5 Changes between 1.0.0j and 1.0.0k [xx XXX xxxx]
7 *) Return an error when checking OCSP signatures when key is NULL.
8 This fixes a DoS attack. (CVE-2013-0166)
11 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
12 the right response is stapled. Also change SSL_get_certificate()
13 so it returns the certificate actually sent.
14 See http://rt.openssl.org/Ticket/Display.html?id=2836.
16 [Rob Stradling <rob.stradling@comodo.com>]
18 *) Fix possible deadlock when decoding public keys.
21 Changes between 1.0.0i and 1.0.0j [10 May 2012]
23 *) Sanity check record length before skipping explicit IV in DTLS
26 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
27 fuzzing as a service testing platform.
31 *) Initialise tkeylen properly when encrypting CMS messages.
32 Thanks to Solar Designer of Openwall for reporting this issue.
35 Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
37 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
38 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
39 in CRYPTO_realloc_clean.
41 Thanks to Tavis Ormandy, Google Security Team, for discovering this
42 issue and to Adam Langley <agl@chromium.org> for fixing it.
44 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
46 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
48 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
49 in CMS and PKCS7 code. When RSA decryption fails use a random key for
50 content decryption and always return the same error. Note: this attack
51 needs on average 2^20 messages so it only affects automated senders. The
52 old behaviour can be reenabled in the CMS code by setting the
53 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
54 an MMA defence is not necessary.
55 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
56 this issue. (CVE-2012-0884)
59 *) Fix CVE-2011-4619: make sure we really are receiving a
60 client hello before rejecting multiple SGC restarts. Thanks to
61 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
64 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
66 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
67 Thanks to Antonio Martin, Enterprise Secure Access Research and
68 Development, Cisco Systems, Inc. for discovering this bug and
69 preparing a fix. (CVE-2012-0050)
72 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
74 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
75 of the Vaudenay padding oracle attack on CBC mode encryption
76 which enables an efficient plaintext recovery attack against
77 the OpenSSL implementation of DTLS. Their attack exploits timing
78 differences arising during decryption processing. A research
79 paper describing this attack can be found at:
80 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
81 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
82 Security Group at Royal Holloway, University of London
83 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
84 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
85 for preparing the fix. (CVE-2011-4108)
86 [Robin Seggelmann, Michael Tuexen]
88 *) Clear bytes used for block padding of SSL 3.0 records.
90 [Adam Langley (Google)]
92 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
93 Kadianakis <desnacked@gmail.com> for discovering this issue and
94 Adam Langley for preparing the fix. (CVE-2011-4619)
95 [Adam Langley (Google)]
97 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
98 [Andrey Kulikov <amdeich@gmail.com>]
100 *) Prevent malformed RFC3779 data triggering an assertion failure.
101 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
102 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
103 [Rob Austein <sra@hactrn.net>]
105 *) Improved PRNG seeding for VOS.
106 [Paul Green <Paul.Green@stratus.com>]
108 *) Fix ssl_ciph.c set-up race.
109 [Adam Langley (Google)]
111 *) Fix spurious failures in ecdsatest.c.
112 [Emilia Käsper (Google)]
114 *) Fix the BIO_f_buffer() implementation (which was mixing different
115 interpretations of the '..._len' fields).
116 [Adam Langley (Google)]
118 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
119 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
120 threads won't reuse the same blinding coefficients.
122 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
123 lock to call BN_BLINDING_invert_ex, and avoids one use of
124 BN_BLINDING_update for each BN_BLINDING structure (previously,
125 the last update always remained unused).
126 [Emilia Käsper (Google)]
128 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
129 [Bob Buckholz (Google)]
131 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
133 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
134 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
135 [Kaspar Brand <ossl@velox.ch>]
137 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
138 for multi-threaded use of ECDH. (CVE-2011-3210)
139 [Adam Langley (Google)]
141 *) Fix x509_name_ex_d2i memory leak on bad inputs.
144 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
145 signature public key algorithm by using OID xref utilities instead.
146 Before this you could only use some ECC ciphersuites with SHA1 only.
149 *) Add protection against ECDSA timing attacks as mentioned in the paper
150 by Billy Bob Brumley and Nicola Tuveri, see:
152 http://eprint.iacr.org/2011/232.pdf
154 [Billy Bob Brumley and Nicola Tuveri]
156 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
158 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
159 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
161 *) Fix bug in string printing code: if *any* escaping is enabled we must
162 escape the escape character (backslash) or the resulting string is
166 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
168 *) Disable code workaround for ancient and obsolete Netscape browsers
169 and servers: an attacker can use it in a ciphersuite downgrade attack.
170 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
173 *) Fixed J-PAKE implementation error, originally discovered by
174 Sebastien Martini, further info and confirmation from Stefan
175 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
178 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
180 *) Fix extension code to avoid race conditions which can result in a buffer
181 overrun vulnerability: resumed sessions must not be modified as they can
182 be shared by multiple threads. CVE-2010-3864
185 *) Fix WIN32 build system to correctly link an ENGINE directory into
189 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
191 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
193 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
195 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
197 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
198 context. The operation can be customised via the ctrl mechanism in
199 case ENGINEs want to include additional functionality.
202 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
205 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
206 output hashes compatible with older versions of OpenSSL.
207 [Willy Weisz <weisz@vcpc.univie.ac.at>]
209 *) Fix compression algorithm handling: if resuming a session use the
210 compression algorithm of the resumed session instead of determining
211 it from client hello again. Don't allow server to change algorithm.
214 *) Add load_crls() function to apps tidying load_certs() too. Add option
215 to verify utility to allow additional CRLs to be included.
218 *) Update OCSP request code to permit adding custom headers to the request:
219 some responders need this.
222 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
224 [Julia Lawall <julia@diku.dk>]
226 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
227 needlessly dereferenced structures, used obsolete functions and
228 didn't handle all updated verify codes correctly.
231 *) Disable MD2 in the default configuration.
234 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
235 indicate the initial BIO being pushed or popped. This makes it possible
236 to determine whether the BIO is the one explicitly called or as a result
237 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
238 it handles reference counts correctly and doesn't zero out the I/O bio
239 when it is not being explicitly popped. WARNING: applications which
240 included workarounds for the old buggy behaviour will need to be modified
241 or they could free up already freed BIOs.
244 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
245 renaming to all platforms (within the 0.9.8 branch, this was
246 done conditionally on Netware platforms to avoid a name clash).
247 [Guenter <lists@gknw.net>]
249 *) Add ECDHE and PSK support to DTLS.
250 [Michael Tuexen <tuexen@fh-muenster.de>]
252 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
256 *) Add "missing" function EVP_MD_flags() (without this the only way to
257 retrieve a digest flags is by accessing the structure directly. Update
258 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
259 or cipher is registered as in the "from" argument. Print out all
260 registered digests in the dgst usage message instead of manually
261 attempting to work them out.
264 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
265 this allows the use of compression and extensions. Change default cipher
266 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
267 by default unless an application cipher string requests it.
270 *) Alter match criteria in PKCS12_parse(). It used to try to use local
271 key ids to find matching certificates and keys but some PKCS#12 files
272 don't follow the (somewhat unwritten) rules and this strategy fails.
273 Now just gather all certificates together and the first private key
274 then look for the first certificate that matches the key.
277 *) Support use of registered digest and cipher names for dgst and cipher
278 commands instead of having to add each one as a special case. So now
285 openssl dgst -sha256 foo
287 and this works for ENGINE based algorithms too.
291 *) Update Gost ENGINE to support parameter files.
292 [Victor B. Wagner <vitus@cryptocom.ru>]
294 *) Support GeneralizedTime in ca utility.
295 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
297 *) Enhance the hash format used for certificate directory links. The new
298 form uses the canonical encoding (meaning equivalent names will work
299 even if they aren't identical) and uses SHA1 instead of MD5. This form
300 is incompatible with the older format and as a result c_rehash should
301 be used to rebuild symbolic links.
304 *) Make PKCS#8 the default write format for private keys, replacing the
305 traditional format. This form is standardised, more secure and doesn't
306 include an implicit MD5 dependency.
309 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
310 committed to OpenSSL should pass this lot as a minimum.
313 *) Add session ticket override functionality for use by EAP-FAST.
314 [Jouni Malinen <j@w1.fi>]
316 *) Modify HMAC functions to return a value. Since these can be implemented
317 in an ENGINE errors can occur.
320 *) Type-checked OBJ_bsearch_ex.
323 *) Type-checked OBJ_bsearch. Also some constification necessitated
324 by type-checking. Still to come: TXT_DB, bsearch(?),
325 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
329 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
330 seconds to a tm structure directly, instead of going through OS
331 specific date routines. This avoids any issues with OS routines such
332 as the year 2038 bug. New *_adj() functions for ASN1 time structures
333 and X509_time_adj_ex() to cover the extended range. The existing
334 X509_time_adj() is still usable and will no longer have any date issues.
337 *) Delta CRL support. New use deltas option which will attempt to locate
338 and search any appropriate delta CRLs available.
340 This work was sponsored by Google.
343 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
344 code and add additional score elements. Validate alternate CRL paths
345 as part of the CRL checking and indicate a new error "CRL path validation
346 error" in this case. Applications wanting additional details can use
347 the verify callback and check the new "parent" field. If this is not
348 NULL CRL path validation is taking place. Existing applications wont
349 see this because it requires extended CRL support which is off by
352 This work was sponsored by Google.
355 *) Support for freshest CRL extension.
357 This work was sponsored by Google.
360 *) Initial indirect CRL support. Currently only supported in the CRLs
361 passed directly and not via lookup. Process certificate issuer
362 CRL entry extension and lookup CRL entries by bother issuer name
363 and serial number. Check and process CRL issuer entry in IDP extension.
365 This work was sponsored by Google.
368 *) Add support for distinct certificate and CRL paths. The CRL issuer
369 certificate is validated separately in this case. Only enabled if
370 an extended CRL support flag is set: this flag will enable additional
371 CRL functionality in future.
373 This work was sponsored by Google.
376 *) Add support for policy mappings extension.
378 This work was sponsored by Google.
381 *) Fixes to pathlength constraint, self issued certificate handling,
382 policy processing to align with RFC3280 and PKITS tests.
384 This work was sponsored by Google.
387 *) Support for name constraints certificate extension. DN, email, DNS
388 and URI types are currently supported.
390 This work was sponsored by Google.
393 *) To cater for systems that provide a pointer-based thread ID rather
394 than numeric, deprecate the current numeric thread ID mechanism and
395 replace it with a structure and associated callback type. This
396 mechanism allows a numeric "hash" to be extracted from a thread ID in
397 either case, and on platforms where pointers are larger than 'long',
398 mixing is done to help ensure the numeric 'hash' is usable even if it
399 can't be guaranteed unique. The default mechanism is to use "&errno"
400 as a pointer-based thread ID to distinguish between threads.
402 Applications that want to provide their own thread IDs should now use
403 CRYPTO_THREADID_set_callback() to register a callback that will call
404 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
406 Note that ERR_remove_state() is now deprecated, because it is tied
407 to the assumption that thread IDs are numeric. ERR_remove_state(0)
408 to free the current thread's error state should be replaced by
409 ERR_remove_thread_state(NULL).
411 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
412 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
413 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
414 application was previously providing a numeric thread callback that
415 was inappropriate for distinguishing threads, then uniqueness might
416 have been obtained with &errno that happened immediately in the
417 intermediate development versions of OpenSSL; this is no longer the
418 case, the numeric thread callback will now override the automatic use
420 [Geoff Thorpe, with help from Bodo Moeller]
422 *) Initial support for different CRL issuing certificates. This covers a
423 simple case where the self issued certificates in the chain exist and
424 the real CRL issuer is higher in the existing chain.
426 This work was sponsored by Google.
429 *) Removed effectively defunct crypto/store from the build.
432 *) Revamp of STACK to provide stronger type-checking. Still to come:
433 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
434 ASN1_STRING, CONF_VALUE.
437 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
438 RAM on SSL connections. This option can save about 34k per idle SSL.
441 *) Revamp of LHASH to provide stronger type-checking. Still to come:
442 STACK, TXT_DB, bsearch, qsort.
445 *) Initial support for Cryptographic Message Syntax (aka CMS) based
446 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
447 support for data, signedData, compressedData, digestedData and
448 encryptedData, envelopedData types included. Scripts to check against
449 RFC4134 examples draft and interop and consistency checks of many
450 content types and variants.
453 *) Add options to enc utility to support use of zlib compression BIO.
456 *) Extend mk1mf to support importing of options and assembly language
457 files from Configure script, currently only included in VC-WIN32.
458 The assembly language rules can now optionally generate the source
459 files from the associated perl scripts.
462 *) Implement remaining functionality needed to support GOST ciphersuites.
463 Interop testing has been performed using CryptoPro implementations.
464 [Victor B. Wagner <vitus@cryptocom.ru>]
466 *) s390x assembler pack.
469 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
473 *) Implement Opaque PRF Input TLS extension as specified in
474 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
475 official specification yet and no extension type assignment by
476 IANA exists, this extension (for now) will have to be explicitly
477 enabled when building OpenSSL by providing the extension number
478 to use. For example, specify an option
480 -DTLSEXT_TYPE_opaque_prf_input=0x9527
482 to the "config" or "Configure" script to enable the extension,
483 assuming extension number 0x9527 (which is a completely arbitrary
484 and unofficial assignment based on the MD5 hash of the Internet
485 Draft). Note that by doing so, you potentially lose
486 interoperability with other TLS implementations since these might
487 be using the same extension number for other purposes.
489 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
490 opaque PRF input value to use in the handshake. This will create
491 an interal copy of the length-'len' string at 'src', and will
492 return non-zero for success.
494 To get more control and flexibility, provide a callback function
497 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
498 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
502 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
505 Callback function 'cb' will be called in handshakes, and is
506 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
507 Argument 'arg' is for application purposes (the value as given to
508 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
509 be provided to the callback function). The callback function
510 has to return non-zero to report success: usually 1 to use opaque
511 PRF input just if possible, or 2 to enforce use of the opaque PRF
512 input. In the latter case, the library will abort the handshake
513 if opaque PRF input is not successfully negotiated.
515 Arguments 'peerinput' and 'len' given to the callback function
516 will always be NULL and 0 in the case of a client. A server will
517 see the client's opaque PRF input through these variables if
518 available (NULL and 0 otherwise). Note that if the server
519 provides an opaque PRF input, the length must be the same as the
520 length of the client's opaque PRF input.
522 Note that the callback function will only be called when creating
523 a new session (session resumption can resume whatever was
524 previously negotiated), and will not be called in SSL 2.0
525 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
526 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
527 for applications that need to enforce opaque PRF input.
531 *) Update ssl code to support digests other than SHA1+MD5 for handshake
534 [Victor B. Wagner <vitus@cryptocom.ru>]
536 *) Add RFC4507 support to OpenSSL. This includes the corrections in
537 RFC4507bis. The encrypted ticket format is an encrypted encoded
538 SSL_SESSION structure, that way new session features are automatically
541 If a client application caches session in an SSL_SESSION structure
542 support is transparent because tickets are now stored in the encoded
545 The SSL_CTX structure automatically generates keys for ticket
546 protection in servers so again support should be possible
547 with no application modification.
549 If a client or server wishes to disable RFC4507 support then the option
550 SSL_OP_NO_TICKET can be set.
552 Add a TLS extension debugging callback to allow the contents of any client
553 or server extensions to be examined.
555 This work was sponsored by Google.
558 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
559 OpenSSL should now compile cleanly on gcc 4.2
560 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
562 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
563 support including streaming MAC support: this is required for GOST
565 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
567 *) Add option -stream to use PKCS#7 streaming in smime utility. New
568 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
569 to output in BER and PEM format.
572 *) Experimental support for use of HMAC via EVP_PKEY interface. This
573 allows HMAC to be handled via the EVP_DigestSign*() interface. The
574 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
575 ENGINE support for HMAC keys which are unextractable. New -mac and
576 -macopt options to dgst utility.
579 *) New option -sigopt to dgst utility. Update dgst to use
580 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
581 alternative signing paramaters such as X9.31 or PSS in the dgst
585 *) Change ssl_cipher_apply_rule(), the internal function that does
586 the work each time a ciphersuite string requests enabling
587 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
588 removing ("!foo+bar") a class of ciphersuites: Now it maintains
589 the order of disabled ciphersuites such that those ciphersuites
590 that most recently went from enabled to disabled not only stay
591 in order with respect to each other, but also have higher priority
592 than other disabled ciphersuites the next time ciphersuites are
595 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
596 the same ciphersuites as with "HIGH" alone, but in a specific
597 order where the PSK ciphersuites come first (since they are the
598 most recently disabled ciphersuites when "HIGH" is parsed).
600 Also, change ssl_create_cipher_list() (using this new
601 funcionality) such that between otherwise identical
602 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
606 *) Change ssl_create_cipher_list() so that it automatically
607 arranges the ciphersuites in reasonable order before starting
608 to process the rule string. Thus, the definition for "DEFAULT"
609 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
610 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
611 This makes it much easier to arrive at a reasonable default order
612 in applications for which anonymous ciphers are OK (meaning
613 that you can't actually use DEFAULT).
614 [Bodo Moeller; suggested by Victor Duchovni]
616 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
617 processing) into multiple integers instead of setting
618 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
619 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
620 (These masks as well as the individual bit definitions are hidden
621 away into the non-exported interface ssl/ssl_locl.h, so this
622 change to the definition of the SSL_CIPHER structure shouldn't
623 affect applications.) This give us more bits for each of these
624 categories, so there is no longer a need to coagulate AES128 and
625 AES256 into a single algorithm bit, and to coagulate Camellia128
626 and Camellia256 into a single algorithm bit, which has led to all
629 Thus, among other things, the kludge introduced in 0.9.7m and
630 0.9.8e for masking out AES256 independently of AES128 or masking
631 out Camellia256 independently of AES256 is not needed here in 0.9.9.
633 With the change, we also introduce new ciphersuite aliases that
634 so far were missing: "AES128", "AES256", "CAMELLIA128", and
638 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
639 Use the leftmost N bytes of the signature input if the input is
640 larger than the prime q (with N being the size in bytes of q).
643 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
644 it yet and it is largely untested.
647 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
650 *) Initial incomplete changes to avoid need for function casts in OpenSSL
651 some compilers (gcc 4.2 and later) reject their use. Safestack is
652 reimplemented. Update ASN1 to avoid use of legacy functions.
655 *) Win32/64 targets are linked with Winsock2.
658 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
659 to external functions. This can be used to increase CRL handling
660 efficiency especially when CRLs are very large by (for example) storing
661 the CRL revoked certificates in a database.
664 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
665 new CRLs added to a directory can be used. New command line option
666 -verify_return_error to s_client and s_server. This causes real errors
667 to be returned by the verify callback instead of carrying on no matter
668 what. This reflects the way a "real world" verify callback would behave.
671 *) GOST engine, supporting several GOST algorithms and public key formats.
672 Kindly donated by Cryptocom.
675 *) Partial support for Issuing Distribution Point CRL extension. CRLs
676 partitioned by DP are handled but no indirect CRL or reason partitioning
677 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
678 selected via a scoring technique which handles IDP and AKID in CRLs.
681 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
682 will ultimately be used for all verify operations: this will remove the
683 X509_STORE dependency on certificate verification and allow alternative
684 lookup methods. X509_STORE based implementations of these two callbacks.
687 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
688 Modify get_crl() to find a valid (unexpired) CRL if possible.
691 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
692 this would be called X509_CRL_cmp() but that name is already used by
693 a function that just compares CRL issuer names. Cache several CRL
694 extensions in X509_CRL structure and cache CRLDP in X509.
697 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
698 this maps equivalent X509_NAME structures into a consistent structure.
699 Name comparison can then be performed rapidly using memcmp().
702 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
706 *) Allow digests to supply their own micalg string for S/MIME type using
707 the ctrl EVP_MD_CTRL_MICALG.
710 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
711 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
712 ctrl. It can then customise the structure before and/or after signing
716 *) New function OBJ_add_sigid() to allow application defined signature OIDs
717 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
718 to free up any added signature OIDs.
721 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
722 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
723 digest and cipher tables. New options added to openssl utility:
724 list-message-digest-algorithms and list-cipher-algorithms.
727 *) Change the array representation of binary polynomials: the list
728 of degrees of non-zero coefficients is now terminated with -1.
729 Previously it was terminated with 0, which was also part of the
730 value; thus, the array representation was not applicable to
731 polynomials where t^0 has coefficient zero. This change makes
732 the array representation useful in a more general context.
