc.
8296 *) Improvements for VMS support.
8300 Changes between 0.9.2b and 0.9.3 [24 May 1999]
8302 *) Bignum library bug fix. IRIX 6 passes "make test" now!
8303 This also avoids the problems with SC4.2 and unpatched SC5.
8304 [Andy Polyakov <appro@fy.chalmers.se>]
8306 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
8307 These are required because of the typesafe stack would otherwise break
8308 existing code. If old code used a structure member which used to be STACK
8309 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
8310 sk_num or sk_value it would produce an error because the num, data members
8311 are not present in STACK_OF. Now it just produces a warning. sk_set
8312 replaces the old method of assigning a value to sk_value
8313 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
8314 that does this will no longer work (and should use sk_set instead) but
8315 this could be regarded as a "questionable" behaviour anyway.
8318 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
8319 correctly handle encrypted S/MIME data.
8322 *) Change type of various DES function arguments from des_cblock
8323 (which means, in function argument declarations, pointer to char)
8324 to des_cblock * (meaning pointer to array with 8 char elements),
8325 which allows the compiler to do more typechecking; it was like
8326 that back in SSLeay, but with lots of ugly casts.
8328 Introduce new type const_des_cblock.
8331 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
8332 problems: find RecipientInfo structure that matches recipient certificate
8333 and initialise the ASN1 structures properly based on passed cipher.
8336 *) Belatedly make the BN tests actually check the results.
8339 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
8340 to and from BNs: it was completely broken. New compilation option
8341 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
8342 key elements as negative integers.
8345 *) Reorganize and speed up MD5.
8346 [Andy Polyakov <appro@fy.chalmers.se>]
8349 [Richard Levitte <richard@levitte.org>]
8351 *) New option -out to asn1parse to allow the parsed structure to be
8352 output to a file. This is most useful when combined with the -strparse
8353 option to examine the output of things like OCTET STRINGS.
8356 *) Make SSL library a little more fool-proof by not requiring any longer
8357 that SSL_set_{accept,connect}_state be called before
8358 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
8359 in many applications because usually everything *appeared* to work as
8360 intended anyway -- now it really works as intended).
8363 *) Move openssl.cnf out of lib/.
8366 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
8367 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
8368 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
8369 [Ralf S. Engelschall]
8371 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
8372 handle PKCS#7 enveloped data properly.
8373 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
8375 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
8376 copying pointers. The cert_st handling is changed by this in
8377 various ways (and thus what used to be known as ctx->default_cert
8378 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
8379 any longer when s->cert does not give us what we need).
8380 ssl_cert_instantiate becomes obsolete by this change.
8381 As soon as we've got the new code right (possibly it already is?),
8382 we have solved a couple of bugs of the earlier code where s->cert
8383 was used as if it could not have been shared with other SSL structures.
8385 Note that using the SSL API in certain dirty ways now will result
8386 in different behaviour than observed with earlier library versions:
8387 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
8388 does not influence s as it used to.
8390 In order to clean up things more thoroughly, inside SSL_SESSION
8391 we don't use CERT any longer, but a new structure SESS_CERT
8392 that holds per-session data (if available); currently, this is
8393 the peer's certificate chain and, for clients, the server's certificate
8394 and temporary key. CERT holds only those values that can have
8395 meaningful defaults in an SSL_CTX.
8398 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
8399 from the internal representation. Various PKCS#7 fixes: remove some
8400 evil casts and set the enc_dig_alg field properly based on the signing
8404 *) Allow PKCS#12 password to be set from the command line or the
8405 environment. Let 'ca' get its config file name from the environment
8406 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
8410 *) Allow certificate policies extension to use an IA5STRING for the
8411 organization field. This is contrary to the PKIX definition but
8412 VeriSign uses it and IE5 only recognises this form. Document 'x509'
8416 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
8417 without disallowing inline assembler and the like for non-pedantic builds.
8420 *) Support Borland C++ builder.
8421 [Janez Jere <jj@void.si>, modified by Ulf Möller]
8426 *) SHA-1 cleanups and performance enhancements.
8427 [Andy Polyakov <appro@fy.chalmers.se>]
8429 *) Sparc v8plus assembler for the bignum library.
8430 [Andy Polyakov <appro@fy.chalmers.se>]
8432 *) Accept any -xxx and +xxx compiler options in Configure.
8435 *) Update HPUX configuration.