735 *) Various modifications and fixes to SSL/TLS cipher string
736 handling. For ECC, the code now distinguishes between fixed ECDH
737 with RSA certificates on the one hand and with ECDSA certificates
738 on the other hand, since these are separate ciphersuites. The
739 unused code for Fortezza ciphersuites has been removed.
741 For consistency with EDH, ephemeral ECDH is now called "EECDH"
742 (not "ECDHE"). For consistency with the code for DH
743 certificates, use of ECDH certificates is now considered ECDH
744 authentication, not RSA or ECDSA authentication (the latter is
745 merely the CA's signing algorithm and not actively used in the
748 The temporary ciphersuite alias "ECCdraft" is no longer
749 available, and ECC ciphersuites are no longer excluded from "ALL"
750 and "DEFAULT". The following aliases now exist for RFC 4492
751 ciphersuites, most of these by analogy with the DH case:
753 kECDHr - ECDH cert, signed with RSA
754 kECDHe - ECDH cert, signed with ECDSA
755 kECDH - ECDH cert (signed with either RSA or ECDSA)
756 kEECDH - ephemeral ECDH
757 ECDH - ECDH cert or ephemeral ECDH
763 AECDH - anonymous ECDH
764 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
768 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
769 Use correct micalg parameters depending on digest(s) in signed message.
772 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
773 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
776 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
777 an engine to register a method. Add ENGINE lookups for methods and
778 functional reference processing.
781 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
782 EVP_{Sign,Verify}* which allow an application to customise the signature
786 *) New -resign option to smime utility. This adds one or more signers
787 to an existing PKCS#7 signedData structure. Also -md option to use an
788 alternative message digest algorithm for signing.
791 *) Tidy up PKCS#7 routines and add new functions to make it easier to
792 create PKCS7 structures containing multiple signers. Update smime
793 application to support multiple signers.
796 *) New -macalg option to pkcs12 utility to allow setting of an alternative
800 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
801 Reorganize PBE internals to lookup from a static table using NIDs,
802 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
803 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
804 PRF which will be automatically used with PBES2.
807 *) Replace the algorithm specific calls to generate keys in "req" with the
811 *) Update PKCS#7 enveloped data routines to use new API. This is now
812 supported by any public key method supporting the encrypt operation. A
813 ctrl is added to allow the public key algorithm to examine or modify
814 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
818 *) Add a ctrl to asn1 method to allow a public key algorithm to express
819 a default digest type to use. In most cases this will be SHA1 but some
820 algorithms (such as GOST) need to specify an alternative digest. The
821 return value indicates how strong the prefernce is 1 means optional and
822 2 is mandatory (that is it is the only supported type). Modify
823 ASN1_item_sign() to accept a NULL digest argument to indicate it should
824 use the default md. Update openssl utilities to use the default digest
825 type for signing if it is not explicitly indicated.
828 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
829 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
830 signing method from the key type. This effectively removes the link
831 between digests and public key types.
834 *) Add an OID cross reference table and utility functions. Its purpose is to
835 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
836 rsaEncryption. This will allow some of the algorithm specific hackery
837 needed to use the correct OID to be removed.
840 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
841 structures for PKCS7_sign(). They are now set up by the relevant public
845 *) Add provisional EC pkey method with support for ECDSA and ECDH.
848 *) Add support for key derivation (agreement) in the API, DH method and
852 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
853 public and private key formats. As a side effect these add additional
854 command line functionality not previously available: DSA signatures can be
855 generated and verified using pkeyutl and DH key support and generation in
860 [Oliver Tappe <zooey@hirschkaefer.de>]
862 *) New make target "install_html_docs" installs HTML renditions of the
864 [Oliver Tappe <zooey@hirschkaefer.de>]
866 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
867 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
868 support key and parameter generation and add initial key generation
869 functionality for RSA.
872 *) Add functions for main EVP_PKEY_method operations. The undocumented
873 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
874 EVP_PKEY_{encrypt,decrypt}_old.
877 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
878 key API, doesn't do much yet.
881 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
882 public key algorithms. New option to openssl utility:
883 "list-public-key-algorithms" to print out info.
886 *) Implement the Supported Elliptic Curves Extension for
887 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
890 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
891 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
894 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
895 utilities such as rsa, dsa, dsaparam etc except they process any key
899 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
900 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
901 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
905 *) Initial support for pluggable public key ASN1.
906 De-spaghettify the public key ASN1 handling. Move public and private
907 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
908 algorithm specific handling to a single module within the relevant
909 algorithm directory. Add functions to allow (near) opaque processing
910 of public and private key structures.
913 *) Implement the Supported Point Formats Extension for
914 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
917 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
918 for the psk identity [hint] and the psk callback functions to the
919 SSL_SESSION, SSL and SSL_CTX structure.
922 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
926 SSL_CTX_use_psk_identity_hint
927 SSL_get_psk_identity_hint
929 SSL_use_psk_identity_hint
931 [Mika Kousa and Pasi Eronen of Nokia Corporation]
933 *) Add RFC 3161 compliant time stamp request creation, response generation
934 and response verification functionality.
935 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
937 *) Add initial support for TLS extensions, specifically for the server_name
938 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
939 have new members for a host name. The SSL data structure has an
940 additional member SSL_CTX *initial_ctx so that new sessions can be
941 stored in that context to allow for session resumption, even after the
942 SSL has been switched to a new SSL_CTX in reaction to a client's
943 server_name extension.
945 New functions (subject to change):
948 SSL_get_servername_type()
951 New CTRL codes and macros (subject to change):
953 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
954 - SSL_CTX_set_tlsext_servername_callback()
955 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
956 - SSL_CTX_set_tlsext_servername_arg()
957 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
959 openssl s_client has a new '-servername ...' option.
961 openssl s_server has new options '-servername_host ...', '-cert2 ...',
962 '-key2 ...', '-servername_fatal' (subject to change). This allows
963 testing the HostName extension for a specific single host name ('-cert'
964 and '-key' remain fallbacks for handshakes without HostName
965 negotiation). If the unrecogninzed_name alert has to be sent, this by
966 default is a warning; it becomes fatal with the '-servername_fatal'
969 [Peter Sylvester, Remy Allais, Christophe Renou]
971 *) Whirlpool hash implementation is added.
974 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
975 bn(64,32). Because of instruction set limitations it doesn't have
976 any negative impact on performance. This was done mostly in order
977 to make it possible to share assembler modules, such as bn_mul_mont
978 implementations, between 32- and 64-bit builds without hassle.
981 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
982 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
986 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
987 dedicated Montgomery multiplication procedure, is introduced.
988 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
989 "64-bit" performance on certain 32-bit targets.
992 *) New option SSL_OP_NO_COMP to disable use of compression selectively
993 in SSL structures. New SSL ctrl to set maximum send fragment size.
994 Save memory by seeting the I/O buffer sizes dynamically instead of
995 using the maximum available value.
998 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
999 in addition to the text details.
1002 *) Very, very preliminary EXPERIMENTAL support for printing of general
1003 ASN1 structures. This currently produces rather ugly output and doesn't
1004 handle several customised structures at all.
1007 *) Integrated support for PVK file format and some related formats such
1008 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1009 these in the 'rsa' and 'dsa' utilities.
1012 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1015 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1016 place for the (very old) "NETSCAPE" format certificates which are now
1017 handled using new ASN1 code equivalents.
1020 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1021 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1022 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1025 *) Modify CRL distribution points extension code to print out previously
1026 unsupported fields. Enhance extension setting code to allow setting of
1030 *) Add print and set support for Issuing Distribution Point CRL extension.
1033 *) Change 'Configure' script to enable Camellia by default.
1036 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
1038 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1039 Thanks to Antonio Martin, Enterprise Secure Access Research and
1040 Development, Cisco Systems, Inc. for discovering this bug and
1041 preparing a fix. (CVE-2012-0050)
1044 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
1046 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1047 of the Vaudenay padding oracle attack on CBC mode encryption
1048 which enables an efficient plaintext recovery attack against
1049 the OpenSSL implementation of DTLS. Their attack exploits timing
1050 differences arising during decryption processing. A research
1051 paper describing this attack can be found at:
1052 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1053 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1054 Security Group at Royal Holloway, University of London
1055 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1056 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1057 for preparing the fix. (CVE-2011-4108)
1058 [Robin Seggelmann, Michael Tuexen]
1060 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
1061 [Ben Laurie, Kasper <ekasper@google.com>]
1063 *) Clear bytes used for block padding of SSL 3.0 records.
1065 [Adam Langley (Google)]
1067 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1068 Kadianakis <desnacked@gmail.com> for discovering this issue and
1069 Adam Langley for preparing the fix. (CVE-2011-4619)
1070 [Adam Langley (Google)]
1072 *) Prevent malformed RFC3779 data triggering an assertion failure.
1073 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1074 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1075 [Rob Austein <sra@hactrn.net>]
1077 *) Fix ssl_ciph.c set-up race.
1078 [Adam Langley (Google)]
1080 *) Fix spurious failures in ecdsatest.c.
1081 [Emilia Käsper (Google)]
1083 *) Fix the BIO_f_buffer() implementation (which was mixing different
1084 interpretations of the '..._len' fields).
1085 [Adam Langley (Google)]
1087 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1088 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1089 threads won't reuse the same blinding coefficients.
1091 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1092 lock to call BN_BLINDING_invert_ex, and avoids one use of
1093 BN_BLINDING_update for each BN_BLINDING structure (previously,
1094 the last update always remained unused).
1095 [Emilia Käsper (Google)]
1097 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1098 for multi-threaded use of ECDH.
1099 [Adam Langley (Google)]
1101 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1104 *) Add protection against ECDSA timing attacks as mentioned in the paper
1105 by Billy Bob Brumley and Nicola Tuveri, see:
1107 http://eprint.iacr.org/2011/232.pdf
1109 [Billy Bob Brumley and Nicola Tuveri]
1111 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
1113 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1114 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1116 *) Fix bug in string printing code: if *any* escaping is enabled we must
1117 escape the escape character (backslash) or the resulting string is
1121 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
1123 *) Disable code workaround for ancient and obsolete Netscape browsers
1124 and servers: an attacker can use it in a ciphersuite downgrade attack.
1125 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1128 *) Fixed J-PAKE implementation error, originally discovered by
1129 Sebastien Martini, further info and confirmation from Stefan
1130 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1133 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
1135 *) Fix extension code to avoid race conditions which can result in a buffer
1136 overrun vulnerability: resumed sessions must not be modified as they can
1137 be shared by multiple threads. CVE-2010-3864
1140 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
1143 *) Don't reencode certificate when calculating signature: cache and use
1144 the original encoding instead. This makes signature verification of
1145 some broken encodings work correctly.
1148 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
1149 is also one of the inputs.
1150 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1152 *) Don't repeatedly append PBE algorithms to table if they already exist.
1153 Sort table on each new add. This effectively makes the table read only
1154 after all algorithms are added and subsequent calls to PKCS12_pbe_add
1158 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
1160 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
1163 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
1164 access or freeing data twice (CVE-2010-0742)
1165 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
1167 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
1168 common in certificates and some applications which only call
1169 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
1173 Reduce copying into .apps and .test in makevms.com
1174 Don't try to use blank CA certificate in CA.com
1175 Allow use of C files from original directories in maketests.com
1176 [Steven M. Schweda" <sms@antinode.info>]
1178 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1180 *) When rejecting SSL/TLS records due to an incorrect version number, never
1181 update s->server with a new major version number. As of
1182 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1183 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1184 the previous behavior could result in a read attempt at NULL when
1185 receiving specific incorrect SSL/TLS records once record payload
1186 protection is active. (CVE-2010-0740)
1187 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1189 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1190 could be crashed if the relevant tables were not present (e.g. chrooted).
1191 [Tomas Hoger <thoger@redhat.com>]
1193 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1195 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1196 [Martin Olsson, Neel Mehta]
1198 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1199 accommodate for stack sorting, always a write lock!).
1202 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1203 excessive delays in the RAND_poll(): over a minute. As a workaround
1204 include a time check in the inner Heap32Next loop too.
1207 *) The code that handled flushing of data in SSL/TLS originally used the
1208 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1209 the problem outlined in PR#1949. The fix suggested there however can
1210 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1211 of Apache). So instead simplify the code to flush unconditionally.
1212 This should be fine since flushing with no data to flush is a no op.
1215 *) Handle TLS versions 2.0 and later properly and correctly use the
1216 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1217 off ancient servers have a habit of sticking around for a while...
1220 *) Modify compression code so it frees up structures without using the
1221 ex_data callbacks. This works around a problem where some applications
1222 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1223 restarting) then use compression (e.g. SSL with compression) later.
1224 This results in significant per-connection memory leaks and
1225 has caused some security issues including CVE-2008-1678 and
1229 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1230 change when encrypting or decrypting.
1233 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1234 connect and renegotiate with servers which do not support RI.
1235 Until RI is more widely deployed this option is enabled by default.
1238 *) Add "missing" ssl ctrls to clear options and mode.
1241 *) If client attempts to renegotiate and doesn't support RI respond with
1242 a no_renegotiation alert as required by RFC5746. Some renegotiating
1243 TLS clients will continue a connection gracefully when they receive
1244 the alert. Unfortunately OpenSSL mishandled this alert and would hang
1245 waiting for a server hello which it will never receive. Now we treat a
1246 received no_renegotiation alert as a fatal error. This is because
1247 applications requesting a renegotiation might well expect it to succeed
1248 and would have no code in place to handle the server denying it so the
1249 only safe thing to do is to terminate the connection.
1252 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
1253 peer supports secure renegotiation and 0 otherwise. Print out peer
1254 renegotiation support in s_client/s_server.
1257 *) Replace the highly broken and deprecated SPKAC certification method with
1258 the updated NID creation version. This should correctly handle UTF8.
1261 *) Implement RFC5746. Re-enable renegotiation but require the extension
1262 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
1263 turns out to be a bad idea. It has been replaced by
1264 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
1265 SSL_CTX_set_options(). This is really not recommended unless you
1266 know what you are doing.
1267 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
1269 *) Fixes to stateless session resumption handling. Use initial_ctx when
1270 issuing and attempting to decrypt tickets in case it has changed during
1271 servername handling. Use a non-zero length session ID when attempting
1272 stateless session resumption: this makes it possible to determine if
1273 a resumption has occurred immediately after receiving server hello
1274 (several places in OpenSSL subtly assume this) instead of later in
1278 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
1279 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
1280 fixes for a few places where the return code is not checked
1282 [Julia Lawall <julia@diku.dk>]
1284 *) Add --strict-warnings option to Configure script to include devteam
1285 warnings in other configurations.
1288 *) Add support for --libdir option and LIBDIR variable in makefiles. This
1289 makes it possible to install openssl libraries in locations which
1290 have names other than "lib", for example "/usr/lib64" which some
1292 [Steve Henson, based on patch from Jeremy Utley]
1294 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
1295 X690 8.9.12 and can produce some misleading textual output of OIDs.
1296 [Steve Henson, reported by Dan Kaminsky]
1298 *) Delete MD2 from algorithm tables. This follows the recommendation in
1299 several standards that it is not used in new applications due to
1300 several cryptographic weaknesses. For binary compatibility reasons
1301 the MD2 API is still compiled in by default.
1304 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
1308 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
1309 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
1311 [Guenter <lists@gknw.net>]
1313 *) Fix the server certificate chain building code to use X509_verify_cert(),
1314 it used to have an ad-hoc builder which was unable to cope with anything
1315 other than a simple chain.
1316 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
1318 *) Don't check self signed certificate signatures in X509_verify_cert()
1319 by default (a flag can override this): it just wastes time without
1320 adding any security. As a useful side effect self signed root CAs
1321 with non-FIPS digests are now usable in FIPS mode.
1324 *) In dtls1_process_out_of_seq_message() the check if the current message
1325 is already buffered was missing. For every new message was memory
1326 allocated, allowing an attacker to perform an denial of service attack
1327 with sending out of seq handshake messages until there is no memory
1328 left. Additionally every future messege was buffered, even if the
1329 sequence number made no sense and would be part of another handshake.
1330 So only messages with sequence numbers less than 10 in advance will be
1331 buffered. (CVE-2009-1378)
1332 [Robin Seggelmann, discovered by Daniel Mentz]
1334 *) Records are buffered if they arrive with a future epoch to be
1335 processed after finishing the corresponding handshake. There is
1336 currently no limitation to this buffer allowing an attacker to perform
1337 a DOS attack with sending records with future epochs until there is no
1338 memory left. This patch adds the pqueue_size() function to detemine
1339 the size of a buffer and limits the record buffer to 100 entries.
1341 [Robin Seggelmann, discovered by Daniel Mentz]
1343 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
1344 parent structure is freed. (CVE-2009-1379)
1347 *) Handle non-blocking I/O properly in SSL_shutdown() call.
1348 [Darryl Miles <darryl-mailinglists@netbauds.net>]
1351 [Ilya O. <vrghost@gmail.com>]
1353 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
1355 *) Disable renegotiation completely - this fixes a severe security
1356 problem (CVE-2009-3555) at the cost of breaking all
1357 renegotiation. Renegotiation can be re-enabled by setting
1358 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
1359 run-time. This is really not recommended unless you know what
1363 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
1365 *) Don't set val to NULL when freeing up structures, it is freed up by
1366 underlying code. If sizeof(void *) > sizeof(long) this can result in
1367 zeroing past the valid field. (CVE-2009-0789)
1368 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
1370 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
1371 checked correctly. This would allow some invalid signed attributes to
1372 appear to verify correctly. (CVE-2009-0591)
1373 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1375 *) Reject UniversalString and BMPString types with invalid lengths. This
1376 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
1377 a legal length. (CVE-2009-0590)
1380 *) Set S/MIME signing as the default purpose rather than setting it
1381 unconditionally. This allows applications to override it at the store
1385 *) Permit restricted recursion of ASN1 strings. This is needed in practice
1386 to handle some structures.
1389 *) Improve efficiency of mem_gets: don't search whole buffer each time
1391 [Jeremy Shapiro <jnshapir@us.ibm.com>]
1393 *) New -hex option for openssl rand.
1396 *) Print out UTF8String and NumericString when parsing ASN1.
1399 *) Support NumericString type for name components.
1402 *) Allow CC in the environment to override the automatically chosen
1403 compiler. Note that nothing is done to ensure flags work with the
1407 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
1409 *) Properly check EVP_VerifyFinal() and similar return values
1411 [Ben Laurie, Bodo Moeller, Google Security Team]
1413 *) Enable TLS extensions by default.
1416 *) Allow the CHIL engine to be loaded, whether the application is
1417 multithreaded or not. (This does not release the developer from the
1418 obligation to set up the dynamic locking callbacks.)
1419 [Sander Temme <sander@temme.net>]
1421 *) Use correct exit code if there is an error in dgst command.
1422 [Steve Henson; problem pointed out by Roland Dirlewanger]
1424 *) Tweak Configure so that you need to say "experimental-jpake" to enable
1425 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
1428 *) Add experimental JPAKE support, including demo authentication in
1429 s_client and s_server.
1432 *) Set the comparison function in v3_addr_canonize().
1433 [Rob Austein <sra@hactrn.net>]
1435 *) Add support for XMPP STARTTLS in s_client.
1436 [Philip Paeps <philip@freebsd.org>]
1438 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
1439 to ensure that even with this option, only ciphersuites in the
1440 server's preference list will be accepted. (Note that the option
1441 applies only when resuming a session, so the earlier behavior was
1442 just about the algorithm choice for symmetric cryptography.)
1445 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
1447 *) Fix NULL pointer dereference if a DTLS server received
1448 ChangeCipherSpec as first record (CVE-2009-1386).
1451 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
1452 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
1455 *) The fix in 0.9.8c that supposedly got rid of unsafe
1456 double-checked locking was incomplete for RSA blinding,
1457 addressing just one layer of what turns out to have been
1458 doubly unsafe triple-checked locking.
1460 So now fix this for real by retiring the MONT_HELPER macro
1461 in crypto/rsa/rsa_eay.c.
1463 [Bodo Moeller; problem pointed out by Marius Schilder]
1465 *) Various precautionary measures:
1467 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
1469 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
1470 (NB: This would require knowledge of the secret session ticket key
1471 to exploit, in which case you'd be SOL either way.)
1473 - Change bn_nist.c so that it will properly handle input BIGNUMs
1474 outside the expected range.