8438 *) Add missing sk_<type>_unshift() function to safestack.h
8439 [Ralf S. Engelschall]
8441 *) New function SSL_CTX_use_certificate_chain_file that sets the
8442 "extra_cert"s in addition to the certificate. (This makes sense
8443 only for "PEM" format files, as chains as a whole are not
8447 *) Support verify_depth from the SSL API.
8448 x509_vfy.c had what can be considered an off-by-one-error:
8449 Its depth (which was not part of the external interface)
8450 was actually counting the number of certificates in a chain;
8451 now it really counts the depth.
8454 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
8455 instead of X509err, which often resulted in confusing error
8456 messages since the error codes are not globally unique
8457 (e.g. an alleged error in ssl3_accept when a certificate
8458 didn't match the private key).
8460 *) New function SSL_CTX_set_session_id_context that allows to set a default
8461 value (so that you don't need SSL_set_session_id_context for each
8462 connection using the SSL_CTX).
8465 *) OAEP decoding bug fix.
8468 *) Support INSTALL_PREFIX for package builders, as proposed by
8472 *) New Configure options "threads" and "no-threads". For systems
8473 where the proper compiler options are known (currently Solaris
8474 and Linux), "threads" is the default.
8477 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
8480 *) Install various scripts to $(OPENSSLDIR)/misc, not to
8481 $(INSTALLTOP)/bin -- they shouldn't clutter directories
8482 such as /usr/local/bin.
8485 *) "make linux-shared" to build shared libraries.
8486 [Niels Poppe <niels@netbox.org>]
8488 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
8491 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
8492 extension adding in x509 utility.
8495 *) Remove NOPROTO sections and error code comments.
8498 *) Partial rewrite of the DEF file generator to now parse the ANSI
8502 *) New Configure options --prefix=DIR and --openssldir=DIR.
8505 *) Complete rewrite of the error code script(s). It is all now handled
8506 by one script at the top level which handles error code gathering,
8507 header rewriting and C source file generation. It should be much better
8508 than the old method: it now uses a modified version of Ulf's parser to
8509 read the ANSI prototypes in all header files (thus the old K&R definitions
8510 aren't needed for error creation any more) and do a better job of
8511 translating function codes into names. The old 'ASN1 error code imbedded
8512 in a comment' is no longer necessary and it doesn't use .err files which
8513 have now been deleted. Also the error code call doesn't have to appear all
8514 on one line (which resulted in some large lines...).
8517 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
8520 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
8521 0 (which usually indicates a closed connection), but continue reading.
8524 *) Fix some race conditions.
8527 *) Add support for CRL distribution points extension. Add Certificate
8528 Policies and CRL distribution points documentation.
8531 *) Move the autogenerated header file parts to crypto/opensslconf.h.
8534 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
8535 8 of keying material. Merlin has also confirmed interop with this fix
8536 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
8537 [Merlin Hughes <merlin@baltimore.ie>]
8539 *) Fix lots of warnings.
8540 [Richard Levitte <levitte@stacken.kth.se>]
8542 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
8543 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
8544 [Richard Levitte <levitte@stacken.kth.se>]
8546 *) Fix problems with sizeof(long) == 8.
8547 [Andy Polyakov <appro@fy.chalmers.se>]
8549 *) Change functions to ANSI C.
8552 *) Fix typos in error codes.
8553 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
8555 *) Remove defunct assembler files from Configure.
8558 *) SPARC v8 assembler BIGNUM implementation.
8559 [Andy Polyakov <appro@fy.chalmers.se>]
8561 *) Support for Certificate Policies extension: both print and set.
8562 Various additions to support the r2i method this uses.
8565 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
8566 return a const string when you are expecting an allocated buffer.
8569 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
8570 types DirectoryString and DisplayText.
8573 *) Add code to allow r2i extensions to access the configuration database,
8574 add an LHASH database driver and add several ctx helper functions.
8577 *) Fix an evil bug in bn_expand2() which caused various BN functions to
8578 fail when they extended the size of a BIGNUM.
8581 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
8582 support typesafe stack.
8585 *) Fix typo in SSL_[gs]et_options().
8586 [Nils Frostberg <nils@medcom.se>]
8588 *) Delete various functions and files that belonged to the (now obsolete)
8589 old X509V3 handling code.
8592 *) New Configure option "rsaref".
8595 *) Don't auto-generate pem.h.
8598 *) Introduce type-safe ASN.1 SETs.