1476 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
1479 [Neel Mehta, Bodo Moeller]
1481 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
1482 the load fails. Useful for distros.
1483 [Ben Laurie and the FreeBSD team]
1485 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
1488 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
1491 *) Expand ENGINE to support engine supplied SSL client certificate functions.
1493 This work was sponsored by Logica.
1496 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
1497 keystores. Support for SSL/TLS client authentication too.
1498 Not compiled unless enable-capieng specified to Configure.
1500 This work was sponsored by Logica.
1503 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
1504 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
1505 attribute creation routines such as certifcate requests and PKCS#12
1509 Changes between 0.9.8g and 0.9.8h [28 May 2008]
1511 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
1512 handshake which could lead to a cilent crash as found using the
1513 Codenomicon TLS test suite (CVE-2008-1672)
1514 [Steve Henson, Mark Cox]
1516 *) Fix double free in TLS server name extensions which could lead to
1517 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
1520 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
1522 Clear the error queue to ensure that error entries left from
1523 older function calls do not interfere with the correct operation.
1524 [Lutz Jaenicke, Erik de Castro Lopo]
1526 *) Remove root CA certificates of commercial CAs:
1528 The OpenSSL project does not recommend any specific CA and does not
1529 have any policy with respect to including or excluding any CA.
1530 Therefore it does not make any sense to ship an arbitrary selection
1531 of root CA certificates with the OpenSSL software.
1534 *) RSA OAEP patches to fix two separate invalid memory reads.
1535 The first one involves inputs when 'lzero' is greater than
1536 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
1537 before the beginning of from). The second one involves inputs where
1538 the 'db' section contains nothing but zeroes (there is a one-byte
1539 invalid read after the end of 'db').
1540 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1542 *) Partial backport from 0.9.9-dev:
1544 Introduce bn_mul_mont (dedicated Montgomery multiplication
1545 procedure) as a candidate for BIGNUM assembler implementation.
1546 While 0.9.9-dev uses assembler for various architectures, only
1547 x86_64 is available by default here in the 0.9.8 branch, and
1548 32-bit x86 is available through a compile-time setting.
1550 To try the 32-bit x86 assembler implementation, use Configure
1551 option "enable-montasm" (which exists only for this backport).
1553 As "enable-montasm" for 32-bit x86 disclaims code stability
1554 anyway, in this constellation we activate additional code
1555 backported from 0.9.9-dev for further performance improvements,
1556 namely BN_from_montgomery_word. (To enable this otherwise,
1557 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
1559 [Andy Polyakov (backport partially by Bodo Moeller)]
1561 *) Add TLS session ticket callback. This allows an application to set
1562 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
1563 values. This is useful for key rollover for example where several key
1564 sets may exist with different names.
1567 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
1568 This was broken until now in 0.9.8 releases, such that the only way
1569 a registered ENGINE could be used (assuming it initialises
1570 successfully on the host) was to explicitly set it as the default
1571 for the relevant algorithms. This is in contradiction with 0.9.7
1572 behaviour and the documentation. With this fix, when an ENGINE is
1573 registered into a given algorithm's table of implementations, the
1574 'uptodate' flag is reset so that auto-discovery will be used next
1575 time a new context for that algorithm attempts to select an
1577 [Ian Lister (tweaked by Geoff Thorpe)]
1579 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
1580 implemention in the following ways:
1582 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
1585 Lack of BER streaming support means one pass streaming processing is
1586 only supported if data is detached: setting the streaming flag is
1587 ignored for embedded content.
1589 CMS support is disabled by default and must be explicitly enabled
1590 with the enable-cms configuration option.
1593 *) Update the GMP engine glue to do direct copies between BIGNUM and
1594 mpz_t when openssl and GMP use the same limb size. Otherwise the
1595 existing "conversion via a text string export" trick is still used.
1596 [Paul Sheer <paulsheer@gmail.com>]
1598 *) Zlib compression BIO. This is a filter BIO which compressed and
1599 uncompresses any data passed through it.
1602 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
1603 RFC3394 compatible AES key wrapping.
1606 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
1607 sets string data without copying. X509_ALGOR_set0() and
1608 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
1609 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
1610 from an X509_ATTRIBUTE structure optionally checking it occurs only
1611 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
1615 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
1616 to get the expected BN_FLG_CONSTTIME behavior.
1617 [Bodo Moeller (Google)]
1621 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
1622 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
1623 - added some more tests to do_tests.pl
1624 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
1625 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
1626 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
1627 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
1628 - various changes to netware.pl to enable gcc-cross builds on Win32
1630 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
1631 - various changes to fix missing prototype warnings
1632 - fixed x86nasm.pl to create correct asm files for NASM COFF output
1633 - added AES, WHIRLPOOL and CPUID assembler code to build files
1634 - added missing AES assembler make rules to mk1mf.pl
1635 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
1636 [Guenter Knauf <eflash@gmx.net>]
1638 *) Implement certificate status request TLS extension defined in RFC3546.
1639 A client can set the appropriate parameters and receive the encoded
1640 OCSP response via a callback. A server can query the supplied parameters
1641 and set the encoded OCSP response in the callback. Add simplified examples
1642 to s_client and s_server.
1645 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
1647 *) Fix various bugs:
1648 + Binary incompatibility of ssl_ctx_st structure
1649 + DTLS interoperation with non-compliant servers
1650 + Don't call get_session_cb() without proposed session
1651 + Fix ia64 assembler code
1652 [Andy Polyakov, Steve Henson]
1654 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
1656 *) DTLS Handshake overhaul. There were longstanding issues with
1657 OpenSSL DTLS implementation, which were making it impossible for
1658 RFC 4347 compliant client to communicate with OpenSSL server.
1659 Unfortunately just fixing these incompatibilities would "cut off"
1660 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
1661 server keeps tolerating non RFC compliant syntax. The opposite is
1662 not true, 0.9.8f client can not communicate with earlier server.
1663 This update even addresses CVE-2007-4995.
1666 *) Changes to avoid need for function casts in OpenSSL: some compilers
1667 (gcc 4.2 and later) reject their use.
1668 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
1671 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1672 RFC4507bis. The encrypted ticket format is an encrypted encoded
1673 SSL_SESSION structure, that way new session features are automatically
1676 If a client application caches session in an SSL_SESSION structure
1677 support is transparent because tickets are now stored in the encoded
1680 The SSL_CTX structure automatically generates keys for ticket
1681 protection in servers so again support should be possible
1682 with no application modification.
1684 If a client or server wishes to disable RFC4507 support then the option
1685 SSL_OP_NO_TICKET can be set.
1687 Add a TLS extension debugging callback to allow the contents of any client
1688 or server extensions to be examined.
1690 This work was sponsored by Google.
1693 *) Add initial support for TLS extensions, specifically for the server_name
1694 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1695 have new members for a host name. The SSL data structure has an
1696 additional member SSL_CTX *initial_ctx so that new sessions can be
1697 stored in that context to allow for session resumption, even after the
1698 SSL has been switched to a new SSL_CTX in reaction to a client's
1699 server_name extension.
1701 New functions (subject to change):
1703 SSL_get_servername()
1704 SSL_get_servername_type()
1707 New CTRL codes and macros (subject to change):
1709 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1710 - SSL_CTX_set_tlsext_servername_callback()
1711 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1712 - SSL_CTX_set_tlsext_servername_arg()
1713 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1715 openssl s_client has a new '-servername ...' option.
1717 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1718 '-key2 ...', '-servername_fatal' (subject to change). This allows
1719 testing the HostName extension for a specific single host name ('-cert'
1720 and '-key' remain fallbacks for handshakes without HostName
1721 negotiation). If the unrecogninzed_name alert has to be sent, this by
1722 default is a warning; it becomes fatal with the '-servername_fatal'
1725 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
1727 *) Add AES and SSE2 assembly language support to VC++ build.
1730 *) Mitigate attack on final subtraction in Montgomery reduction.
1733 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
1734 (which previously caused an internal error).
1737 *) Squeeze another 10% out of IGE mode when in != out.
1740 *) AES IGE mode speedup.
1741 [Dean Gaudet (Google)]
1743 *) Add the Korean symmetric 128-bit cipher SEED (see
1744 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
1745 add SEED ciphersuites from RFC 4162:
1747 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
1748 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
1749 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
1750 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
1752 To minimize changes between patchlevels in the OpenSSL 0.9.8
1753 series, SEED remains excluded from compilation unless OpenSSL
1754 is configured with 'enable-seed'.
1755 [KISA, Bodo Moeller]
1757 *) Mitigate branch prediction attacks, which can be practical if a
1758 single processor is shared, allowing a spy process to extract
1759 information. For detailed background information, see
1760 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
1761 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
1762 and Necessary Software Countermeasures"). The core of the change
1763 are new versions BN_div_no_branch() and
1764 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
1765 respectively, which are slower, but avoid the security-relevant
1766 conditional branches. These are automatically called by BN_div()
1767 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
1768 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
1769 remove a conditional branch.
1771 BN_FLG_CONSTTIME is the new name for the previous
1772 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
1773 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
1774 in the exponent causes BN_mod_exp_mont() to use the alternative
1775 implementation in BN_mod_exp_mont_consttime().) The old name
1776 remains as a deprecated alias.
1778 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
1779 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
1780 constant-time implementations for more than just exponentiation.
1781 Here too the old name is kept as a deprecated alias.
1783 BN_BLINDING_new() will now use BN_dup() for the modulus so that
1784 the BN_BLINDING structure gets an independent copy of the
1785 modulus. This means that the previous "BIGNUM *m" argument to
1786 BN_BLINDING_new() and to BN_BLINDING_create_param() now
1787 essentially becomes "const BIGNUM *m", although we can't actually
1788 change this in the header file before 0.9.9. It allows
1789 RSA_setup_blinding() to use BN_with_flags() on the modulus to
1790 enable BN_FLG_CONSTTIME.
1792 [Matthew D Wood (Intel Corp)]
1794 *) In the SSL/TLS server implementation, be strict about session ID
1795 context matching (which matters if an application uses a single
1796 external cache for different purposes). Previously,
1797 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
1798 set. This did ensure strict client verification, but meant that,
1799 with applications using a single external cache for quite
1800 different requirements, clients could circumvent ciphersuite
1801 restrictions for a given session ID context by starting a session
1802 in a different context.
1805 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
1806 a ciphersuite string such as "DEFAULT:RSA" cannot enable
1807 authentication-only ciphersuites.
1810 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
1811 not complete and could lead to a possible single byte overflow
1812 (CVE-2007-5135) [Ben Laurie]
1814 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
1816 *) Since AES128 and AES256 (and similarly Camellia128 and
1817 Camellia256) share a single mask bit in the logic of
1818 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
1819 kludge to work properly if AES128 is available and AES256 isn't
1820 (or if Camellia128 is available and Camellia256 isn't).
1823 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
1824 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
1825 When a point or a seed is encoded in a BIT STRING, we need to
1826 prevent the removal of trailing zero bits to get the proper DER
1827 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
1828 of a NamedBitList, for which trailing 0 bits need to be removed.)
1831 *) Have SSL/TLS server implementation tolerate "mismatched" record
1832 protocol version while receiving ClientHello even if the
1833 ClientHello is fragmented. (The server can't insist on the
1834 particular protocol version it has chosen before the ServerHello
1835 message has informed the client about his choice.)
1838 *) Add RFC 3779 support.
1839 [Rob Austein for ARIN, Ben Laurie]
1841 *) Load error codes if they are not already present instead of using a
1842 static variable. This allows them to be cleanly unloaded and reloaded.
1843 Improve header file function name parsing.
1846 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
1847 or CAPABILITY handshake as required by RFCs.
1850 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
1852 *) Introduce limits to prevent malicious keys being able to
1853 cause a denial of service. (CVE-2006-2940)
1854 [Steve Henson, Bodo Moeller]
1856 *) Fix ASN.1 parsing of certain invalid structures that can result
1857 in a denial of service. (CVE-2006-2937) [Steve Henson]
1859 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
1860 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
1862 *) Fix SSL client code which could crash if connecting to a
1863 malicious SSLv2 server. (CVE-2006-4343)
1864 [Tavis Ormandy and Will Drewry, Google Security Team]
1866 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
1867 match only those. Before that, "AES256-SHA" would be interpreted
1868 as a pattern and match "AES128-SHA" too (since AES128-SHA got
1869 the same strength classification in 0.9.7h) as we currently only
1870 have a single AES bit in the ciphersuite description bitmap.
1871 That change, however, also applied to ciphersuite strings such as
1872 "RC4-MD5" that intentionally matched multiple ciphersuites --
1873 namely, SSL 2.0 ciphersuites in addition to the more common ones
1874 from SSL 3.0/TLS 1.0.
1876 So we change the selection algorithm again: Naming an explicit
1877 ciphersuite selects this one ciphersuite, and any other similar
1878 ciphersuite (same bitmap) from *other* protocol versions.
1879 Thus, "RC4-MD5" again will properly select both the SSL 2.0
1880 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
1882 Since SSL 2.0 does not have any ciphersuites for which the
1883 128/256 bit distinction would be relevant, this works for now.
1884 The proper fix will be to use different bits for AES128 and
1885 AES256, which would have avoided the problems from the beginning;
1886 however, bits are scarce, so we can only do this in a new release
1887 (not just a patchlevel) when we can change the SSL_CIPHER
1888 definition to split the single 'unsigned long mask' bitmap into
1889 multiple values to extend the available space.
1893 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
1895 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
1896 (CVE-2006-4339) [Ben Laurie and Google Security Team]
1898 *) Add AES IGE and biIGE modes.
1901 *) Change the Unix randomness entropy gathering to use poll() when
1902 possible instead of select(), since the latter has some
1903 undesirable limitations.
1904 [Darryl Miles via Richard Levitte and Bodo Moeller]
1906 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
1907 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
1908 cannot be implicitly activated as part of, e.g., the "AES" alias.
1909 However, please upgrade to OpenSSL 0.9.9[-dev] for
1910 non-experimental use of the ECC ciphersuites to get TLS extension
1911 support, which is required for curve and point format negotiation
1912 to avoid potential handshake problems.
1915 *) Disable rogue ciphersuites:
1917 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
1918 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
1919 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
1921 The latter two were purportedly from
1922 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
1925 Also deactivate the remaining ciphersuites from
1926 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
1927 unofficial, and the ID has long expired.
1930 *) Fix RSA blinding Heisenbug (problems sometimes occured on
1931 dual-core machines) and other potential thread-safety issues.
1934 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
1935 versions), which is now available for royalty-free use
1936 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
1937 Also, add Camellia TLS ciphersuites from RFC 4132.
1939 To minimize changes between patchlevels in the OpenSSL 0.9.8
1940 series, Camellia remains excluded from compilation unless OpenSSL
1941 is configured with 'enable-camellia'.
1944 *) Disable the padding bug check when compression is in use. The padding
1945 bug check assumes the first packet is of even length, this is not
1946 necessarily true if compresssion is enabled and can result in false
1947 positives causing handshake failure. The actual bug test is ancient
1948 code so it is hoped that implementations will either have fixed it by
1949 now or any which still have the bug do not support compression.
1952 Changes between 0.9.8a and 0.9.8b [04 May 2006]
1954 *) When applying a cipher rule check to see if string match is an explicit
1955 cipher suite and only match that one cipher suite if it is.
1958 *) Link in manifests for VC++ if needed.
1959 [Austin Ziegler <halostatue@gmail.com>]
1961 *) Update support for ECC-based TLS ciphersuites according to
1962 draft-ietf-tls-ecc-12.txt with proposed changes (but without
1963 TLS extensions, which are supported starting with the 0.9.9
1964 branch, not in the OpenSSL 0.9.8 branch).
1967 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
1968 opaque EVP_CIPHER_CTX handling.
1971 *) Fixes and enhancements to zlib compression code. We now only use
1972 "zlib1.dll" and use the default __cdecl calling convention on Win32
1973 to conform with the standards mentioned here:
1974 http://www.zlib.net/DLL_FAQ.txt
1975 Static zlib linking now works on Windows and the new --with-zlib-include
1976 --with-zlib-lib options to Configure can be used to supply the location
1977 of the headers and library. Gracefully handle case where zlib library
1981 *) Several fixes and enhancements to the OID generation code. The old code
1982 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
1983 handle numbers larger than ULONG_MAX, truncated printing and had a
1984 non standard OBJ_obj2txt() behaviour.
1987 *) Add support for building of engines under engine/ as shared libraries
1988 under VC++ build system.
1991 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
1992 Hopefully, we will not see any false combination of paths any more.
1995 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
1997 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1998 (part of SSL_OP_ALL). This option used to disable the
1999 countermeasure against man-in-the-middle protocol-version
2000 rollback in the SSL 2.0 server implementation, which is a bad
2001 idea. (CVE-2005-2969)
2003 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2004 for Information Security, National Institute of Advanced Industrial
2005 Science and Technology [AIST], Japan)]
2007 *) Add two function to clear and return the verify parameter flags.
2010 *) Keep cipherlists sorted in the source instead of sorting them at
2011 runtime, thus removing the need for a lock.
2014 *) Avoid some small subgroup attacks in Diffie-Hellman.
2015 [Nick Mathewson and Ben Laurie]
2017 *) Add functions for well-known primes.
2020 *) Extended Windows CE support.
2021 [Satoshi Nakamura and Andy Polyakov]
2023 *) Initialize SSL_METHOD structures at compile time instead of during
2024 runtime, thus removing the need for a lock.
2027 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2028 attempting to decrypt each encrypted key in turn. Add support to
2032 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2034 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2037 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2040 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2041 key into the same file any more.
2044 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2047 *) Add -utf8 command line and config file option to 'ca'.
2048 [Stefan <stf@udoma.org]
2050 *) Removed the macro des_crypt(), as it seems to conflict with some
2051 libraries. Use DES_crypt().
2054 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2055 involves renaming the source and generated shared-libs for
2056 both. The engines will accept the corrected or legacy ids
2057 ('ncipher' and '4758_cca' respectively) when binding. NB,
2058 this only applies when building 'shared'.
2059 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2061 *) Add attribute functions to EVP_PKEY structure. Modify
2062 PKCS12_create() to recognize a CSP name attribute and
2063 use it. Make -CSP option work again in pkcs12 utility.
2066 *) Add new functionality to the bn blinding code:
2067 - automatic re-creation of the BN_BLINDING parameters after
2068 a fixed number of uses (currently 32)
2069 - add new function for parameter creation
2070 - introduce flags to control the update behaviour of the
2071 BN_BLINDING parameters
2072 - hide BN_BLINDING structure
2073 Add a second BN_BLINDING slot to the RSA structure to improve
2074 performance when a single RSA object is shared among several
2078 *) Add support for DTLS.
2079 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2081 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2082 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2085 *) Remove buggy and incompletet DH cert support from
2086 ssl/ssl_rsa.c and ssl/s3_both.c
2089 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2090 the apps/openssl applications.
2093 *) Compile clean with "-Wall -Wmissing-prototypes
2094 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2095 DEBUG_SAFESTACK must also be set.
2098 *) Change ./Configure so that certain algorithms can be disabled by default.
2099 The new counterpiece to "no-xxx" is "enable-xxx".
2101 The patented RC5 and MDC2 algorithms will now be disabled unless
2102 "enable-rc5" and "enable-mdc2", respectively, are specified.
2104 (IDEA remains enabled despite being patented. This is because IDEA
2105 is frequently required for interoperability, and there is no license
2106 fee for non-commercial use. As before, "no-idea" can be used to
2107 avoid this algorithm.)
2111 *) Add processing of proxy certificates (see RFC 3820). This work was
2112 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2113 EGEE (Enabling Grids for E-science in Europe).
2116 *) RC4 performance overhaul on modern architectures/implementations, such
2117 as Intel P4, IA-64 and AMD64.
2120 *) New utility extract-section.pl. This can be used specify an alternative
2121 section number in a pod file instead of having to treat each file as
2122 a separate case in Makefile. This can be done by adding two lines to the
2125 =for comment openssl_section:XXX
2127 The blank line is mandatory.
2131 *) New arguments -certform, -keyform and -pass for s_client and s_server
2132 to allow alternative format key and certificate files and passphrase
2136 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2137 update associated structures and add various utility functions.
2139 Add new policy related verify parameters, include policy checking in
2140 standard verify code. Enhance 'smime' application with extra parameters
2141 to support policy checking and print out.
2144 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2145 Nehemiah processors. These extensions support AES encryption in hardware
2146 as well as RNG (though RNG support is currently disabled).