8601 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
8602 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
8604 *) Introduce type-safe STACKs. This will almost certainly break lots of code
8605 that links with OpenSSL (well at least cause lots of warnings), but fear
8606 not: the conversion is trivial, and it eliminates loads of evil casts. A
8607 few STACKed things have been converted already. Feel free to convert more.
8608 In the fullness of time, I'll do away with the STACK type altogether.
8611 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
8612 specified in <certfile> by updating the entry in the index.txt file.
8613 This way one no longer has to edit the index.txt file manually for
8614 revoking a certificate. The -revoke option does the gory details now.
8615 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
8617 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
8618 `-text' option at all and this way the `-noout -text' combination was
8619 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
8620 [Ralf S. Engelschall]
8622 *) Make sure a corresponding plain text error message exists for the
8623 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
8624 verify callback function determined that a certificate was revoked.
8625 [Ralf S. Engelschall]
8627 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
8628 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
8629 all available cipers including rc5, which was forgotten until now.
8630 In order to let the testing shell script know which algorithms
8631 are available, a new (up to now undocumented) command
8632 "openssl list-cipher-commands" is used.
8635 *) Bugfix: s_client occasionally would sleep in select() when
8636 it should have checked SSL_pending() first.
8639 *) New functions DSA_do_sign and DSA_do_verify to provide access to
8640 the raw DSA values prior to ASN.1 encoding.
8643 *) Tweaks to Configure
8644 [Niels Poppe <niels@netbox.org>]
8646 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
8650 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
8653 *) New config option to avoid instructions that are illegal on the 80386.
8654 The default code is faster, but requires at least a 486.
8657 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
8658 SSL2_SERVER_VERSION (not used at all) macros, which are now the
8659 same as SSL2_VERSION anyway.
8662 *) New "-showcerts" option for s_client.
8665 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
8666 application. Various cleanups and fixes.
8669 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
8670 modify error routines to work internally. Add error codes and PBE init
8671 to library startup routines.
8674 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
8675 packing functions to asn1 and evp. Changed function names and error
8676 codes along the way.
8679 *) PKCS12 integration: and so it begins... First of several patches to
8680 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
8681 objects to objects.h
8684 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
8685 and display support for Thawte strong extranet extension.
8688 *) Add LinuxPPC support.
8689 [Jeff Dubrule <igor@pobox.org>]
8691 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
8692 bn_div_words in alpha.s.
8693 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
8695 *) Make sure the RSA OAEP test is skipped under -DRSAref because
8696 OAEP isn't supported when OpenSSL is built with RSAref.
8697 [Ulf Moeller <ulf@fitug.de>]
8699 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
8700 so they no longer are missing under -DNOPROTO.
8701 [Soren S. Jorvang <soren@t.dk>]
8704 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
8706 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
8707 doesn't work when the session is reused. Coming soon!
8710 *) Fix a security hole, that allows sessions to be reused in the wrong
8711 context thus bypassing client cert protection! All software that uses
8712 client certs and session caches in multiple contexts NEEDS PATCHING to
8713 allow session reuse! A fuller solution is in the works.
8714 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
8716 *) Some more source tree cleanups (removed obsolete files
8717 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
8718 permission on "config" script to be executable) and a fix for the INSTALL
8720 [Ulf Moeller <ulf@fitug.de>]
8722 *) Remove some legacy and erroneous uses of malloc, free instead of
8724 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
8726 *) Make rsa_oaep_test return non-zero on error.
8727 [Ulf Moeller <ulf@fitug.de>]
8729 *) Add support for native Solaris shared libraries. Configure
8730 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
8731 if someone would make that last step automatic.
8732 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
8734 *) ctx_size was not built with the right compiler during "make links". Fixed.
8737 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
8738 except NULL ciphers". This means the default cipher list will no longer
8739 enable NULL ciphers. They need to be specifically enabled e.g. with
8740 the string "DEFAULT:eNULL".
8743 *) Fix to RSA private encryption routines: if p < q then it would
8744 occasionally produce an invalid result. This will only happen with
8745 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
8748 *) Be less restrictive and allow also `perl util/perlpath.pl
8749 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
8750 because this way one can also use an interpreter named `perl5' (which is
8751 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
8752 installed as `perl').
8753 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8755 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
8756 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8758 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
8759 advapi32.lib to Win32 build and change the pem test comparision
8760 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
8761 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
8762 and crypto/des/ede_cbcm_enc.c.