2147 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2149 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2152 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2153 [Andy Polyakov and a number of other people]
2155 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2156 implementation contributed by IBM.
2157 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2159 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2160 exponent rather than 'unsigned long'. There is a corresponding change to
2161 the new 'rsa_keygen' element of the RSA_METHOD structure.
2162 [Jelte Jansen, Geoff Thorpe]
2164 *) Functionality for creating the initial serial number file is now
2165 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2167 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2168 number file to 1, which is bound to cause problems. To avoid
2169 the problems while respecting compatibility between different 0.9.7
2170 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2171 CA.pl for serial number initialization. With the new release 0.9.8,
2172 we can fix the problem directly in the 'ca' utility.)
2175 *) Reduced header interdepencies by declaring more opaque objects in
2176 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2177 give fewer recursive includes, which could break lazy source code - so
2178 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2179 developers should define this symbol when building and using openssl to
2180 ensure they track the recommended behaviour, interfaces, [etc], but
2181 backwards-compatible behaviour prevails when this isn't defined.
2184 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2187 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2188 This will generate a random key of the appropriate length based on the
2189 cipher context. The EVP_CIPHER can provide its own random key generation
2190 routine to support keys of a specific form. This is used in the des and
2191 3des routines to generate a key of the correct parity. Update S/MIME
2192 code to use new functions and hence generate correct parity DES keys.
2193 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2194 valid (weak or incorrect parity).
2197 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2198 as looking them up. This is useful when the verified structure may contain
2199 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2200 present unless the new PKCS7_NO_CRL flag is asserted.
2203 *) Extend ASN1 oid configuration module. It now additionally accepts the
2206 shortName = some long name, 1.2.3.4
2209 *) Reimplemented the BN_CTX implementation. There is now no more static
2210 limitation on the number of variables it can handle nor the depth of the
2211 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2212 information can now expand as required, and rather than having a single
2213 static array of bignums, BN_CTX now uses a linked-list of such arrays
2214 allowing it to expand on demand whilst maintaining the usefulness of
2215 BN_CTX's "bundling".
2218 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2219 to allow all RSA operations to function using a single BN_CTX.
2222 *) Preliminary support for certificate policy evaluation and checking. This
2223 is initially intended to pass the tests outlined in "Conformance Testing
2224 of Relying Party Client Certificate Path Processing Logic" v1.07.
2227 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2228 remained unused and not that useful. A variety of other little bignum
2229 tweaks and fixes have also been made continuing on from the audit (see
2233 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2234 associated ASN1, EVP and SSL functions and old ASN1 macros.
2237 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2238 and this should never fail. So the return value from the use of
2239 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2240 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2243 *) BN_CTX_get() should return zero-valued bignums, providing the same
2244 initialised value as BN_new().
2245 [Geoff Thorpe, suggested by Ulf Möller]
2247 *) Support for inhibitAnyPolicy certificate extension.
2250 *) An audit of the BIGNUM code is underway, for which debugging code is
2251 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
2252 is considered valid when processing BIGNUMs, and causes execution to
2253 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
2254 further steps are taken to deliberately pollute unused data in BIGNUM
2255 structures to try and expose faulty code further on. For now, openssl will
2256 (in its default mode of operation) continue to tolerate the inconsistent
2257 forms that it has tolerated in the past, but authors and packagers should
2258 consider trying openssl and their own applications when compiled with
2259 these debugging symbols defined. It will help highlight potential bugs in
2260 their own code, and will improve the test coverage for OpenSSL itself. At
2261 some point, these tighter rules will become openssl's default to improve
2262 maintainability, though the assert()s and other overheads will remain only
2263 in debugging configurations. See bn.h for more details.
2264 [Geoff Thorpe, Nils Larsch, Ulf Möller]
2266 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
2267 that can only be obtained through BN_CTX_new() (which implicitly
2268 initialises it). The presence of this function only made it possible
2269 to overwrite an existing structure (and cause memory leaks).
2272 *) Because of the callback-based approach for implementing LHASH as a
2273 template type, lh_insert() adds opaque objects to hash-tables and
2274 lh_doall() or lh_doall_arg() are typically used with a destructor callback
2275 to clean up those corresponding objects before destroying the hash table
2276 (and losing the object pointers). So some over-zealous constifications in
2277 LHASH have been relaxed so that lh_insert() does not take (nor store) the
2278 objects as "const" and the lh_doall[_arg] callback wrappers are not
2279 prototyped to have "const" restrictions on the object pointers they are
2280 given (and so aren't required to cast them away any more).
2283 *) The tmdiff.h API was so ugly and minimal that our own timing utility
2284 (speed) prefers to use its own implementation. The two implementations
2285 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
2286 its object type properly exposed (MS_TM) instead of casting to/from "char
2287 *". This may still change yet if someone realises MS_TM and "ms_time_***"
2288 aren't necessarily the greatest nomenclatures - but this is what was used
2289 internally to the implementation so I've used that for now.
2292 *) Ensure that deprecated functions do not get compiled when
2293 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
2294 the self-tests were still using deprecated key-generation functions so
2295 these have been updated also.
2298 *) Reorganise PKCS#7 code to separate the digest location functionality
2299 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
2300 New function PKCS7_set_digest() to set the digest type for PKCS#7
2301 digestedData type. Add additional code to correctly generate the
2302 digestedData type and add support for this type in PKCS7 initialization
2306 *) New function PKCS7_set0_type_other() this initializes a PKCS7
2307 structure of type "other".
2310 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
2311 sure the loop does correctly stop and breaking ("division by zero")
2312 modulus operations are not performed. The (pre-generated) prime
2313 table crypto/bn/bn_prime.h was already correct, but it could not be
2314 re-generated on some platforms because of the "division by zero"
2315 situation in the script.
2316 [Ralf S. Engelschall]
2318 *) Update support for ECC-based TLS ciphersuites according to
2319 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
2320 SHA-1 now is only used for "small" curves (where the
2321 representation of a field element takes up to 24 bytes); for
2322 larger curves, the field element resulting from ECDH is directly
2323 used as premaster secret.
2324 [Douglas Stebila (Sun Microsystems Laboratories)]
2326 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
2327 curve secp160r1 to the tests.
2328 [Douglas Stebila (Sun Microsystems Laboratories)]
2330 *) Add the possibility to load symbols globally with DSO.
2331 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
2333 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
2334 control of the error stack.
2337 *) Add support for STORE in ENGINE.
2340 *) Add the STORE type. The intention is to provide a common interface
2341 to certificate and key stores, be they simple file-based stores, or
2342 HSM-type store, or LDAP stores, or...
2343 NOTE: The code is currently UNTESTED and isn't really used anywhere.
2346 *) Add a generic structure called OPENSSL_ITEM. This can be used to
2347 pass a list of arguments to any function as well as provide a way
2348 for a function to pass data back to the caller.
2351 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
2352 works like BUF_strdup() but can be used to duplicate a portion of
2353 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
2357 *) Add the function sk_find_ex() which works like sk_find(), but will
2358 return an index to an element even if an exact match couldn't be
2359 found. The index is guaranteed to point at the element where the
2360 searched-for key would be inserted to preserve sorting order.
2363 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
2364 takes an extra flags argument for optional functionality. Currently,
2365 the following flags are defined:
2367 OBJ_BSEARCH_VALUE_ON_NOMATCH
2368 This one gets OBJ_bsearch_ex() to return a pointer to the first
2369 element where the comparing function returns a negative or zero
2372 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
2373 This one gets OBJ_bsearch_ex() to return a pointer to the first
2374 element where the comparing function returns zero. This is useful
2375 if there are more than one element where the comparing function
2379 *) Make it possible to create self-signed certificates with 'openssl ca'
2380 in such a way that the self-signed certificate becomes part of the
2381 CA database and uses the same mechanisms for serial number generation
2382 as all other certificate signing. The new flag '-selfsign' enables
2383 this functionality. Adapt CA.sh and CA.pl.in.
2386 *) Add functionality to check the public key of a certificate request
2387 against a given private. This is useful to check that a certificate
2388 request can be signed by that key (self-signing).
2391 *) Make it possible to have multiple active certificates with the same
2392 subject in the CA index file. This is done only if the keyword
2393 'unique_subject' is set to 'no' in the main CA section (default
2394 if 'CA_default') of the configuration file. The value is saved
2395 with the database itself in a separate index attribute file,
2396 named like the index file with '.attr' appended to the name.
2399 *) Generate muti valued AVAs using '+' notation in config files for
2403 *) Support for nameConstraints certificate extension.
2406 *) Support for policyConstraints certificate extension.
2409 *) Support for policyMappings certificate extension.
2412 *) Make sure the default DSA_METHOD implementation only uses its
2413 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
2414 and change its own handlers to be NULL so as to remove unnecessary
2415 indirection. This lets alternative implementations fallback to the
2416 default implementation more easily.
2419 *) Support for directoryName in GeneralName related extensions
2423 *) Make it possible to link applications using Makefile.shared.
2424 Make that possible even when linking against static libraries!
2427 *) Support for single pass processing for S/MIME signing. This now
2428 means that S/MIME signing can be done from a pipe, in addition
2429 cleartext signing (multipart/signed type) is effectively streaming
2430 and the signed data does not need to be all held in memory.
2432 This is done with a new flag PKCS7_STREAM. When this flag is set
2433 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
2434 is done after the data is output (and digests calculated) in
2435 SMIME_write_PKCS7().
2438 *) Add full support for -rpath/-R, both in shared libraries and
2439 applications, at least on the platforms where it's known how
2443 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
2444 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
2445 will now compute a table of multiples of the generator that
2446 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
2447 faster (notably in the case of a single point multiplication,
2448 scalar * generator).
2449 [Nils Larsch, Bodo Moeller]
2451 *) IPv6 support for certificate extensions. The various extensions
2452 which use the IP:a.b.c.d can now take IPv6 addresses using the
2453 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
2457 *) Added an ENGINE that implements RSA by performing private key
2458 exponentiations with the GMP library. The conversions to and from
2459 GMP's mpz_t format aren't optimised nor are any montgomery forms
2460 cached, and on x86 it appears OpenSSL's own performance has caught up.
2461 However there are likely to be other architectures where GMP could
2462 provide a boost. This ENGINE is not built in by default, but it can be
2463 specified at Configure time and should be accompanied by the necessary
2464 linker additions, eg;
2465 ./config -DOPENSSL_USE_GMP -lgmp
2468 *) "openssl engine" will not display ENGINE/DSO load failure errors when
2469 testing availability of engines with "-t" - the old behaviour is
2470 produced by increasing the feature's verbosity with "-tt".
2473 *) ECDSA routines: under certain error conditions uninitialized BN objects
2474 could be freed. Solution: make sure initialization is performed early
2475 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
2479 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
2480 and DH_METHOD (eg. by ENGINE implementations) to override the normal
2481 software implementations. For DSA and DH, parameter generation can
2482 also be overriden by providing the appropriate method callbacks.
2485 *) Change the "progress" mechanism used in key-generation and
2486 primality testing to functions that take a new BN_GENCB pointer in
2487 place of callback/argument pairs. The new API functions have "_ex"
2488 postfixes and the older functions are reimplemented as wrappers for
2489 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
2490 declarations of the old functions to help (graceful) attempts to
2491 migrate to the new functions. Also, the new key-generation API
2492 functions operate on a caller-supplied key-structure and return
2493 success/failure rather than returning a key or NULL - this is to
2494 help make "keygen" another member function of RSA_METHOD etc.
2496 Example for using the new callback interface:
2498 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
2502 BN_GENCB_set(&my_cb, my_callback, my_arg);
2504 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
2505 /* For the meaning of a, b in calls to my_callback(), see the
2506 * documentation of the function that calls the callback.
2507 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
2508 * my_callback should return 1 if it wants BN_is_prime_ex()
2509 * to continue, or 0 to stop.
2514 *) Change the ZLIB compression method to be stateful, and make it
2515 available to TLS with the number defined in
2516 draft-ietf-tls-compression-04.txt.
2519 *) Add the ASN.1 structures and functions for CertificatePair, which
2520 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
2522 CertificatePair ::= SEQUENCE {
2523 forward [0] Certificate OPTIONAL,
2524 reverse [1] Certificate OPTIONAL,
2525 -- at least one of the pair shall be present -- }
2527 Also implement the PEM functions to read and write certificate
2528 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
2530 This needed to be defined, mostly for the sake of the LDAP
2531 attribute crossCertificatePair, but may prove useful elsewhere as
2535 *) Make it possible to inhibit symlinking of shared libraries in
2536 Makefile.shared, for Cygwin's sake.
2539 *) Extend the BIGNUM API by creating a function
2540 void BN_set_negative(BIGNUM *a, int neg);
2541 and a macro that behave like
2542 int BN_is_negative(const BIGNUM *a);
2544 to avoid the need to access 'a->neg' directly in applications.
2547 *) Implement fast modular reduction for pseudo-Mersenne primes
2548 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
2549 EC_GROUP_new_curve_GFp() will now automatically use this
2551 [Nils Larsch <nla@trustcenter.de>]
2553 *) Add new lock type (CRYPTO_LOCK_BN).
2556 *) Change the ENGINE framework to automatically load engines
2557 dynamically from specific directories unless they could be
2558 found to already be built in or loaded. Move all the
2559 current engines except for the cryptodev one to a new
2561 The engines in engines/ are built as shared libraries if
2562 the "shared" options was given to ./Configure or ./config.
2563 Otherwise, they are inserted in libcrypto.a.
2564 /usr/local/ssl/engines is the default directory for dynamic
2565 engines, but that can be overriden at configure time through
2566 the usual use of --prefix and/or --openssldir, and at run
2567 time with the environment variable OPENSSL_ENGINES.
2568 [Geoff Thorpe and Richard Levitte]
2570 *) Add Makefile.shared, a helper makefile to build shared
2571 libraries. Addapt Makefile.org.
2574 *) Add version info to Win32 DLLs.
2575 [Peter 'Luna' Runestig" <peter@runestig.com>]
2577 *) Add new 'medium level' PKCS#12 API. Certificates and keys
2578 can be added using this API to created arbitrary PKCS#12
2579 files while avoiding the low level API.
2581 New options to PKCS12_create(), key or cert can be NULL and
2582 will then be omitted from the output file. The encryption
2583 algorithm NIDs can be set to -1 for no encryption, the mac
2584 iteration count can be set to 0 to omit the mac.
2586 Enhance pkcs12 utility by making the -nokeys and -nocerts
2587 options work when creating a PKCS#12 file. New option -nomac
2588 to omit the mac, NONE can be set for an encryption algorithm.
2589 New code is modified to use the enhanced PKCS12_create()
2590 instead of the low level API.
2593 *) Extend ASN1 encoder to support indefinite length constructed
2594 encoding. This can output sequences tags and octet strings in
2595 this form. Modify pk7_asn1.c to support indefinite length
2596 encoding. This is experimental and needs additional code to
2597 be useful, such as an ASN1 bio and some enhanced streaming
2600 Extend template encode functionality so that tagging is passed
2601 down to the template encoder.
2604 *) Let 'openssl req' fail if an argument to '-newkey' is not
2605 recognized instead of using RSA as a default.
2608 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
2609 As these are not official, they are not included in "ALL";
2610 the "ECCdraft" ciphersuite group alias can be used to select them.
2611 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
2613 *) Add ECDH engine support.
2614 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
2616 *) Add ECDH in new directory crypto/ecdh/.
2617 [Douglas Stebila (Sun Microsystems Laboratories)]
2619 *) Let BN_rand_range() abort with an error after 100 iterations
2620 without success (which indicates a broken PRNG).
2623 *) Change BN_mod_sqrt() so that it verifies that the input value
2624 is really the square of the return value. (Previously,
2625 BN_mod_sqrt would show GIGO behaviour.)
2628 *) Add named elliptic curves over binary fields from X9.62, SECG,
2629 and WAP/WTLS; add OIDs that were still missing.
2631 [Sheueling Chang Shantz and Douglas Stebila
2632 (Sun Microsystems Laboratories)]
2634 *) Extend the EC library for elliptic curves over binary fields
2635 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
2638 EC_GF2m_simple_method
2642 EC_GROUP_new_curve_GF2m
2643 EC_GROUP_set_curve_GF2m
2644 EC_GROUP_get_curve_GF2m
2645 EC_POINT_set_affine_coordinates_GF2m
2646 EC_POINT_get_affine_coordinates_GF2m
2647 EC_POINT_set_compressed_coordinates_GF2m
2649 Point compression for binary fields is disabled by default for
2650 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
2653 As binary polynomials are represented as BIGNUMs, various members
2654 of the EC_GROUP and EC_POINT data structures can be shared
2655 between the implementations for prime fields and binary fields;
2656 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
2657 are essentially identical to their ..._GFp counterparts.
2658 (For simplicity, the '..._GFp' prefix has been dropped from
2659 various internal method names.)
2661 An internal 'field_div' method (similar to 'field_mul' and
2662 'field_sqr') has been added; this is used only for binary fields.
2664 [Sheueling Chang Shantz and Douglas Stebila
2665 (Sun Microsystems Laboratories)]
2667 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
2668 through methods ('mul', 'precompute_mult').
2670 The generic implementations (now internally called 'ec_wNAF_mul'
2671 and 'ec_wNAF_precomputed_mult') remain the default if these
2672 methods are undefined.
2674 [Sheueling Chang Shantz and Douglas Stebila
2675 (Sun Microsystems Laboratories)]
2677 *) New function EC_GROUP_get_degree, which is defined through
2678 EC_METHOD. For curves over prime fields, this returns the bit
2679 length of the modulus.
2681 [Sheueling Chang Shantz and Douglas Stebila
2682 (Sun Microsystems Laboratories)]
2684 *) New functions EC_GROUP_dup, EC_POINT_dup.
2685 (These simply call ..._new and ..._copy).
2687 [Sheueling Chang Shantz and Douglas Stebila
2688 (Sun Microsystems Laboratories)]
2690 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
2691 Polynomials are represented as BIGNUMs (where the sign bit is not
2692 used) in the following functions [macros]:
2695 BN_GF2m_sub [= BN_GF2m_add]
2696 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
2697 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
2698 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
2700 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
2701 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
2702 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
2703 BN_GF2m_cmp [= BN_ucmp]
2705 (Note that only the 'mod' functions are actually for fields GF(2^m).
2706 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
2708 For some functions, an the irreducible polynomial defining a
2709 field can be given as an 'unsigned int[]' with strictly
2710 decreasing elements giving the indices of those bits that are set;
2711 i.e., p[] represents the polynomial
2712 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
2714 p[0] > p[1] > ... > p[k] = 0.
2715 This applies to the following functions:
2720 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
2721 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
2723 BN_GF2m_mod_sqrt_arr
2724 BN_GF2m_mod_solve_quad_arr
2728 Conversion can be performed by the following functions:
2733 bntest.c has additional tests for binary polynomial arithmetic.
2735 Two implementations for BN_GF2m_mod_div() are available.
2736 The default algorithm simply uses BN_GF2m_mod_inv() and
2737 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
2738 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
2739 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
2741 [Sheueling Chang Shantz and Douglas Stebila
2742 (Sun Microsystems Laboratories)]
2744 *) Add new error code 'ERR_R_DISABLED' that can be used when some
2745 functionality is disabled at compile-time.
2746 [Douglas Stebila <douglas.stebila@sun.com>]
2748 *) Change default behaviour of 'openssl asn1parse' so that more
2749 information is visible when viewing, e.g., a certificate:
2751 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
2752 mode the content of non-printable OCTET STRINGs is output in a
2753 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
2754 avoid the appearance of a printable string.
2755 [Nils Larsch <nla@trustcenter.de>]
2757 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
2759 EC_GROUP_set_asn1_flag()
2760 EC_GROUP_get_asn1_flag()
2761 EC_GROUP_set_point_conversion_form()
2762 EC_GROUP_get_point_conversion_form()
2763 These control ASN1 encoding details:
2764 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
2765 has been set to OPENSSL_EC_NAMED_CURVE.
2766 - Points are encoded in uncompressed form by default; options for
2767 asn1_for are as for point2oct, namely
2768 POINT_CONVERSION_COMPRESSED
2769 POINT_CONVERSION_UNCOMPRESSED
2770 POINT_CONVERSION_HYBRID
2772 Also add 'seed' and 'seed_len' members to EC_GROUP with access
2775 EC_GROUP_get0_seed()
2776 EC_GROUP_get_seed_len()
2777 This is used only for ASN1 purposes (so far).
2778 [Nils Larsch <nla@trustcenter.de>]
2780 *) Add 'field_type' member to EC_METHOD, which holds the NID
2781 of the appropriate field type OID. The new function
2782 EC_METHOD_get_field_type() returns this value.
2783 [Nils Larsch <nla@trustcenter.de>]
2788 EC_POINT_point2hex()
2789 EC_POINT_hex2point()
2790 providing useful interfaces to EC_POINT_point2oct() and
2791 EC_POINT_oct2point().
2792 [Nils Larsch <nla@trustcenter.de>]
2794 *) Change internals of the EC library so that the functions
2795 EC_GROUP_set_generator()
2796 EC_GROUP_get_generator()
2797 EC_GROUP_get_order()
2798 EC_GROUP_get_cofactor()
2799 are implemented directly in crypto/ec/ec_lib.c and not dispatched
2800 to methods, which would lead to unnecessary code duplication when
2801 adding different types of curves.
2802 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
2804 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
2805 arithmetic, and such that modified wNAFs are generated
2806 (which avoid length expansion in many cases).
2809 *) Add a function EC_GROUP_check_discriminant() (defined via
2810 EC_METHOD) that verifies that the curve discriminant is non-zero.
2812 Add a function EC_GROUP_check() that makes some sanity tests
2813 on a EC_GROUP, its generator and order. This includes
2814 EC_GROUP_check_discriminant().
2815 [Nils Larsch <nla@trustcenter.de>]
2817 *) Add ECDSA in new directory crypto/ecdsa/.
2819 Add applications 'openssl ecparam' and 'openssl ecdsa'
2820 (these are based on 'openssl dsaparam' and 'openssl dsa').
2822 ECDSA support is also included in various other files across the
2823 library. Most notably,
2824 - 'openssl req' now has a '-newkey ecdsa:file' option;
2825 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
2826 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
2827 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
2828 them suitable for ECDSA where domain parameters must be
2829 extracted before the specific public key;
2830 - ECDSA engine support has been added.
2831 [Nils Larsch <nla@trustcenter.de>]
2833 *) Include some named elliptic curves, and add OIDs from X9.62,
2834 SECG, and WAP/WTLS. Each curve can be obtained from the new
2836 EC_GROUP_new_by_curve_name(),
2837 and the list of available named curves can be obtained with
2838 EC_get_builtin_curves().
2839 Also add a 'curve_name' member to EC_GROUP objects, which can be
2841 EC_GROUP_set_curve_name()
2842 EC_GROUP_get_curve_name()
2843 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
2845 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2846 was actually never needed) and in BN_mul(). The removal in BN_mul()
2847 required a small change in bn_mul_part_recursive() and the addition
2848 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2849 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2850 bn_sub_words() and bn_add_words() except they take arrays with
2854 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
2856 *) Cleanse PEM buffers before freeing them since they may contain
2858 [Benjamin Bennett <ben@psc.edu>]
2860 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2861 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2862 authentication-only ciphersuites.
2865 *) Since AES128 and AES256 share a single mask bit in the logic of
2866 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2867 kludge to work properly if AES128 is available and AES256 isn't.
2870 *) Expand security boundary to match 1.1.1 module.
2873 *) Remove redundant features: hash file source, editing of test vectors
2874 modify fipsld to use external fips_premain.c signature.
2877 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
2878 run algorithm test programs.
2881 *) Make algorithm test programs more tolerant of whitespace.
2884 *) Have SSL/TLS server implementation tolerate "mismatched" record
2885 protocol version while receiving ClientHello even if the
2886 ClientHello is fragmented. (The server can't insist on the
2887 particular protocol version it has chosen before the ServerHello
2888 message has informed the client about his choice.)
2891 *) Load error codes if they are not already present instead of using a
2892 static variable. This allows them to be cleanly unloaded and reloaded.
2895 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
2897 *) Introduce limits to prevent malicious keys being able to
2898 cause a denial of service. (CVE-2006-2940)
2899 [Steve Henson, Bodo Moeller]
2901 *) Fix ASN.1 parsing of certain invalid structures that can result
2902 in a denial of service. (CVE-2006-2937) [Steve Henson]
2904 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2905 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2907 *) Fix SSL client code which could crash if connecting to a
2908 malicious SSLv2 server. (CVE-2006-4343)
2909 [Tavis Ormandy and Will Drewry, Google Security Team]
2911 *) Change ciphersuite string processing so that an explicit
2912 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
2913 will no longer include "AES128-SHA"), and any other similar
2914 ciphersuite (same bitmap) from *other* protocol versions (so that
2915 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
2916 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
2917 changes from 0.9.8b and 0.9.8d.
2920 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
2922 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2923 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2925 *) Change the Unix randomness entropy gathering to use poll() when
2926 possible instead of select(), since the latter has some
2927 undesirable limitations.
2928 [Darryl Miles via Richard Levitte and Bodo Moeller]
2930 *) Disable rogue ciphersuites:
2932 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2933 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2934 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2936 The latter two were purportedly from
2937 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2940 Also deactive the remaining ciphersuites from
2941 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2942 unofficial, and the ID has long expired.
2945 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2946 dual-core machines) and other potential thread-safety issues.
2949 Changes between 0.9.7i and 0.9.7j [04 May 2006]
2951 *) Adapt fipsld and the build system to link against the validated FIPS
2952 module in FIPS mode.
2955 *) Fixes for VC++ 2005 build under Windows.
2958 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
2959 from a Windows bash shell such as MSYS. It is autodetected from the
2960 "config" script when run from a VC++ environment. Modify standard VC++
2961 build to use fipscanister.o from the GNU make build.
2964 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
2966 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
2967 The value now differs depending on if you build for FIPS or not.
2968 BEWARE! A program linked with a shared FIPSed libcrypto can't be
2969 safely run with a non-FIPSed libcrypto, as it may crash because of
2970 the difference induced by this change.
2973 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
2975 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2976 (part of SSL_OP_ALL). This option used to disable the
2977 countermeasure against man-in-the-middle protocol-version
2978 rollback in the SSL 2.0 server implementation, which is a bad
2979 idea. (CVE-2005-2969)
2981 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2982 for Information Security, National Institute of Advanced Industrial
2983 Science and Technology [AIST], Japan)]
2985 *) Minimal support for X9.31 signatures and PSS padding modes. This is
2986 mainly for FIPS compliance and not fully integrated at this stage.
2989 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
2990 the exponentiation using a fixed-length exponent. (Otherwise,
2991 the information leaked through timing could expose the secret key
2992 after many signatures; cf. Bleichenbacher's attack on DSA with
2996 *) Make a new fixed-window mod_exp implementation the default for
2997 RSA, DSA, and DH private-key operations so that the sequence of
2998 squares and multiplies and the memory access pattern are
2999 independent of the particular secret key. This will mitigate
3000 cache-timing and potential related attacks.
3002 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3003 and this is automatically used by BN_mod_exp_mont() if the new flag
3004 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3005 will use this BN flag for private exponents unless the flag
3006 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3007 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3009 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3011 *) Change the client implementation for SSLv23_method() and
3012 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3013 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3014 (Previously, the SSL 2.0 backwards compatible Client Hello
3015 message format would be used even with SSL_OP_NO_SSLv2.)
3018 *) Add support for smime-type MIME parameter in S/MIME messages which some
3022 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3023 a threadsafe manner. Modify rsa code to use new function and add calls
3024 to dsa and dh code (which had race conditions before).
3027 *) Include the fixed error library code in the C error file definitions
3028 instead of fixing them up at runtime. This keeps the error code
3029 structures constant.
3032 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3034 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3037 *) Fixes for newer kerberos headers. NB: the casts are needed because
3038 the 'length' field is signed on one version and unsigned on another
3039 with no (?) obvious way to tell the difference, without these VC++
3040 complains. Also the "definition" of FAR (blank) is no longer included
3041 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3042 some needed definitions.
3045 *) Undo Cygwin change.
3048 *) Added support for proxy certificates according to RFC 3820.
3049 Because they may be a security thread to unaware applications,
3050 they must be explicitely allowed in run-time. See
3051 docs/HOWTO/proxy_certificates.txt for further information.
3054 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3056 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3057 server and client random values. Previously
3058 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3059 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3061 This change has negligible security impact because:
3063 1. Server and client random values still have 24 bytes of pseudo random
3066 2. Server and client random values are sent in the clear in the initial
3069 3. The master secret is derived using the premaster secret (48 bytes in
3070 size for static RSA ciphersuites) as well as client server and random
3073 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3076 [Stephen Henson, reported by UK NISCC]
3078 *) Use Windows randomness collection on Cygwin.
3081 *) Fix hang in EGD/PRNGD query when communication socket is closed
3082 prematurely by EGD/PRNGD.
3083 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3085 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3088 *) Back-port of selected performance improvements from development
3089 branch, as well as improved support for PowerPC platforms.
3092 *) Add lots of checks for memory allocation failure, error codes to indicate
3093 failure and freeing up memory if a failure occurs.
3094 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3096 *) Add new -passin argument to dgst.
3099 *) Perform some character comparisons of different types in X509_NAME_cmp:
3100 this is needed for some certificates that reencode DNs into UTF8Strings
3101 (in violation of RFC3280) and can't or wont issue name rollover
3105 *) Make an explicit check during certificate validation to see that
3106 the CA setting in each certificate on the chain is correct. As a
3107 side effect always do the following basic checks on extensions,
3108 not just when there's an associated purpose to the check:
3110 - if there is an unhandled critical extension (unless the user
3111 has chosen to ignore this fault)
3112 - if the path length has been exceeded (if one is set at all)
3113 - that certain extensions fit the associated purpose (if one has
3117 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3119 *) Avoid a race condition when CRLs are checked in a multi threaded
3120 environment. This would happen due to the reordering of the revoked
3121 entries during signature checking and serial number lookup. Now the
3122 encoding is cached and the serial number sort performed under a lock.
3123 Add new STACK function sk_is_sorted().
3126 *) Add Delta CRL to the extension code.
3129 *) Various fixes to s3_pkt.c so alerts are sent properly.
3130 [David Holmes <d.holmes@f5.com>]
3132 *) Reduce the chances of duplicate issuer name and serial numbers (in
3133 violation of RFC3280) using the OpenSSL certificate creation utilities.
3134 This is done by creating a random 64 bit value for the initial serial
3135 number when a serial number file is created or when a self signed
3136 certificate is created using 'openssl req -x509'. The initial serial
3137 number file is created using 'openssl x509 -next_serial' in CA.pl
3138 rather than being initialized to 1.
3141 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3143 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3144 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3145 [Joe Orton, Steve Henson]
3147 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3149 [Joe Orton, Steve Henson]
3151 *) Make it possible to have multiple active certificates with the same
3152 subject in the CA index file. This is done only if the keyword
3153 'unique_subject' is set to 'no' in the main CA section (default
3154 if 'CA_default') of the configuration file. The value is saved
3155 with the database itself in a separate index attribute file,
3156 named like the index file with '.attr' appended to the name.
3159 *) X509 verify fixes. Disable broken certificate workarounds when
3160 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3161 keyUsage extension present. Don't accept CRLs with unhandled critical
3162 extensions: since verify currently doesn't process CRL extensions this
3163 rejects a CRL with *any* critical extensions. Add new verify error codes
3167 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
3168 A clarification of RFC2560 will require the use of OCTET STRINGs and
3169 some implementations cannot handle the current raw format. Since OpenSSL
3170 copies and compares OCSP nonces as opaque blobs without any attempt at
3171 parsing them this should not create any compatibility issues.
3174 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
3175 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
3176 this HMAC (and other) operations are several times slower than OpenSSL
3180 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
3181 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3183 *) Use the correct content when signing type "other".
3186 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
3188 *) Fix various bugs revealed by running the NISCC test suite:
3190 Stop out of bounds reads in the ASN1 code when presented with
3191 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3193 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
3195 If verify callback ignores invalid public key errors don't try to check
3196 certificate signature with the NULL public key.
3200 *) New -ignore_err option in ocsp application to stop the server
3201 exiting on the first error in a request.
3204 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3205 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3209 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3210 extra data after the compression methods not only for TLS 1.0
3211 but also for SSL 3.0 (as required by the specification).
3212 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3214 *) Change X509_certificate_type() to mark the key as exported/exportable
3215 when it's 512 *bits* long, not 512 bytes.
3218 *) Change AES_cbc_encrypt() so it outputs exact multiple of
3219 blocks during encryption.
3222 *) Various fixes to base64 BIO and non blocking I/O. On write
3223 flushes were not handled properly if the BIO retried. On read
3224 data was not being buffered properly and had various logic bugs.
3225 This also affects blocking I/O when the data being decoded is a
3229 *) Various S/MIME bugfixes and compatibility changes:
3230 output correct application/pkcs7 MIME type if
3231 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
3232 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
3233 of files as .eml work). Correctly handle very long lines in MIME
3237 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
3239 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3240 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3241 a protocol version number mismatch like a decryption error
3242 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3245 *) Turn on RSA blinding by default in the default implementation
3246 to avoid a timing attack. Applications that don't want it can call
3247 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3248 They would be ill-advised to do so in most cases.
3249 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3251 *) Change RSA blinding code so that it works when the PRNG is not
3252 seeded (in this case, the secret RSA exponent is abused as
3253 an unpredictable seed -- if it is not unpredictable, there
3254 is no point in blinding anyway). Make RSA blinding thread-safe
3255 by remembering the creator's thread ID in rsa->blinding and
3256 having all other threads use local one-time blinding factors
3257 (this requires more computation than sharing rsa->blinding, but
3258 avoids excessive locking; and if an RSA object is not shared
3259 between threads, blinding will still be very fast).
3262 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
3263 ENGINE as defaults for all supported algorithms irrespective of
3264 the 'flags' parameter. 'flags' is now honoured, so applications
3265 should make sure they are passing it correctly.
3268 *) Target "mingw" now allows native Windows code to be generated in
3269 the Cygwin environment as well as with the MinGW compiler.
3272 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
3274 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3275 via timing by performing a MAC computation even if incorrrect
3276 block cipher padding has been found. This is a countermeasure
3277 against active attacks where the attacker has to distinguish
3278 between bad padding and a MAC verification error. (CVE-2003-0078)
3280 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3281 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3282 Martin Vuagnoux (EPFL, Ilion)]
3284 *) Make the no-err option work as intended. The intention with no-err
3285 is not to have the whole error stack handling routines removed from
3286 libcrypto, it's only intended to remove all the function name and
3287 reason texts, thereby removing some of the footprint that may not
3288 be interesting if those errors aren't displayed anyway.
3290 NOTE: it's still possible for any application or module to have it's
3291 own set of error texts inserted. The routines are there, just not
3292 used by default when no-err is given.
3295 *) Add support for FreeBSD on IA64.
3296 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
3298 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
3299 Kerberos function mit_des_cbc_cksum(). Before this change,
3300 the value returned by DES_cbc_cksum() was like the one from
3301 mit_des_cbc_cksum(), except the bytes were swapped.
3302 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
3304 *) Allow an application to disable the automatic SSL chain building.
3305 Before this a rather primitive chain build was always performed in
3306 ssl3_output_cert_chain(): an application had no way to send the
3307 correct chain if the automatic operation produced an incorrect result.
3309 Now the chain builder is disabled if either:
3311 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
3313 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
3315 The reasoning behind this is that an application would not want the
3316 auto chain building to take place if extra chain certificates are
3317 present and it might also want a means of sending no additional
3318 certificates (for example the chain has two certificates and the
3322 *) Add the possibility to build without the ENGINE framework.
3323 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3325 *) Under Win32 gmtime() can return NULL: check return value in
3326 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
3329 *) DSA routines: under certain error conditions uninitialized BN objects
3330 could be freed. Solution: make sure initialization is performed early
3331 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
3332 Nils Larsch <nla@trustcenter.de> via PR#459)
3335 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
3336 checked on reconnect on the client side, therefore session resumption
3337 could still fail with a "ssl session id is different" error. This
3338 behaviour is masked when SSL_OP_ALL is used due to
3339 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
3340 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3341 followup to PR #377.
3344 *) IA-32 assembler support enhancements: unified ELF targets, support
3345 for SCO/Caldera platforms, fix for Cygwin shared build.
3348 *) Add support for FreeBSD on sparc64. As a consequence, support for
3349 FreeBSD on non-x86 processors is separate from x86 processors on
3350 the config script, much like the NetBSD support.
3351 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
3353 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
3355 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
3358 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
3359 code (06) was taken as the first octet of the session ID and the last
3360 octet was ignored consequently. As a result SSLv2 client side session
3361 caching could not have worked due to the session ID mismatch between
3363 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3367 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
3368 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
3372 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
3373 seems that in spite of existing for more than a year, many application
3374 author have done nothing to provide the necessary callbacks, which
3375 means that this particular engine will not work properly anywhere.
3376 This is a very unfortunate situation which forces us, in the name
3377 of usability, to give the hw_ncipher.c a static lock, which is part
3379 NOTE: This is for the 0.9.7 series ONLY. This hack will never
3380 appear in 0.9.8 or later. We EXPECT application authors to have
3381 dealt properly with this when 0.9.8 is released (unless we actually
3382 make such changes in the libcrypto locking code that changes will
3383 have to be made anyway).
3386 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
3387 octets have been read, EOF or an error occurs. Without this change
3388 some truncated ASN1 structures will not produce an error.
3391 *) Disable Heimdal support, since it hasn't been fully implemented.
3392 Still give the possibility to force the use of Heimdal, but with
3393 warnings and a request that patches get sent to openssl-dev.
3396 *) Add the VC-CE target, introduce the WINCE sysname, and add
3397 INSTALL.WCE and appropriate conditionals to make it build.
3398 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3400 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
3401 cygssl-x.y.z.dll, where x, y and z are the major, minor and
3402 edit numbers of the version.
3403 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3405 *) Introduce safe string copy and catenation functions
3406 (BUF_strlcpy() and BUF_strlcat()).
3407 [Ben Laurie (CHATS) and Richard Levitte]
3409 *) Avoid using fixed-size buffers for one-line DNs.
3410 [Ben Laurie (CHATS)]
3412 *) Add BUF_MEM_grow_clean() to avoid information leakage when
3413 resizing buffers containing secrets, and use where appropriate.
3414 [Ben Laurie (CHATS)]
3416 *) Avoid using fixed size buffers for configuration file location.
3417 [Ben Laurie (CHATS)]
3419 *) Avoid filename truncation for various CA files.
3420 [Ben Laurie (CHATS)]
3422 *) Use sizeof in preference to magic numbers.
3423 [Ben Laurie (CHATS)]
3425 *) Avoid filename truncation in cert requests.
3426 [Ben Laurie (CHATS)]
3428 *) Add assertions to check for (supposedly impossible) buffer
3430 [Ben Laurie (CHATS)]
3432 *) Don't cache truncated DNS entries in the local cache (this could
3433 potentially lead to a spoofing attack).
3434 [Ben Laurie (CHATS)]
3436 *) Fix various buffers to be large enough for hex/decimal
3437 representations in a platform independent manner.
3438 [Ben Laurie (CHATS)]
3440 *) Add CRYPTO_realloc_clean() to avoid information leakage when
3441 resizing buffers containing secrets, and use where appropriate.
3442 [Ben Laurie (CHATS)]
3444 *) Add BIO_indent() to avoid much slightly worrying code to do
3446 [Ben Laurie (CHATS)]
3448 *) Convert sprintf()/BIO_puts() to BIO_printf().
3449 [Ben Laurie (CHATS)]
3451 *) buffer_gets() could terminate with the buffer only half
3453 [Ben Laurie (CHATS)]
3455 *) Add assertions to prevent user-supplied crypto functions from
3456 overflowing internal buffers by having large block sizes, etc.
3457 [Ben Laurie (CHATS)]
3459 *) New OPENSSL_assert() macro (similar to assert(), but enabled
3461 [Ben Laurie (CHATS)]
3463 *) Eliminate unused copy of key in RC4.
3464 [Ben Laurie (CHATS)]
3466 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
3467 [Ben Laurie (CHATS)]
3469 *) Fix off-by-one error in EGD path.
3470 [Ben Laurie (CHATS)]
3472 *) If RANDFILE path is too long, ignore instead of truncating.
3473 [Ben Laurie (CHATS)]
3475 *) Eliminate unused and incorrectly sized X.509 structure
3477 [Ben Laurie (CHATS)]
3479 *) Eliminate unused and dangerous function knumber().