8765 *) DES quad checksum was broken on big-endian architectures. Fixed.
8768 *) Comment out two functions in bio.h that aren't implemented. Fix up the
8769 Win32 test batch file so it (might) work again. The Win32 test batch file
8770 is horrible: I feel ill....
8773 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
8774 in e_os.h. Audit of header files to check ANSI and non ANSI
8775 sections: 10 functions were absent from non ANSI section and not exported
8776 from Windows DLLs. Fixed up libeay.num for new functions.
8779 *) Make `openssl version' output lines consistent.
8780 [Ralf S. Engelschall]
8782 *) Fix Win32 symbol export lists for BIO functions: Added
8783 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
8784 to ms/libeay{16,32}.def.
8785 [Ralf S. Engelschall]
8787 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
8788 fine under Unix and passes some trivial tests I've now added. But the
8789 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
8790 added to make sure no one expects that this stuff really works in the
8791 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
8792 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
8794 [Ralf S. Engelschall]
8796 *) Fix the generation of two part addresses in perl.
8797 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
8799 *) Add config entry for Linux on MIPS.
8800 [John Tobey <jtobey@channel1.com>]
8802 *) Make links whenever Configure is run, unless we are on Windoze.
8805 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
8806 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
8810 *) Add a useful kludge to allow package maintainers to specify compiler and
8811 other platforms details on the command line without having to patch the
8812 Configure script everytime: One now can use ``perl Configure
8813 <id>:<details>'', i.e. platform ids are allowed to have details appended
8814 to them (seperated by colons). This is treated as there would be a static
8815 pre-configured entry in Configure's %table under key <id> with value
8816 <details> and ``perl Configure <id>'' is called. So, when you want to
8817 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
8818 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
8819 now, which overrides the FreeBSD-elf entry on-the-fly.
8820 [Ralf S. Engelschall]
8822 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
8825 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
8826 on the `perl Configure ...' command line. This way one can compile
8827 OpenSSL libraries with Position Independent Code (PIC) which is needed
8828 for linking it into DSOs.
8829 [Ralf S. Engelschall]
8831 *) Remarkably, export ciphers were totally broken and no-one had noticed!
8835 *) Cleaned up the LICENSE document: The official contact for any license
8836 questions now is the OpenSSL core team under openssl-core@openssl.org.
8837 And add a paragraph about the dual-license situation to make sure people
8838 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
8839 to the OpenSSL toolkit.
8840 [Ralf S. Engelschall]
8842 *) General source tree makefile cleanups: Made `making xxx in yyy...'
8843 display consistent in the source tree and replaced `/bin/rm' by `rm'.
8844 Additonally cleaned up the `make links' target: Remove unnecessary
8845 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
8846 to speed processing and no longer clutter the display with confusing
8847 stuff. Instead only the actually done links are displayed.
8848 [Ralf S. Engelschall]
8850 *) Permit null encryption ciphersuites, used for authentication only. It used
8851 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
8852 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
8856 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
8857 signed attributes when verifying signatures (this would break them),
8858 the detached data encoding was wrong and public keys obtained using
8859 X509_get_pubkey() weren't freed.
8862 *) Add text documentation for the BUFFER functions. Also added a work around
8863 to a Win95 console bug. This was triggered by the password read stuff: the
8864 last character typed gets carried over to the next fread(). If you were
8865 generating a new cert request using 'req' for example then the last
8866 character of the passphrase would be CR which would then enter the first
8870 *) Added the new `Includes OpenSSL Cryptography Software' button as
8871 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
8872 button and can be used by applications based on OpenSSL to show the
8873 relationship to the OpenSSL project.
8874 [Ralf S. Engelschall]
8876 *) Remove confusing variables in function signatures in files
8877 ssl/ssl_lib.c and ssl/ssl.h.
8878 [Lennart Bong <lob@kulthea.stacken.kth.se>]
8880 *) Don't install bss_file.c under PREFIX/include/
8881 [Lennart Bong <lob@kulthea.stacken.kth.se>]
8883 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
8884 functions that return function pointers and has support for NT specific
8885 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
8886 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
8887 unsigned to signed types: this was killing the Win32 compile.
8890 *) Add new certificate file to stack functions,
8891 SSL_add_dir_cert_subjects_to_stack() and
8892 SSL_add_file_cert_subjects_to_stack(). These largely supplant
8893 SSL_load_client_CA_file(), and can be used to add multiple certs easily
8894 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
8895 This means that Apache-SSL and similar packages don't have to mess around
8896 to add as many CAs as they want to the preferred list.