3480 [Ben Laurie (CHATS)]
3482 *) Eliminate unused and dangerous structure, KSSL_ERR.
3483 [Ben Laurie (CHATS)]
3485 *) Protect against overlong session ID context length in an encoded
3486 session object. Since these are local, this does not appear to be
3488 [Ben Laurie (CHATS)]
3490 *) Change from security patch (see 0.9.6e below) that did not affect
3491 the 0.9.6 release series:
3493 Remote buffer overflow in SSL3 protocol - an attacker could
3494 supply an oversized master key in Kerberos-enabled versions.
3496 [Ben Laurie (CHATS)]
3498 *) Change the SSL kerb5 codes to match RFC 2712.
3501 *) Make -nameopt work fully for req and add -reqopt switch.
3502 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
3504 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
3505 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
3507 *) Make sure tests can be performed even if the corresponding algorithms
3508 have been removed entirely. This was also the last step to make
3509 OpenSSL compilable with DJGPP under all reasonable conditions.
3510 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
3512 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
3513 to allow version independent disabling of normally unselected ciphers,
3514 which may be activated as a side-effect of selecting a single cipher.
3516 (E.g., cipher list string "RSA" enables ciphersuites that are left
3517 out of "ALL" because they do not provide symmetric encryption.
3518 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
3519 [Lutz Jaenicke, Bodo Moeller]
3521 *) Add appropriate support for separate platform-dependent build
3522 directories. The recommended way to make a platform-dependent
3523 build directory is the following (tested on Linux), maybe with
3526 # Place yourself outside of the OpenSSL source tree. In
3527 # this example, the environment variable OPENSSL_SOURCE
3528 # is assumed to contain the absolute OpenSSL source directory.
3529 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
3530 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
3531 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
3532 mkdir -p `dirname $F`
3533 ln -s $OPENSSL_SOURCE/$F $F
3536 To be absolutely sure not to disturb the source tree, a "make clean"
3537 is a good thing. If it isn't successfull, don't worry about it,
3538 it probably means the source directory is very clean.
3541 *) Make sure any ENGINE control commands make local copies of string
3542 pointers passed to them whenever necessary. Otherwise it is possible
3543 the caller may have overwritten (or deallocated) the original string
3544 data when a later ENGINE operation tries to use the stored values.
3545 [Götz Babin-Ebell <babinebell@trustcenter.de>]
3547 *) Improve diagnostics in file reading and command-line digests.
3548 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
3550 *) Add AES modes CFB and OFB to the object database. Correct an
3551 error in AES-CFB decryption.
3554 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
3555 allows existing EVP_CIPHER_CTX structures to be reused after
3556 calling EVP_*Final(). This behaviour is used by encryption
3557 BIOs and some applications. This has the side effect that
3558 applications must explicitly clean up cipher contexts with
3559 EVP_CIPHER_CTX_cleanup() or they will leak memory.
3562 *) Check the values of dna and dnb in bn_mul_recursive before calling
3563 bn_mul_comba (a non zero value means the a or b arrays do not contain
3564 n2 elements) and fallback to bn_mul_normal if either is not zero.
3567 *) Fix escaping of non-ASCII characters when using the -subj option
3568 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
3571 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
3572 form for "surname", serialNumber has no short form.
3573 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
3574 therefore remove "mail" short name for "internet 7".
3575 The OID for unique identifiers in X509 certificates is
3576 x500UniqueIdentifier, not uniqueIdentifier.
3577 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
3580 *) Add an "init" command to the ENGINE config module and auto initialize
3581 ENGINEs. Without any "init" command the ENGINE will be initialized
3582 after all ctrl commands have been executed on it. If init=1 the
3583 ENGINE is initailized at that point (ctrls before that point are run
3584 on the uninitialized ENGINE and after on the initialized one). If
3585 init=0 then the ENGINE will not be iniatialized at all.
3588 *) Fix the 'app_verify_callback' interface so that the user-defined
3589 argument is actually passed to the callback: In the
3590 SSL_CTX_set_cert_verify_callback() prototype, the callback
3591 declaration has been changed from
3594 int (*cb)(X509_STORE_CTX *,void *);
3595 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
3596 i=s->ctx->app_verify_callback(&ctx)
3597 has been changed into
3598 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
3600 To update applications using SSL_CTX_set_cert_verify_callback(),
3601 a dummy argument can be added to their callback functions.
3602 [D. K. Smetters <smetters@parc.xerox.com>]
3604 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
3605 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
3607 *) Add and OPENSSL_LOAD_CONF define which will cause
3608 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
3609 This allows older applications to transparently support certain
3610 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
3611 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
3612 load the config file and OPENSSL_add_all_algorithms_conf() which will
3613 always load it have also been added.
3616 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
3617 Adjust NIDs and EVP layer.
3618 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3620 *) Config modules support in openssl utility.
3622 Most commands now load modules from the config file,
3623 though in a few (such as version) this isn't done
3624 because it couldn't be used for anything.
3626 In the case of ca and req the config file used is
3627 the same as the utility itself: that is the -config
3628 command line option can be used to specify an
3632 *) Move default behaviour from OPENSSL_config(). If appname is NULL
3633 use "openssl_conf" if filename is NULL use default openssl config file.
3636 *) Add an argument to OPENSSL_config() to allow the use of an alternative
3637 config section name. Add a new flag to tolerate a missing config file
3638 and move code to CONF_modules_load_file().
3641 *) Support for crypto accelerator cards from Accelerated Encryption
3642 Processing, www.aep.ie. (Use engine 'aep')
3643 The support was copied from 0.9.6c [engine] and adapted/corrected
3644 to work with the new engine framework.
3645 [AEP Inc. and Richard Levitte]
3647 *) Support for SureWare crypto accelerator cards from Baltimore
3648 Technologies. (Use engine 'sureware')
3649 The support was copied from 0.9.6c [engine] and adapted
3650 to work with the new engine framework.
3653 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
3654 make the newer ENGINE framework commands for the CHIL engine work.
3655 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
3657 *) Make it possible to produce shared libraries on ReliantUNIX.
3658 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
3660 *) Add the configuration target debug-linux-ppro.
3661 Make 'openssl rsa' use the general key loading routines
3662 implemented in apps.c, and make those routines able to
3663 handle the key format FORMAT_NETSCAPE and the variant
3665 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3667 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3668 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3670 *) Add -keyform to rsautl, and document -engine.
3671 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
3673 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
3674 BIO_R_NO_SUCH_FILE error code rather than the generic
3675 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
3678 *) Add new functions
3680 ERR_peek_last_error_line
3681 ERR_peek_last_error_line_data.
3682 These are similar to
3685 ERR_peek_error_line_data,
3686 but report on the latest error recorded rather than the first one
3687 still in the error queue.
3688 [Ben Laurie, Bodo Moeller]
3690 *) default_algorithms option in ENGINE config module. This allows things
3692 default_algorithms = ALL
3693 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
3696 *) Prelminary ENGINE config module.
3699 *) New experimental application configuration code.
3702 *) Change the AES code to follow the same name structure as all other
3703 symmetric ciphers, and behave the same way. Move everything to
3704 the directory crypto/aes, thereby obsoleting crypto/rijndael.
3705 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3707 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
3708 [Ben Laurie and Theo de Raadt]
3710 *) Add option to output public keys in req command.
3711 [Massimiliano Pala madwolf@openca.org]
3713 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
3714 (up to about 10% better than before for P-192 and P-224).
3717 *) New functions/macros
3719 SSL_CTX_set_msg_callback(ctx, cb)
3720 SSL_CTX_set_msg_callback_arg(ctx, arg)
3721 SSL_set_msg_callback(ssl, cb)
3722 SSL_set_msg_callback_arg(ssl, arg)
3724 to request calling a callback function
3726 void cb(int write_p, int version, int content_type,
3727 const void *buf, size_t len, SSL *ssl, void *arg)
3729 whenever a protocol message has been completely received
3730 (write_p == 0) or sent (write_p == 1). Here 'version' is the
3731 protocol version according to which the SSL library interprets
3732 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
3733 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
3734 the content type as defined in the SSL 3.0/TLS 1.0 protocol
3735 specification (change_cipher_spec(20), alert(21), handshake(22)).
3736 'buf' and 'len' point to the actual message, 'ssl' to the
3737 SSL object, and 'arg' is the application-defined value set by
3738 SSL[_CTX]_set_msg_callback_arg().
3740 'openssl s_client' and 'openssl s_server' have new '-msg' options
3741 to enable a callback that displays all protocol messages.
3744 *) Change the shared library support so shared libraries are built as
3745 soon as the corresponding static library is finished, and thereby get
3746 openssl and the test programs linked against the shared library.
3747 This still only happens when the keyword "shard" has been given to
3748 the configuration scripts.
3750 NOTE: shared library support is still an experimental thing, and
3751 backward binary compatibility is still not guaranteed.
3752 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
3754 *) Add support for Subject Information Access extension.
3755 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3757 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
3758 additional bytes when new memory had to be allocated, not just
3759 when reusing an existing buffer.
3762 *) New command line and configuration option 'utf8' for the req command.
3763 This allows field values to be specified as UTF8 strings.
3766 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
3767 runs for the former and machine-readable output for the latter.
3770 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
3771 of the e-mail address in the DN (i.e., it will go into a certificate
3772 extension only). The new configuration file option 'email_in_dn = no'
3773 has the same effect.
3774 [Massimiliano Pala madwolf@openca.org]
3776 *) Change all functions with names starting with des_ to be starting
3777 with DES_ instead. Add wrappers that are compatible with libdes,
3778 but are named _ossl_old_des_*. Finally, add macros that map the
3779 des_* symbols to the corresponding _ossl_old_des_* if libdes
3780 compatibility is desired. If OpenSSL 0.9.6c compatibility is
3781 desired, the des_* symbols will be mapped to DES_*, with one
3784 Since we provide two compatibility mappings, the user needs to
3785 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
3786 compatibility is desired. The default (i.e., when that macro
3787 isn't defined) is OpenSSL 0.9.6c compatibility.
3789 There are also macros that enable and disable the support of old
3790 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
3791 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
3792 are defined, the default will apply: to support the old des routines.
3794 In either case, one must include openssl/des.h to get the correct
3795 definitions. Do not try to just include openssl/des_old.h, that
3798 NOTE: This is a major break of an old API into a new one. Software
3799 authors are encouraged to switch to the DES_ style functions. Some
3800 time in the future, des_old.h and the libdes compatibility functions
3801 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
3802 default), and then completely removed.
3805 *) Test for certificates which contain unsupported critical extensions.
3806 If such a certificate is found during a verify operation it is
3807 rejected by default: this behaviour can be overridden by either
3808 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
3809 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
3810 X509_supported_extension() has also been added which returns 1 if a
3811 particular extension is supported.
3814 *) Modify the behaviour of EVP cipher functions in similar way to digests
3815 to retain compatibility with existing code.
3818 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
3819 compatibility with existing code. In particular the 'ctx' parameter does
3820 not have to be to be initialized before the call to EVP_DigestInit() and
3821 it is tidied up after a call to EVP_DigestFinal(). New function
3822 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
3823 EVP_MD_CTX_copy() changed to not require the destination to be
3824 initialized valid and new function EVP_MD_CTX_copy_ex() added which
3825 requires the destination to be valid.
3827 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
3828 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
3831 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
3832 so that complete 'Handshake' protocol structures are kept in memory
3833 instead of overwriting 'msg_type' and 'length' with 'body' data.
3836 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
3837 [Massimo Santin via Richard Levitte]
3839 *) Major restructuring to the underlying ENGINE code. This includes
3840 reduction of linker bloat, separation of pure "ENGINE" manipulation
3841 (initialisation, etc) from functionality dealing with implementations
3842 of specific crypto iterfaces. This change also introduces integrated
3843 support for symmetric ciphers and digest implementations - so ENGINEs
3844 can now accelerate these by providing EVP_CIPHER and EVP_MD
3845 implementations of their own. This is detailed in crypto/engine/README
3846 as it couldn't be adequately described here. However, there are a few
3847 API changes worth noting - some RSA, DSA, DH, and RAND functions that
3848 were changed in the original introduction of ENGINE code have now
3849 reverted back - the hooking from this code to ENGINE is now a good
3850 deal more passive and at run-time, operations deal directly with
3851 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
3852 dereferencing through an ENGINE pointer any more. Also, the ENGINE
3853 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
3854 they were not being used by the framework as there is no concept of a
3855 BIGNUM_METHOD and they could not be generalised to the new
3856 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
3857 ENGINE_cpy() has been removed as it cannot be consistently defined in
3861 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
3864 *) Change mkdef.pl to sort symbols that get the same entry number,
3865 and make sure the automatically generated functions ERR_load_*
3866 become part of libeay.num as well.
3869 *) New function SSL_renegotiate_pending(). This returns true once
3870 renegotiation has been requested (either SSL_renegotiate() call
3871 or HelloRequest/ClientHello receveived from the peer) and becomes
3872 false once a handshake has been completed.
3873 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
3874 sends a HelloRequest, but does not ensure that a handshake takes
3875 place. SSL_renegotiate_pending() is useful for checking if the
3876 client has followed the request.)
3879 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
3880 By default, clients may request session resumption even during
3881 renegotiation (if session ID contexts permit); with this option,
3882 session resumption is possible only in the first handshake.
3884 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
3885 more bits available for options that should not be part of
3886 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
3889 *) Add some demos for certificate and certificate request creation.
3892 *) Make maximum certificate chain size accepted from the peer application
3893 settable (SSL*_get/set_max_cert_list()), as proposed by
3894 "Douglas E. Engert" <deengert@anl.gov>.
3897 *) Add support for shared libraries for Unixware-7
3898 (Boyd Lynn Gerber <gerberb@zenez.com>).
3901 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
3902 be done prior to destruction. Use this to unload error strings from
3903 ENGINEs that load their own error strings. NB: This adds two new API
3904 functions to "get" and "set" this destroy handler in an ENGINE.
3907 *) Alter all existing ENGINE implementations (except "openssl" and
3908 "openbsd") to dynamically instantiate their own error strings. This
3909 makes them more flexible to be built both as statically-linked ENGINEs
3910 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
3911 Also, add stub code to each that makes building them as self-contained
3912 shared-libraries easier (see README.ENGINE).
3915 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
3916 implementations into applications that are completely implemented in
3917 self-contained shared-libraries. The "dynamic" ENGINE exposes control
3918 commands that can be used to configure what shared-library to load and
3919 to control aspects of the way it is handled. Also, made an update to
3920 the README.ENGINE file that brings its information up-to-date and
3921 provides some information and instructions on the "dynamic" ENGINE
3922 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
3925 *) Make it possible to unload ranges of ERR strings with a new
3926 "ERR_unload_strings" function.
3929 *) Add a copy() function to EVP_MD.
3932 *) Make EVP_MD routines take a context pointer instead of just the
3933 md_data void pointer.
3936 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
3937 that the digest can only process a single chunk of data
3938 (typically because it is provided by a piece of
3939 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
3940 is only going to provide a single chunk of data, and hence the
3941 framework needn't accumulate the data for oneshot drivers.
3944 *) As with "ERR", make it possible to replace the underlying "ex_data"
3945 functions. This change also alters the storage and management of global
3946 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
3947 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
3948 index counters. The API functions that use this state have been changed
3949 to take a "class_index" rather than pointers to the class's local STACK
3950 and counter, and there is now an API function to dynamically create new
3951 classes. This centralisation allows us to (a) plug a lot of the
3952 thread-safety problems that existed, and (b) makes it possible to clean
3953 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
3954 such data would previously have always leaked in application code and
3955 workarounds were in place to make the memory debugging turn a blind eye
3956 to it. Application code that doesn't use this new function will still
3957 leak as before, but their memory debugging output will announce it now
3958 rather than letting it slide.
3960 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
3961 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
3962 has a return value to indicate success or failure.
3965 *) Make it possible to replace the underlying "ERR" functions such that the
3966 global state (2 LHASH tables and 2 locks) is only used by the "default"
3967 implementation. This change also adds two functions to "get" and "set"
3968 the implementation prior to it being automatically set the first time
3969 any other ERR function takes place. Ie. an application can call "get",
3970 pass the return value to a module it has just loaded, and that module
3971 can call its own "set" function using that value. This means the
3972 module's "ERR" operations will use (and modify) the error state in the
3973 application and not in its own statically linked copy of OpenSSL code.
3976 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
3977 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
3978 the operation, and provides a more encapsulated way for external code
3979 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
3980 to use these functions rather than manually incrementing the counts.
3982 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
3985 *) Add EVP test program.
3988 *) Add symmetric cipher support to ENGINE. Expect the API to change!
3991 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
3992 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
3993 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
3994 These allow a CRL to be built without having to access X509_CRL fields
3995 directly. Modify 'ca' application to use new functions.
3998 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
3999 bug workarounds. Rollback attack detection is a security feature.
4000 The problem will only arise on OpenSSL servers when TLSv1 is not
4001 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4002 Software authors not wanting to support TLSv1 will have special reasons
4003 for their choice and can explicitly enable this option.
4004 [Bodo Moeller, Lutz Jaenicke]
4006 *) Rationalise EVP so it can be extended: don't include a union of
4007 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4008 (similar to those existing for EVP_CIPHER_CTX).
4013 EVP_MD_CTX_init(&md); /* new function call */
4014 EVP_DigestInit(&md, EVP_sha1());
4015 EVP_DigestUpdate(&md, in, len);
4016 EVP_DigestFinal(&md, out, NULL);
4017 EVP_MD_CTX_cleanup(&md); /* new function call */
4021 *) Make DES key schedule conform to the usual scheme, as well as
4022 correcting its structure. This means that calls to DES functions
4023 now have to pass a pointer to a des_key_schedule instead of a
4024 plain des_key_schedule (which was actually always a pointer
4027 des_key_schedule ks;
4029 des_set_key_checked(..., &ks);
4030 des_ncbc_encrypt(..., &ks, ...);
4032 (Note that a later change renames 'des_...' into 'DES_...'.)
4035 *) Initial reduction of linker bloat: the use of some functions, such as
4036 PEM causes large amounts of unused functions to be linked in due to
4037 poor organisation. For example pem_all.c contains every PEM function
4038 which has a knock on effect of linking in large amounts of (unused)
4039 ASN1 code. Grouping together similar functions and splitting unrelated
4040 functions prevents this.
4043 *) Cleanup of EVP macros.
4046 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4047 correct _ecb suffix.
4050 *) Add initial OCSP responder support to ocsp application. The
4051 revocation information is handled using the text based index
4052 use by the ca application. The responder can either handle
4053 requests generated internally, supplied in files (for example
4054 via a CGI script) or using an internal minimal server.
4057 *) Add configuration choices to get zlib compression for TLS.
4060 *) Changes to Kerberos SSL for RFC 2712 compliance:
4061 1. Implemented real KerberosWrapper, instead of just using
4062 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4063 2. Implemented optional authenticator field of KerberosWrapper.
4065 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4066 and authenticator structs; see crypto/krb5/.
4068 Generalized Kerberos calls to support multiple Kerberos libraries.
4069 [Vern Staats <staatsvr@asc.hpc.mil>,
4070 Jeffrey Altman <jaltman@columbia.edu>
4071 via Richard Levitte]
4073 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4074 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4075 values for each of the key sizes rather than having just
4076 parameters (and 'speed' generating keys each time).
4079 *) Speed up EVP routines.
4082 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4083 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4084 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4085 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4087 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4088 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4089 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4092 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4094 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4097 *) Added the OS2-EMX target.
4098 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4100 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4101 to support NCONF routines in extension code. New function CONF_set_nconf()
4102 to allow functions which take an NCONF to also handle the old LHASH
4103 structure: this means that the old CONF compatible routines can be
4104 retained (in particular wrt extensions) without having to duplicate the
4105 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4108 *) Enhance the general user interface with mechanisms for inner control
4109 and with possibilities to have yes/no kind of prompts.
4112 *) Change all calls to low level digest routines in the library and
4113 applications to use EVP. Add missing calls to HMAC_cleanup() and
4114 don't assume HMAC_CTX can be copied using memcpy().
4115 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4117 *) Add the possibility to control engines through control names but with
4118 arbitrary arguments instead of just a string.
4119 Change the key loaders to take a UI_METHOD instead of a callback
4120 function pointer. NOTE: this breaks binary compatibility with earlier
4121 versions of OpenSSL [engine].