8899 *) Experiment with doxygen documentation. Currently only partially applied to
8901 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
8902 openssl.doxy as the configuration file.
8905 *) Get rid of remaining C++-style comments which strict C compilers hate.
8906 [Ralf S. Engelschall, pointed out by Carlos Amengual]
8908 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
8909 compiled in by default: it has problems with large keys.
8912 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
8913 DH private keys and/or callback functions which directly correspond to
8914 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
8915 is needed for applications which have to configure certificates on a
8916 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
8918 For the RSA certificate situation is makes no difference, but
8919 for the DSA certificate situation this fixes the "no shared cipher"
8920 problem where the OpenSSL cipher selection procedure failed because the
8921 temporary keys were not overtaken from the context and the API provided
8922 no way to reconfigure them.
8923 The new functions now let applications reconfigure the stuff and they
8924 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
8925 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
8926 non-public-API function ssl_cert_instantiate() is used as a helper
8927 function and also to reduce code redundancy inside ssl_rsa.c.
8928 [Ralf S. Engelschall]
8930 *) Move s_server -dcert and -dkey options out of the undocumented feature
8931 area because they are useful for the DSA situation and should be
8932 recognized by the users.
8933 [Ralf S. Engelschall]
8935 *) Fix the cipher decision scheme for export ciphers: the export bits are
8936 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
8937 SSL_EXP_MASK. So, the original variable has to be used instead of the
8938 already masked variable.
8939 [Richard Levitte <levitte@stacken.kth.se>]
8941 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
8942 [Richard Levitte <levitte@stacken.kth.se>]
8944 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
8945 from `int' to `unsigned int' because it's a length and initialized by
8946 EVP_DigestFinal() which expects an `unsigned int *'.
8947 [Richard Levitte <levitte@stacken.kth.se>]
8949 *) Don't hard-code path to Perl interpreter on shebang line of Configure
8950 script. Instead use the usual Shell->Perl transition trick.
8951 [Ralf S. Engelschall]
8953 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
8954 (in addition to RSA certificates) to match the behaviour of `openssl dsa
8955 -noout -modulus' as it's already the case for `openssl rsa -noout
8956 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
8957 currently the public key is printed (a decision which was already done by
8958 `openssl dsa -modulus' in the past) which serves a similar purpose.
8959 Additionally the NO_RSA no longer completely removes the whole -modulus
8960 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
8962 [Ralf S. Engelschall]
8964 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
8965 BIO. See the source (crypto/evp/bio_ok.c) for more info.
8966 [Arne Ansper <arne@ats.cyber.ee>]
8968 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
8969 to be added. Now both 'req' and 'ca' can use new objects defined in the
8973 *) Add cool BIO that does syslog (or event log on NT).
8974 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
8976 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
8977 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
8978 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
8979 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
8982 *) Add preliminary config info for new extension code.
8985 *) Make RSA_NO_PADDING really use no padding.
8986 [Ulf Moeller <ulf@fitug.de>]
8988 *) Generate errors when private/public key check is done.
8991 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
8992 for some CRL extensions and new objects added.
8995 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
8996 key usage extension and fuller support for authority key id.
8999 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9000 padding method for RSA, which is recommended for new applications in PKCS
9001 #1 v2.0 (RFC 2437, October 1998).
9002 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9003 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9004 against Bleichbacher's attack on RSA.
9005 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9008 *) Updates to the new SSL compression code
9009 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9011 *) Fix so that the version number in the master secret, when passed
9012 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9013 (because the server will not accept higher), that the version number
9014 is 0x03,0x01, not 0x03,0x00
9015 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9017 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9018 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9019 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
9022 *) Support for RAW extensions where an arbitrary extension can be
9023 created by including its DER encoding. See apps/openssl.cnf for
9027 *) Make sure latest Perl versions don't interpret some generated C array
9028 code as Perl array code in the crypto/err/err_genc.pl script.
9029 [Lars Weber <3weber@informatik.uni-hamburg.de>]
9031 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
9032 not many people have the assembler. Various Win32 compilation fixes and
9033 update to the INSTALL.W32 file with (hopefully) more accurate Win32
9037 *) Modify configure script 'Configure' to automatically create crypto/date.h
9038 file under Win32 and also build pem.h from pem.org. New script
9039 util/mkfiles.pl to create the MINFO file on environments that can't do a
9040 'make files': perl util/mkfiles.pl >MINFO should work.