4122 Adapt the nCipher code for these new conditions and add a card insertion
4126 *) Enhance the general user interface with mechanisms to better support
4127 dialog box interfaces, application-defined prompts, the possibility
4128 to use defaults (for example default passwords from somewhere else)
4129 and interrupts/cancellations.
4132 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4133 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4136 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4137 tidy up some unnecessarily weird code in 'sk_new()').
4138 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4140 *) Change the key loading routines for ENGINEs to use the same kind
4141 callback (pem_password_cb) as all other routines that need this
4145 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4146 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4147 than this minimum value is recommended.
4150 *) New random seeder for OpenVMS, using the system process statistics
4151 that are easily reachable.
4154 *) Windows apparently can't transparently handle global
4155 variables defined in DLLs. Initialisations such as:
4157 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4159 wont compile. This is used by the any applications that need to
4160 declare their own ASN1 modules. This was fixed by adding the option
4161 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4162 needed for static libraries under Win32.
4165 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
4166 setting of purpose and trust fields. New X509_STORE trust and
4167 purpose functions and tidy up setting in other SSL functions.
4170 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
4171 structure. These are inherited by X509_STORE_CTX when it is
4172 initialised. This allows various defaults to be set in the
4173 X509_STORE structure (such as flags for CRL checking and custom
4174 purpose or trust settings) for functions which only use X509_STORE_CTX
4175 internally such as S/MIME.
4177 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
4178 trust settings if they are not set in X509_STORE. This allows X509_STORE
4179 purposes and trust (in S/MIME for example) to override any set by default.
4181 Add command line options for CRL checking to smime, s_client and s_server
4185 *) Initial CRL based revocation checking. If the CRL checking flag(s)
4186 are set then the CRL is looked up in the X509_STORE structure and
4187 its validity and signature checked, then if the certificate is found
4188 in the CRL the verify fails with a revoked error.
4190 Various new CRL related callbacks added to X509_STORE_CTX structure.
4192 Command line options added to 'verify' application to support this.
4194 This needs some additional work, such as being able to handle multiple
4195 CRLs with different times, extension based lookup (rather than just
4196 by subject name) and ultimately more complete V2 CRL extension
4200 *) Add a general user interface API (crypto/ui/). This is designed
4201 to replace things like des_read_password and friends (backward
4202 compatibility functions using this new API are provided).
4203 The purpose is to remove prompting functions from the DES code
4204 section as well as provide for prompting through dialog boxes in
4205 a window system and the like.
4208 *) Add "ex_data" support to ENGINE so implementations can add state at a
4209 per-structure level rather than having to store it globally.
4212 *) Make it possible for ENGINE structures to be copied when retrieved by
4213 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
4214 This causes the "original" ENGINE structure to act like a template,
4215 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
4216 operational state can be localised to each ENGINE structure, despite the
4217 fact they all share the same "methods". New ENGINE structures returned in
4218 this case have no functional references and the return value is the single
4219 structural reference. This matches the single structural reference returned
4220 by ENGINE_by_id() normally, when it is incremented on the pre-existing
4224 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
4225 needs to match any other type at all we need to manually clear the
4229 *) Changes to the "openssl engine" utility to include;
4230 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
4231 about an ENGINE's available control commands.
4232 - executing control commands from command line arguments using the
4233 '-pre' and '-post' switches. '-post' is only used if '-t' is
4234 specified and the ENGINE is successfully initialised. The syntax for
4235 the individual commands are colon-separated, for example;
4236 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
4239 *) New dynamic control command support for ENGINEs. ENGINEs can now
4240 declare their own commands (numbers), names (strings), descriptions,
4241 and input types for run-time discovery by calling applications. A
4242 subset of these commands are implicitly classed as "executable"
4243 depending on their input type, and only these can be invoked through
4244 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
4245 can be based on user input, config files, etc). The distinction is
4246 that "executable" commands cannot return anything other than a boolean
4247 result and can only support numeric or string input, whereas some
4248 discoverable commands may only be for direct use through
4249 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
4250 pointers, or other custom uses. The "executable" commands are to
4251 support parameterisations of ENGINE behaviour that can be
4252 unambiguously defined by ENGINEs and used consistently across any
4253 OpenSSL-based application. Commands have been added to all the
4254 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
4255 control over shared-library paths without source code alterations.
4258 *) Changed all ENGINE implementations to dynamically allocate their
4259 ENGINEs rather than declaring them statically. Apart from this being
4260 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
4261 this also allows the implementations to compile without using the
4262 internal engine_int.h header.
4265 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
4266 'const' value. Any code that should be able to modify a RAND_METHOD
4267 should already have non-const pointers to it (ie. they should only
4268 modify their own ones).
4271 *) Made a variety of little tweaks to the ENGINE code.
4272 - "atalla" and "ubsec" string definitions were moved from header files
4273 to C code. "nuron" string definitions were placed in variables
4274 rather than hard-coded - allowing parameterisation of these values
4275 later on via ctrl() commands.
4276 - Removed unused "#if 0"'d code.
4277 - Fixed engine list iteration code so it uses ENGINE_free() to release
4278 structural references.
4279 - Constified the RAND_METHOD element of ENGINE structures.
4280 - Constified various get/set functions as appropriate and added
4281 missing functions (including a catch-all ENGINE_cpy that duplicates
4282 all ENGINE values onto a new ENGINE except reference counts/state).
4283 - Removed NULL parameter checks in get/set functions. Setting a method
4284 or function to NULL is a way of cancelling out a previously set
4285 value. Passing a NULL ENGINE parameter is just plain stupid anyway
4286 and doesn't justify the extra error symbols and code.
4287 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
4288 flags from engine_int.h to engine.h.
4289 - Changed prototypes for ENGINE handler functions (init(), finish(),
4290 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
4293 *) Implement binary inversion algorithm for BN_mod_inverse in addition
4294 to the algorithm using long division. The binary algorithm can be
4295 used only if the modulus is odd. On 32-bit systems, it is faster
4296 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
4297 roughly 5-15% for 256-bit moduli), so we use it only for moduli
4298 up to 450 bits. In 64-bit environments, the binary algorithm
4299 appears to be advantageous for much longer moduli; here we use it
4300 for moduli up to 2048 bits.
4303 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
4304 could not support the combine flag in choice fields.
4307 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
4308 extensions from a certificate request to the certificate.
4311 *) Allow multiple 'certopt' and 'nameopt' options to be separated
4312 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
4313 file: this allows the display of the certificate about to be
4314 signed to be customised, to allow certain fields to be included
4315 or excluded and extension details. The old system didn't display
4316 multicharacter strings properly, omitted fields not in the policy
4317 and couldn't display additional details such as extensions.
4320 *) Function EC_POINTs_mul for multiple scalar multiplication
4321 of an arbitrary number of elliptic curve points
4322 \sum scalars[i]*points[i],
4323 optionally including the generator defined for the EC_GROUP:
4324 scalar*generator + \sum scalars[i]*points[i].
4326 EC_POINT_mul is a simple wrapper function for the typical case
4327 that the point list has just one item (besides the optional
4331 *) First EC_METHODs for curves over GF(p):
4333 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
4334 operations and provides various method functions that can also
4335 operate with faster implementations of modular arithmetic.
4337 EC_GFp_mont_method() reuses most functions that are part of
4338 EC_GFp_simple_method, but uses Montgomery arithmetic.
4340 [Bodo Moeller; point addition and point doubling
4341 implementation directly derived from source code provided by
4342 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
4344 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
4345 crypto/ec/ec_lib.c):
4347 Curves are EC_GROUP objects (with an optional group generator)
4348 based on EC_METHODs that are built into the library.
4350 Points are EC_POINT objects based on EC_GROUP objects.
4352 Most of the framework would be able to handle curves over arbitrary
4353 finite fields, but as there are no obvious types for fields other
4354 than GF(p), some functions are limited to that for now.
4357 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
4358 that the file contains a complete HTTP response.
4361 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
4362 change the def and num file printf format specifier from "%-40sXXX"
4363 to "%-39s XXX". The latter will always guarantee a space after the
4364 field while the former will cause them to run together if the field
4365 is 40 of more characters long.
4368 *) Constify the cipher and digest 'method' functions and structures
4369 and modify related functions to take constant EVP_MD and EVP_CIPHER
4373 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
4374 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
4377 *) Modify EVP_Digest*() routines so they now return values. Although the
4378 internal software routines can never fail additional hardware versions
4382 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
4384 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
4385 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
4388 ERR_R_NESTED_ASN1_ERROR
4390 ERR_R_MISSING_ASN1_EOS
4391 were 4 .. 9, conflicting with
4392 ERR_LIB_RSA (= ERR_R_RSA_LIB)
4394 ERR_LIB_PEM (= ERR_R_PEM_LIB).
4395 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
4397 Add new error code 'ERR_R_INTERNAL_ERROR'.
4400 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
4404 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
4405 sets the subject name for a new request or supersedes the
4406 subject name in a given request. Formats that can be parsed are
4407 'CN=Some Name, OU=myOU, C=IT'
4409 'CN=Some Name/OU=myOU/C=IT'.
4411 Add options '-batch' and '-verbose' to 'openssl req'.
4412 [Massimiliano Pala <madwolf@hackmasters.net>]
4414 *) Introduce the possibility to access global variables through
4415 functions on platform were that's the best way to handle exporting
4416 global variables in shared libraries. To enable this functionality,
4417 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
4418 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
4419 is normally done by Configure or something similar).
4421 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
4422 in the source file (foo.c) like this:
4424 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
4425 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
4427 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
4428 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
4430 OPENSSL_DECLARE_GLOBAL(int,foo);
4431 #define foo OPENSSL_GLOBAL_REF(foo)
4432 OPENSSL_DECLARE_GLOBAL(double,bar);
4433 #define bar OPENSSL_GLOBAL_REF(bar)
4435 The #defines are very important, and therefore so is including the
4436 header file everywhere where the defined globals are used.
4438 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
4439 of ASN.1 items, but that structure is a bit different.
4441 The largest change is in util/mkdef.pl which has been enhanced with
4442 better and easier to understand logic to choose which symbols should
4443 go into the Windows .def files as well as a number of fixes and code
4444 cleanup (among others, algorithm keywords are now sorted
4445 lexicographically to avoid constant rewrites).
4448 *) In BN_div() keep a copy of the sign of 'num' before writing the
4449 result to 'rm' because if rm==num the value will be overwritten
4450 and produce the wrong result if 'num' is negative: this caused
4451 problems with BN_mod() and BN_nnmod().
4454 *) Function OCSP_request_verify(). This checks the signature on an
4455 OCSP request and verifies the signer certificate. The signer
4456 certificate is just checked for a generic purpose and OCSP request
4460 *) Add OCSP_check_validity() function to check the validity of OCSP
4461 responses. OCSP responses are prepared in real time and may only
4462 be a few seconds old. Simply checking that the current time lies
4463 between thisUpdate and nextUpdate max reject otherwise valid responses
4464 caused by either OCSP responder or client clock inaccuracy. Instead
4465 we allow thisUpdate and nextUpdate to fall within a certain period of
4466 the current time. The age of the response can also optionally be
4467 checked. Two new options -validity_period and -status_age added to
4471 *) If signature or public key algorithm is unrecognized print out its
4472 OID rather that just UNKNOWN.
4475 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
4476 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
4477 ID to be generated from the issuer certificate alone which can then be
4478 passed to OCSP_id_issuer_cmp().
4481 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
4482 ASN1 modules to export functions returning ASN1_ITEM pointers
4483 instead of the ASN1_ITEM structures themselves. This adds several
4484 new macros which allow the underlying ASN1 function/structure to
4485 be accessed transparently. As a result code should not use ASN1_ITEM
4486 references directly (such as &X509_it) but instead use the relevant
4487 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
4488 use of the new ASN1 code on platforms where exporting structures
4489 is problematical (for example in shared libraries) but exporting
4490 functions returning pointers to structures is not.
4493 *) Add support for overriding the generation of SSL/TLS session IDs.
4494 These callbacks can be registered either in an SSL_CTX or per SSL.
4495 The purpose of this is to allow applications to control, if they wish,
4496 the arbitrary values chosen for use as session IDs, particularly as it
4497 can be useful for session caching in multiple-server environments. A
4498 command-line switch for testing this (and any client code that wishes
4499 to use such a feature) has been added to "s_server".
4500 [Geoff Thorpe, Lutz Jaenicke]
4502 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
4503 of the form '#if defined(...) || defined(...) || ...' and
4504 '#if !defined(...) && !defined(...) && ...'. This also avoids
4505 the growing number of special cases it was previously handling.
4508 *) Make all configuration macros available for application by making
4509 sure they are available in opensslconf.h, by giving them names starting
4510 with "OPENSSL_" to avoid conflicts with other packages and by making
4511 sure e_os2.h will cover all platform-specific cases together with
4513 Additionally, it is now possible to define configuration/platform-
4514 specific names (called "system identities"). In the C code, these
4515 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
4516 macro with the name beginning with "OPENSSL_SYS_", which is determined
4517 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
4521 *) New option -set_serial to 'req' and 'x509' this allows the serial
4522 number to use to be specified on the command line. Previously self
4523 signed certificates were hard coded with serial number 0 and the
4524 CA options of 'x509' had to use a serial number in a file which was
4528 *) New options to 'ca' utility to support V2 CRL entry extensions.
4529 Currently CRL reason, invalidity date and hold instruction are
4530 supported. Add new CRL extensions to V3 code and some new objects.
4533 *) New function EVP_CIPHER_CTX_set_padding() this is used to
4534 disable standard block padding (aka PKCS#5 padding) in the EVP
4535 API, which was previously mandatory. This means that the data is
4536 not padded in any way and so the total length much be a multiple
4537 of the block size, otherwise an error occurs.
4540 *) Initial (incomplete) OCSP SSL support.
4543 *) New function OCSP_parse_url(). This splits up a URL into its host,
4544 port and path components: primarily to parse OCSP URLs. New -url
4545 option to ocsp utility.
4548 *) New nonce behavior. The return value of OCSP_check_nonce() now
4549 reflects the various checks performed. Applications can decide
4550 whether to tolerate certain situations such as an absent nonce
4551 in a response when one was present in a request: the ocsp application
4552 just prints out a warning. New function OCSP_add1_basic_nonce()
4553 this is to allow responders to include a nonce in a response even if
4554 the request is nonce-less.
4557 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
4558 skipped when using openssl x509 multiple times on a single input file,
4559 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
4562 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
4563 set string type: to handle setting ASN1_TIME structures. Fix ca
4564 utility to correctly initialize revocation date of CRLs.
4567 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
4568 the clients preferred ciphersuites and rather use its own preferences.
4569 Should help to work around M$ SGC (Server Gated Cryptography) bug in
4570 Internet Explorer by ensuring unchanged hash method during stepup.
4571 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
4574 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
4575 to aes and add a new 'exist' option to print out symbols that don't
4579 *) Additional options to ocsp utility to allow flags to be set and
4580 additional certificates supplied.
4583 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
4584 OCSP client a number of certificate to only verify the response
4588 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
4589 handle the new API. Currently only ECB, CBC modes supported. Add new
4592 Add TLS AES ciphersuites as described in RFC3268, "Advanced
4593 Encryption Standard (AES) Ciphersuites for Transport Layer
4594 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
4595 not enabled by default and were not part of the "ALL" ciphersuite
4596 alias because they were not yet official; they could be
4597 explicitly requested by specifying the "AESdraft" ciphersuite
4598 group alias. In the final release of OpenSSL 0.9.7, the group
4599 alias is called "AES" and is part of "ALL".)
4600 [Ben Laurie, Steve Henson, Bodo Moeller]
4602 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
4603 request to response.
4606 *) Functions for OCSP responders. OCSP_request_onereq_count(),
4607 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
4608 extract information from a certificate request. OCSP_response_create()
4609 creates a response and optionally adds a basic response structure.
4610 OCSP_basic_add1_status() adds a complete single response to a basic
4611 response and returns the OCSP_SINGLERESP structure just added (to allow
4612 extensions to be included for example). OCSP_basic_add1_cert() adds a
4613 certificate to a basic response and OCSP_basic_sign() signs a basic
4614 response with various flags. New helper functions ASN1_TIME_check()
4615 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
4616 (converts ASN1_TIME to GeneralizedTime).
4619 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
4620 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
4621 structure from a certificate. X509_pubkey_digest() digests the public_key
4622 contents: this is used in various key identifiers.
4625 *) Make sk_sort() tolerate a NULL argument.
4626 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
4628 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
4629 passed by the function are trusted implicitly. If any of them signed the
4630 response then it is assumed to be valid and is not verified.
4633 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
4634 to data. This was previously part of the PKCS7 ASN1 code. This
4635 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
4636 [Steve Henson, reported by Kenneth R. Robinette
4637 <support@securenetterm.com>]
4639 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
4640 routines: without these tracing memory leaks is very painful.
4641 Fix leaks in PKCS12 and PKCS7 routines.
4644 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
4645 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
4646 effectively meant GeneralizedTime would never be used. Now it
4647 is initialised to -1 but X509_time_adj() now has to check the value
4648 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
4649 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
4650 [Steve Henson, reported by Kenneth R. Robinette
4651 <support@securenetterm.com>]
4653 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
4654 result in a zero length in the ASN1_INTEGER structure which was
4655 not consistent with the structure when d2i_ASN1_INTEGER() was used
4656 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
4657 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
4658 where it did not print out a minus for negative ASN1_INTEGER.
4661 *) Add summary printout to ocsp utility. The various functions which
4662 convert status values to strings have been renamed to:
4663 OCSP_response_status_str(), OCSP_cert_status_str() and
4664 OCSP_crl_reason_str() and are no longer static. New options
4665 to verify nonce values and to disable verification. OCSP response
4666 printout format cleaned up.
4669 *) Add additional OCSP certificate checks. These are those specified
4670 in RFC2560. This consists of two separate checks: the CA of the
4671 certificate being checked must either be the OCSP signer certificate
4672 or the issuer of the OCSP signer certificate. In the latter case the
4673 OCSP signer certificate must contain the OCSP signing extended key
4674 usage. This check is performed by attempting to match the OCSP
4675 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
4676 in the OCSP_CERTID structures of the response.
4679 *) Initial OCSP certificate verification added to OCSP_basic_verify()
4680 and related routines. This uses the standard OpenSSL certificate
4681 verify routines to perform initial checks (just CA validity) and
4682 to obtain the certificate chain. Then additional checks will be
4683 performed on the chain. Currently the root CA is checked to see
4684 if it is explicitly trusted for OCSP signing. This is used to set
4685 a root CA as a global signing root: that is any certificate that
4686 chains to that CA is an acceptable OCSP signing certificate.
4689 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
4690 extensions from a separate configuration file.
4691 As when reading extensions from the main configuration file,
4692 the '-extensions ...' option may be used for specifying the
4694 [Massimiliano Pala <madwolf@comune.modena.it>]
4696 *) New OCSP utility. Allows OCSP requests to be generated or
4697 read. The request can be sent to a responder and the output
4698 parsed, outputed or printed in text form. Not complete yet:
4699 still needs to check the OCSP response validity.
4702 *) New subcommands for 'openssl ca':
4703 'openssl ca -status <serial>' prints the status of the cert with
4704 the given serial number (according to the index file).
4705 'openssl ca -updatedb' updates the expiry status of certificates
4707 [Massimiliano Pala <madwolf@comune.modena.it>]
4709 *) New '-newreq-nodes' command option to CA.pl. This is like
4710 '-newreq', but calls 'openssl req' with the '-nodes' option
4711 so that the resulting key is not encrypted.
4712 [Damien Miller <djm@mindrot.org>]
4714 *) New configuration for the GNU Hurd.
4715 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
4717 *) Initial code to implement OCSP basic response verify. This
4718 is currently incomplete. Currently just finds the signer's
4719 certificate and verifies the signature on the response.
4722 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
4723 value of OPENSSLDIR. This is available via the new '-d' option
4724 to 'openssl version', and is also included in 'openssl version -a'.
4727 *) Allowing defining memory allocation callbacks that will be given
4728 file name and line number information in additional arguments
4729 (a const char* and an int). The basic functionality remains, as
4730 well as the original possibility to just replace malloc(),
4731 realloc() and free() by functions that do not know about these
4732 additional arguments. To register and find out the current
4733 settings for extended allocation functions, the following
4734 functions are provided:
4736 CRYPTO_set_mem_ex_functions
4737 CRYPTO_set_locked_mem_ex_functions
4738 CRYPTO_get_mem_ex_functions
4739 CRYPTO_get_locked_mem_ex_functions
4741 These work the same way as CRYPTO_set_mem_functions and friends.
4742 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
4743 extended allocation function is enabled.