9043 *) Major rework of DES function declarations, in the pursuit of correctness
9044 and purity. As a result, many evil casts evaporated, and some weirdness,
9045 too. You may find this causes warnings in your code. Zapping your evil
9046 casts will probably fix them. Mostly.
9049 *) Fix for a typo in asn1.h. Bug fix to object creation script
9050 obj_dat.pl. It considered a zero in an object definition to mean
9051 "end of object": none of the objects in objects.h have any zeros
9052 so it wasn't spotted.
9053 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
9055 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
9056 Masking (CBCM). In the absence of test vectors, the best I have been able
9057 to do is check that the decrypt undoes the encrypt, so far. Send me test
9058 vectors if you have them.
9061 *) Correct calculation of key length for export ciphers (too much space was
9062 allocated for null ciphers). This has not been tested!
9065 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
9066 message is now correct (it understands "crypto" and "ssl" on its
9067 command line). There is also now an "update" option. This will update
9068 the util/ssleay.num and util/libeay.num files with any new functions.
9070 perl util/mkdef.pl crypto ssl update
9071 it will update them.
9074 *) Overhauled the Perl interface (perl/*):
9075 - ported BN stuff to OpenSSL's different BN library
9076 - made the perl/ source tree CVS-aware
9077 - renamed the package from SSLeay to OpenSSL (the files still contain
9078 their history because I've copied them in the repository)
9079 - removed obsolete files (the test scripts will be replaced
9080 by better Test::Harness variants in the future)
9081 [Ralf S. Engelschall]
9083 *) First cut for a very conservative source tree cleanup:
9084 1. merge various obsolete readme texts into doc/ssleay.txt
9085 where we collect the old documents and readme texts.
9086 2. remove the first part of files where I'm already sure that we no
9087 longer need them because of three reasons: either they are just temporary
9088 files which were left by Eric or they are preserved original files where
9089 I've verified that the diff is also available in the CVS via "cvs diff
9090 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
9091 the crypto/md/ stuff).
9092 [Ralf S. Engelschall]
9094 *) More extension code. Incomplete support for subject and issuer alt
9095 name, issuer and authority key id. Change the i2v function parameters
9096 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
9097 what that's for :-) Fix to ASN1 macro which messed up
9098 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
9101 *) Preliminary support for ENUMERATED type. This is largely copied from the
9105 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
9106 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9108 *) Make sure `make rehash' target really finds the `openssl' program.
9109 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9111 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
9112 like to hear about it if this slows down other processors.
9115 *) Add CygWin32 platform information to Configure script.
9116 [Alan Batie <batie@aahz.jf.intel.com>]
9118 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
9119 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
9121 *) New program nseq to manipulate netscape certificate sequences
9124 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
9128 *) Fixes to BN code. Previously the default was to define BN_RECURSION
9129 but the BN code had some problems that would cause failures when
9130 doing certificate verification and some other functions.
9131 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9133 *) Add ASN1 and PEM code to support netscape certificate sequences.
9136 *) Add ASN1 and PEM code to support netscape certificate sequences.
9139 *) Add several PKIX and private extended key usage OIDs.
9142 *) Modify the 'ca' program to handle the new extension code. Modify
9143 openssl.cnf for new extension format, add comments.
9146 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
9147 and add a sample to openssl.cnf so req -x509 now adds appropriate
9151 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
9152 error code, add initial support to X509_print() and x509 application.
9155 *) Takes a deep breath and start addding X509 V3 extension support code. Add
9156 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
9157 stuff is currently isolated and isn't even compiled yet.
9160 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
9161 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
9162 Removed the versions check from X509 routines when loading extensions:
9163 this allows certain broken certificates that don't set the version
9164 properly to be processed.
9167 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
9168 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
9169 can still be regenerated with "make depend".
9172 *) Spelling mistake in C version of CAST-128.
9173 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
9175 *) Changes to the error generation code. The perl script err-code.pl
9176 now reads in the old error codes and retains the old numbers, only
9177 adding new ones if necessary. It also only changes the .err files if new
9178 codes are added. The makefiles have been modified to only insert errors
9179 when needed (to avoid needlessly modifying header files). This is done
9180 by only inserting errors if the .err file is newer than the auto generated
9181 C file. To rebuild all the error codes from scratch (the old behaviour)
9182 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
9183 or delete all the .err files.
9186 *) CAST-128 was incorrectly implemented for short keys. The C version has
9187 been fixed, but is untested. The assembler versions are also fixed, but
9188 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
9189 to regenerate it if needed.
9190 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
9191 Hagino <itojun@kame.net>]
9193 *) File was opened incorrectly in randfile.c.
9194 [Ulf Möller <ulf@fitug.de>]
9196 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
9197 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
9198 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
9199 al: it's just almost always a UTCTime. Note this patch adds new error
9200 codes so do a "make errors" if there are problems.
9203 *) Correct Linux 1 recognition in config.
9204 [Ulf Möller <ulf@fitug.de>]
9206 *) Remove pointless MD5 hash when using DSA keys in ca.
9207 [Anonymous <nobody@replay.com>]
9209 *) Generate an error if given an empty string as a cert directory. Also
9210 generate an error if handed NULL (previously returned 0 to indicate an
9211 error, but didn't set one).
9212 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
9214 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
9217 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
9218 parameters. This was causing a warning which killed off the Win32 compile.
9221 *) Remove C++ style comments from crypto/bn/bn_local.h.
9222 [Neil Costigan <neil.costigan@celocom.com>]
9224 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
9225 based on a text string, looking up short and long names and finally
9226 "dot" format. The "dot" format stuff didn't work. Added new function
9227 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
9228 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
9229 OID is not part of the table.
9232 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
9233 X509_LOOKUP_by_alias().
9236 *) Sort openssl functions by name.
9239 *) Get the gendsa program working (hopefully) and add it to app list. Remove
9240 encryption from sample DSA keys (in case anyone is interested the password
9244 *) Make _all_ *_free functions accept a NULL pointer.
9245 [Frans Heymans <fheymans@isaserver.be>]
9247 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
9249 [Anonymous <nobody@replay.com>]
9251 *) s_server should send the CAfile as acceptable CAs, not its own cert.
9252 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9254 *) Don't blow it for numeric -newkey arguments to apps/req.
9255 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9257 *) Temp key "for export" tests were wrong in s3_srvr.c.
9258 [Anonymous <nobody@replay.com>]
9260 *) Add prototype for temp key callback functions
9261 SSL_CTX_set_tmp_{rsa,dh}_callback().
9264 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
9265 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
9268 *) X509_name_add_entry() freed the wrong thing after an error.
9269 [Arne Ansper <arne@ats.cyber.ee>]
9271 *) rsa_eay.c would attempt to free a NULL context.
9272 [Arne Ansper <arne@ats.cyber.ee>]
9274 *) BIO_s_socket() had a broken should_retry() on Windoze.
9275 [Arne Ansper <arne@ats.cyber.ee>]
9277 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
9278 [Arne Ansper <arne@ats.cyber.ee>]
9280 *) Make sure the already existing X509_STORE->depth variable is initialized
9281 in X509_STORE_new(), but document the fact that this variable is still
9282 unused in the certificate verification process.
9283 [Ralf S. Engelschall]
9285 *) Fix the various library and apps files to free up pkeys obtained from
9286 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
9289 *) Fix reference counting in X509_PUBKEY_get(). This makes
9290 demos/maurice/example2.c work, amongst others, probably.
9291 [Steve Henson and Ben Laurie]
9293 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
9294 `openssl' and second, the shortcut symlinks for the `openssl <command>'
9295 are no longer created. This way we have a single and consistent command
9296 line interface `openssl <command>', similar to `cvs <command>'.
9297 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
9299 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
9300 BIT STRING wrapper always have zero unused bits.
9303 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
9306 *) Make the top-level INSTALL documentation easier to understand.
9309 *) Makefiles updated to exit if an error occurs in a sub-directory
9310 make (including if user presses ^C) [Paul Sutton]
9312 *) Make Montgomery context stuff explicit in RSA data structure.
9315 *) Fix build order of pem and err to allow for generated pem.h.
9318 *) Fix renumbering bug in X509_NAME_delete_entry().
9321 *) Enhanced the err-ins.pl script so it makes the error library number
9322 global and can add a library name. This is needed for external ASN1 and
9323 other error libraries.
9326 *) Fixed sk_insert which never worked properly.
9329 *) Fix ASN1 macros so they can handle indefinite length construted
9330 EXPLICIT tags. Some non standard certificates use these: they can now
9334 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
9335 into a single doc/ssleay.txt bundle. This way the information is still
9336 preserved but no longer messes up this directory. Now it's new room for
9337 the new set of documenation files.
9338 [Ralf S. Engelschall]
9340 *) SETs were incorrectly DER encoded. This was a major pain, because they
9341 shared code with SEQUENCEs, which aren't coded the same. This means that
9342 almost everything to do with SETs or SEQUENCEs has either changed name or
9343 number of arguments.
9344 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
9346 *) Fix test data to work with the above.
9349 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
9350 was already fixed by Eric for 0.9.1 it seems.
9351 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
9353 *) Autodetect FreeBSD3.
9356 *) Fix various bugs in Configure. This affects the following platforms:
9360 unixware-2.0-pentium
9364 *) Eliminate generated files from CVS. Reorder tests to regenerate files
9365 before they are needed.
9368 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
9372 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
9374 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
9375 changed SSLeay to OpenSSL in version strings.
9376 [Ralf S. Engelschall]
9378 *) Some fixups to the top-level documents.
9381 *) Fixed the nasty bug where rsaref.h was not found under compile-time
9382 because the symlink to include/ was missing.
9383 [Ralf S. Engelschall]
9385 *) Incorporated the popular no-RSA/DSA-only patches
9386 which allow to compile a RSA-free SSLeay.
9387 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
9389 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
9390 when "ssleay" is still not found.
9391 [Ralf S. Engelschall]
9393 *) Added more platforms to Configure: Cray T3E, HPUX 11,
9394 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
9396 *) Updated the README file.
9397 [Ralf S. Engelschall]
9399 *) Added various .cvsignore files in the CVS repository subdirs
9400 to make a "cvs update" really silent.
9401 [Ralf S. Engelschall]
9403 *) Recompiled the error-definition header files and added
9404 missing symbols to the Win32 linker tables.
9405 [Ralf S. Engelschall]
9407 *) Cleaned up the top-level documents;
9408 o new files: CHANGES and LICENSE
9409 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
9410 o merged COPYRIGHT into LICENSE
9411 o removed obsolete TODO file
9412 o renamed MICROSOFT to INSTALL.W32
9413 [Ralf S. Engelschall]
9415 *) Removed dummy files from the 0.9.1b source tree:
9416 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
9417 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
9418 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
9419 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
9420 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
9421 [Ralf S. Engelschall]
9423 *) Added various platform portability fixes.
9426 *) The Genesis of the OpenSSL rpject:
9427 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
9428 Young and Tim J. Hudson created while they were working for C2Net until
9430 [The OpenSSL Project]
9433 Changes between 0.9.0b and 0.9.1b [not released]
9435 *) Updated a few CA certificates under certs/
9438 *) Changed some BIGNUM api stuff.
9441 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
9442 DGUX x86, Linux Alpha, etc.
9445 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
9446 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
9450 *) Add -strparse option to asn1pars program which parses nested
9452 [Dr Stephen Henson <shenson@bigfoot.com>]
9454 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
9457 *) DSA fix for "ca" program.
9460 *) Added "-genkey" option to "dsaparam" program.
9463 *) Added RIPE MD160 (rmd160) message digest.
9466 *) Added -a (all) option to "ssleay version" command.
9469 *) Added PLATFORM define which is the id given to Configure.
9472 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
9475 *) Extended the ASN.1 parser routines.
9478 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
9481 *) Added a BN_CTX to the BN library.
9484 *) Fixed the weak key values in DES library
9487 *) Changed API in EVP library for cipher aliases.
9490 *) Added support for RC2/64bit cipher.
9493 *) Converted the lhash library to the crypto/mem.c functions.
9496 *) Added more recognized ASN.1 object ids.
9499 *) Added more RSA padding checks for SSL/TLS.
9502 *) Added BIO proxy/filter functionality.
9505 *) Added extra_certs to SSL_CTX which can be used
9506 send extra CA certificates to the client in the CA cert chain sending
9507 process. It can be configured with SSL_CTX_add_extra_chain_cert().
9510 *) Now Fortezza is denied in the authentication phase because
9511 this is key exchange mechanism is not supported by SSLeay at all.
9514 *) Additional PKCS1 checks.
9517 *) Support the string "TLSv1" for all TLS v1 ciphers.
9520 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
9521 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
9524 *) Fixed a few memory leaks.
9527 *) Fixed various code and comment typos.
9530 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
9531 bytes sent in the client random.
9532 [Edward Bishop <ebishop@spyglass.com>]