4744 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
4745 a conventional allocation function is enabled.
4746 [Richard Levitte, Bodo Moeller]
4748 *) Finish off removing the remaining LHASH function pointer casts.
4749 There should no longer be any prototype-casting required when using
4750 the LHASH abstraction, and any casts that remain are "bugs". See
4751 the callback types and macros at the head of lhash.h for details
4752 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
4755 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
4756 If /dev/[u]random devices are not available or do not return enough
4757 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
4759 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
4760 /etc/entropy will be queried once each in this sequence, quering stops
4761 when enough entropy was collected without querying more sockets.
4764 *) Change the Unix RAND_poll() variant to be able to poll several
4765 random devices, as specified by DEVRANDOM, until a sufficient amount
4766 of data has been collected. We spend at most 10 ms on each file
4767 (select timeout) and read in non-blocking mode. DEVRANDOM now
4768 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
4769 (previously it was just the string "/dev/urandom"), so on typical
4770 platforms the 10 ms delay will never occur.
4771 Also separate out the Unix variant to its own file, rand_unix.c.
4772 For VMS, there's a currently-empty rand_vms.c.
4775 *) Move OCSP client related routines to ocsp_cl.c. These
4776 provide utility functions which an application needing
4777 to issue a request to an OCSP responder and analyse the
4778 response will typically need: as opposed to those which an
4779 OCSP responder itself would need which will be added later.
4781 OCSP_request_sign() signs an OCSP request with an API similar
4782 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
4783 response. OCSP_response_get1_basic() extracts basic response
4784 from response. OCSP_resp_find_status(): finds and extracts status
4785 information from an OCSP_CERTID structure (which will be created
4786 when the request structure is built). These are built from lower
4787 level functions which work on OCSP_SINGLERESP structures but
4788 wont normally be used unless the application wishes to examine
4789 extensions in the OCSP response for example.
4791 Replace nonce routines with a pair of functions.
4792 OCSP_request_add1_nonce() adds a nonce value and optionally
4793 generates a random value. OCSP_check_nonce() checks the
4794 validity of the nonce in an OCSP response.
4797 *) Change function OCSP_request_add() to OCSP_request_add0_id().
4798 This doesn't copy the supplied OCSP_CERTID and avoids the
4799 need to free up the newly created id. Change return type
4800 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
4801 This can then be used to add extensions to the request.
4802 Deleted OCSP_request_new(), since most of its functionality
4803 is now in OCSP_REQUEST_new() (and the case insensitive name
4804 clash) apart from the ability to set the request name which
4805 will be added elsewhere.
4808 *) Update OCSP API. Remove obsolete extensions argument from
4809 various functions. Extensions are now handled using the new
4810 OCSP extension code. New simple OCSP HTTP function which
4811 can be used to send requests and parse the response.
4814 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
4815 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
4816 uses the special reorder version of SET OF to sort the attributes
4817 and reorder them to match the encoded order. This resolves a long
4818 standing problem: a verify on a PKCS7 structure just after signing
4819 it used to fail because the attribute order did not match the
4820 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
4821 it uses the received order. This is necessary to tolerate some broken
4822 software that does not order SET OF. This is handled by encoding
4823 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
4824 to produce the required SET OF.
4827 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
4828 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
4829 files to get correct declarations of the ASN.1 item variables.
4832 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
4833 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
4834 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
4835 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
4836 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
4837 ASN1_ITEM and no wrapper functions.
4840 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
4841 replace the old function pointer based I/O routines. Change most of
4842 the *_d2i_bio() and *_d2i_fp() functions to use these.
4845 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
4846 lines, recognice more "algorithms" that can be deselected, and make
4847 it complain about algorithm deselection that isn't recognised.
4850 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
4851 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
4852 to use new functions. Add NO_ASN1_OLD which can be set to remove
4853 some old style ASN1 functions: this can be used to determine if old
4854 code will still work when these eventually go away.
4857 *) New extension functions for OCSP structures, these follow the
4858 same conventions as certificates and CRLs.
4861 *) New function X509V3_add1_i2d(). This automatically encodes and
4862 adds an extension. Its behaviour can be customised with various
4863 flags to append, replace or delete. Various wrappers added for
4864 certifcates and CRLs.
4867 *) Fix to avoid calling the underlying ASN1 print routine when
4868 an extension cannot be parsed. Correct a typo in the
4869 OCSP_SERVICELOC extension. Tidy up print OCSP format.
4872 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
4873 entries for variables.
4876 *) Add functionality to apps/openssl.c for detecting locking
4877 problems: As the program is single-threaded, all we have
4878 to do is register a locking callback using an array for
4879 storing which locks are currently held by the program.
4882 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
4883 SSL_get_ex_data_X509_STORE_idx(), which is used in
4884 ssl_verify_cert_chain() and thus can be called at any time
4885 during TLS/SSL handshakes so that thread-safety is essential.
4886 Unfortunately, the ex_data design is not at all suited
4887 for multi-threaded use, so it probably should be abolished.
4890 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
4891 [Broadcom, tweaked and integrated by Geoff Thorpe]
4893 *) Move common extension printing code to new function
4894 X509V3_print_extensions(). Reorganise OCSP print routines and
4895 implement some needed OCSP ASN1 functions. Add OCSP extensions.
4898 *) New function X509_signature_print() to remove duplication in some
4902 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
4903 set (this was treated exactly the same as SET OF previously). This
4904 is used to reorder the STACK representing the structure to match the
4905 encoding. This will be used to get round a problem where a PKCS7
4906 structure which was signed could not be verified because the STACK
4907 order did not reflect the encoded order.
4910 *) Reimplement the OCSP ASN1 module using the new code.
4913 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
4914 for its ASN1 operations. The old style function pointers still exist
4915 for now but they will eventually go away.
4918 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
4919 completely replaces the old ASN1 functionality with a table driven
4920 encoder and decoder which interprets an ASN1_ITEM structure describing
4921 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
4922 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
4923 has also been converted to the new form.
4926 *) Change BN_mod_exp_recp so that negative moduli are tolerated
4927 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
4928 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
4929 for negative moduli.
4932 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
4933 of not touching the result's sign bit.
4936 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
4940 *) Changed the LHASH code to use prototypes for callbacks, and created
4941 macros to declare and implement thin (optionally static) functions
4942 that provide type-safety and avoid function pointer casting for the
4943 type-specific callbacks.
4946 *) Added Kerberos Cipher Suites to be used with TLS, as written in
4948 [Veers Staats <staatsvr@asc.hpc.mil>,
4949 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
4951 *) Reformat the FAQ so the different questions and answers can be divided
4952 in sections depending on the subject.
4955 *) Have the zlib compression code load ZLIB.DLL dynamically under
4959 *) New function BN_mod_sqrt for computing square roots modulo a prime
4960 (using the probabilistic Tonelli-Shanks algorithm unless
4961 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
4962 be handled deterministically).
4963 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
4965 *) Make BN_mod_inverse faster by explicitly handling small quotients
4966 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
4967 512 bits], about 30% for larger ones [1024 or 2048 bits].)
4970 *) New function BN_kronecker.
4973 *) Fix BN_gcd so that it works on negative inputs; the result is
4974 positive unless both parameters are zero.
4975 Previously something reasonably close to an infinite loop was
4976 possible because numbers could be growing instead of shrinking
4977 in the implementation of Euclid's algorithm.
4980 *) Fix BN_is_word() and BN_is_one() macros to take into account the
4981 sign of the number in question.
4983 Fix BN_is_word(a,w) to work correctly for w == 0.
4985 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
4986 because its test if the absolute value of 'a' equals 'w'.
4987 Note that BN_abs_is_word does *not* handle w == 0 reliably;
4988 it exists mostly for use in the implementations of BN_is_zero(),
4989 BN_is_one(), and BN_is_word().
4992 *) New function BN_swap.
4995 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
4996 the exponentiation functions are more likely to produce reasonable
4997 results on negative inputs.
5000 *) Change BN_mod_mul so that the result is always non-negative.
5001 Previously, it could be negative if one of the factors was negative;
5002 I don't think anyone really wanted that behaviour.
5005 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5006 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5007 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5008 and add new functions:
5017 BN_mod_lshift1_quick
5021 These functions always generate non-negative results.
5023 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5024 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5026 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5027 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5028 be reduced modulo m.
5029 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5032 The following entry accidentily appeared in the CHANGES file
5033 distributed with OpenSSL 0.9.7. The modifications described in
5034 it do *not* apply to OpenSSL 0.9.7.
5036 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5037 was actually never needed) and in BN_mul(). The removal in BN_mul()
5038 required a small change in bn_mul_part_recursive() and the addition
5039 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5040 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5041 bn_sub_words() and bn_add_words() except they take arrays with
5046 *) In 'openssl passwd', verify passwords read from the terminal
5047 unless the '-salt' option is used (which usually means that
5048 verification would just waste user's time since the resulting
5049 hash is going to be compared with some given password hash)
5050 or the new '-noverify' option is used.
5052 This is an incompatible change, but it does not affect
5053 non-interactive use of 'openssl passwd' (passwords on the command
5054 line, '-stdin' option, '-in ...' option) and thus should not
5058 *) Remove all references to RSAref, since there's no more need for it.
5061 *) Make DSO load along a path given through an environment variable
5062 (SHLIB_PATH) with shl_load().
5065 *) Constify the ENGINE code as a result of BIGNUM constification.
5066 Also constify the RSA code and most things related to it. In a
5067 few places, most notable in the depth of the ASN.1 code, ugly
5068 casts back to non-const were required (to be solved at a later
5072 *) Make it so the openssl application has all engines loaded by default.
5075 *) Constify the BIGNUM routines a little more.
5078 *) Add the following functions:
5080 ENGINE_load_cswift()
5082 ENGINE_load_atalla()
5084 ENGINE_load_builtin_engines()
5086 That way, an application can itself choose if external engines that
5087 are built-in in OpenSSL shall ever be used or not. The benefit is
5088 that applications won't have to be linked with libdl or other dso
5089 libraries unless it's really needed.
5091 Changed 'openssl engine' to load all engines on demand.
5092 Changed the engine header files to avoid the duplication of some
5093 declarations (they differed!).
5096 *) 'openssl engine' can now list capabilities.
5099 *) Better error reporting in 'openssl engine'.
5102 *) Never call load_dh_param(NULL) in s_server.
5105 *) Add engine application. It can currently list engines by name and
5106 identity, and test if they are actually available.
5109 *) Improve RPM specification file by forcing symbolic linking and making
5110 sure the installed documentation is also owned by root.root.
5111 [Damien Miller <djm@mindrot.org>]
5113 *) Give the OpenSSL applications more possibilities to make use of
5114 keys (public as well as private) handled by engines.
5117 *) Add OCSP code that comes from CertCo.
5120 *) Add VMS support for the Rijndael code.
5123 *) Added untested support for Nuron crypto accelerator.
5126 *) Add support for external cryptographic devices. This code was
5127 previously distributed separately as the "engine" branch.
5128 [Geoff Thorpe, Richard Levitte]
5130 *) Rework the filename-translation in the DSO code. It is now possible to
5131 have far greater control over how a "name" is turned into a filename
5132 depending on the operating environment and any oddities about the
5133 different shared library filenames on each system.
5136 *) Support threads on FreeBSD-elf in Configure.
5139 *) Fix for SHA1 assembly problem with MASM: it produces
5140 warnings about corrupt line number information when assembling
5141 with debugging information. This is caused by the overlapping
5143 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5146 NCONF_get_number() has no error checking at all. As a replacement,
5147 NCONF_get_number_e() is defined (_e for "error checking") and is
5148 promoted strongly. The old NCONF_get_number is kept around for
5149 binary backward compatibility.
5150 Make it possible for methods to load from something other than a BIO,
5151 by providing a function pointer that is given a name instead of a BIO.
5152 For example, this could be used to load configuration data from an
5156 *) Fix for non blocking accept BIOs. Added new I/O special reason
5157 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5158 with non blocking I/O was not possible because no retry code was
5159 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5163 *) Added the beginnings of Rijndael support.
5166 *) Fix for bug in DirectoryString mask setting. Add support for
5167 X509_NAME_print_ex() in 'req' and X509_print_ex() function
5168 to allow certificate printing to more controllable, additional
5169 'certopt' option to 'x509' to allow new printing options to be
5173 *) Clean old EAY MD5 hack from e_os.h.
5176 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5178 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5179 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5180 [Joe Orton, Steve Henson]
5182 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5184 *) Fix additional bug revealed by the NISCC test suite:
5186 Stop bug triggering large recursion when presented with
5187 certain ASN.1 tags (CVE-2003-0851)
5190 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5192 *) Fix various bugs revealed by running the NISCC test suite:
5194 Stop out of bounds reads in the ASN1 code when presented with
5195 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5197 If verify callback ignores invalid public key errors don't try to check
5198 certificate signature with the NULL public key.
5202 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5203 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5207 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5208 extra data after the compression methods not only for TLS 1.0
5209 but also for SSL 3.0 (as required by the specification).
5210 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5212 *) Change X509_certificate_type() to mark the key as exported/exportable
5213 when it's 512 *bits* long, not 512 bytes.
5216 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5218 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5219 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5220 a protocol version number mismatch like a decryption error
5221 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5224 *) Turn on RSA blinding by default in the default implementation
5225 to avoid a timing attack. Applications that don't want it can call
5226 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5227 They would be ill-advised to do so in most cases.
5228 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5230 *) Change RSA blinding code so that it works when the PRNG is not
5231 seeded (in this case, the secret RSA exponent is abused as
5232 an unpredictable seed -- if it is not unpredictable, there
5233 is no point in blinding anyway). Make RSA blinding thread-safe
5234 by remembering the creator's thread ID in rsa->blinding and
5235 having all other threads use local one-time blinding factors
5236 (this requires more computation than sharing rsa->blinding, but
5237 avoids excessive locking; and if an RSA object is not shared
5238 between threads, blinding will still be very fast).
5241 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5243 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5244 via timing by performing a MAC computation even if incorrrect
5245 block cipher padding has been found. This is a countermeasure
5246 against active attacks where the attacker has to distinguish
5247 between bad padding and a MAC verification error. (CVE-2003-0078)
5249 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5250 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5251 Martin Vuagnoux (EPFL, Ilion)]
5253 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
5255 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
5256 memory from it's contents. This is done with a counter that will
5257 place alternating values in each byte. This can be used to solve
5258 two issues: 1) the removal of calls to memset() by highly optimizing
5259 compilers, and 2) cleansing with other values than 0, since those can
5260 be read through on certain media, for example a swap space on disk.
5263 *) Bugfix: client side session caching did not work with external caching,
5264 because the session->cipher setting was not restored when reloading
5265 from the external cache. This problem was masked, when
5266 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
5267 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
5270 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
5271 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
5272 [Zeev Lieber <zeev-l@yahoo.com>]
5274 *) Undo an undocumented change introduced in 0.9.6e which caused
5275 repeated calls to OpenSSL_add_all_ciphers() and
5276 OpenSSL_add_all_digests() to be ignored, even after calling
5280 *) Change the default configuration reader to deal with last line not
5281 being properly terminated.
5284 *) Change X509_NAME_cmp() so it applies the special rules on handling
5285 DN values that are of type PrintableString, as well as RDNs of type
5286 emailAddress where the value has the type ia5String.
5287 [stefank@valicert.com via Richard Levitte]
5289 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
5290 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
5291 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
5292 the bitwise-OR of the two for use by the majority of applications
5293 wanting this behaviour, and update the docs. The documented
5294 behaviour and actual behaviour were inconsistent and had been
5295 changing anyway, so this is more a bug-fix than a behavioural
5297 [Geoff Thorpe, diagnosed by Nadav Har'El]
5299 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
5300 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
5303 *) Fix initialization code race conditions in
5304 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
5305 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
5306 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
5307 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
5308 ssl2_get_cipher_by_char(),
5309 ssl3_get_cipher_by_char().
5310 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
5312 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
5313 the cached sessions are flushed, as the remove_cb() might use ex_data
5314 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
5315 (see [openssl.org #212]).
5316 [Geoff Thorpe, Lutz Jaenicke]
5318 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
5319 length, instead of the encoding length to d2i_ASN1_OBJECT.
5322 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
5324 *) [In 0.9.6g-engine release:]
5325 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
5326 [Lynn Gazis <lgazis@rainbow.com>]
5328 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
5330 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
5331 and get fix the header length calculation.
5332 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
5333 Alon Kantor <alonk@checkpoint.com> (and others),
5336 *) Use proper error handling instead of 'assertions' in buffer
5337 overflow checks added in 0.9.6e. This prevents DoS (the
5338 assertions could call abort()).
5339 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
5341 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
5343 *) Add various sanity checks to asn1_get_length() to reject
5344 the ASN1 length bytes if they exceed sizeof(long), will appear
5345 negative or the content length exceeds the length of the
5347 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5349 *) Fix cipher selection routines: ciphers without encryption had no flags
5350 for the cipher strength set and where therefore not handled correctly
5351 by the selection routines (PR #130).
5354 *) Fix EVP_dsa_sha macro.
5358 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
5359 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
5360 that was added in OpenSSL 0.9.6d.
5362 As the countermeasure turned out to be incompatible with some
5363 broken SSL implementations, the new option is part of SSL_OP_ALL.
5364 SSL_OP_ALL is usually employed when compatibility with weird SSL
5365 implementations is desired (e.g. '-bugs' option to 's_client' and
5366 's_server'), so the new option is automatically set in many
5370 *) Changes in security patch:
5372 Changes marked "(CHATS)" were sponsored by the Defense Advanced
5373 Research Projects Agency (DARPA) and Air Force Research Laboratory,
5374 Air Force Materiel Command, USAF, under agreement number
5377 *) Add various sanity checks to asn1_get_length() to reject
5378 the ASN1 length bytes if they exceed sizeof(long), will appear
5379 negative or the content length exceeds the length of the
5380 supplied buffer. (CVE-2002-0659)
5381 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5383 *) Assertions for various potential buffer overflows, not known to
5385 [Ben Laurie (CHATS)]
5387 *) Various temporary buffers to hold ASCII versions of integers were
5388 too small for 64 bit platforms. (CVE-2002-0655)
5389 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
5391 *) Remote buffer overflow in SSL3 protocol - an attacker could
5392 supply an oversized session ID to a client. (CVE-2002-0656)
5393 [Ben Laurie (CHATS)]
5395 *) Remote buffer overflow in SSL2 protocol - an attacker could
5396 supply an oversized client master key. (CVE-2002-0656)
5397 [Ben Laurie (CHATS)]
5399 Changes between 0.9.6c and 0.9.6d [9 May 2002]
5401 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
5402 encoded as NULL) with id-dsa-with-sha1.
5403 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
5405 *) Check various X509_...() return values in apps/req.c.
5406 [Nils Larsch <nla@trustcenter.de>]
5408 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
5409 an end-of-file condition would erronously be flagged, when the CRLF
5410 was just at the end of a processed block. The bug was discovered when
5411 processing data through a buffering memory BIO handing the data to a
5412 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
5413 <ptsekov@syntrex.com> and Nedelcho Stanev.
5416 *) Implement a countermeasure against a vulnerability recently found
5417 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
5418 before application data chunks to avoid the use of known IVs
5419 with data potentially chosen by the attacker.
5422 *) Fix length checks in ssl3_get_client_hello().
5425 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
5426 to prevent ssl3_read_internal() from incorrectly assuming that
5427 ssl3_read_bytes() found application data while handshake
5428 processing was enabled when in fact s->s3->in_read_app_data was
5429 merely automatically cleared during the initial handshake.
5430 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
5432 *) Fix object definitions for Private and Enterprise: they were not
5433 recognized in their shortname (=lowercase) representation. Extend
5434 obj_dat.pl to issue an error when using undefined keywords instead
5435 of silently ignoring the problem (Svenning Sorensen
5436 <sss@sss.dnsalias.net>).
5439 *) Fix DH_generate_parameters() so that it works for 'non-standard'
5440 generators, i.e. generators other than 2 and 5. (Previously, the
5441 code did not properly initialise the 'add' and 'rem' values to
5442 BN_generate_prime().)
5444 In the new general case, we do not insist that 'generator' is
5445 actually a primitive root: This requirement is rather pointless;
5446 a generator of the order-q subgroup is just as good, if not
5450 *) Map new X509 verification errors to alerts. Discovered and submitted by
5451 Tom Wu <tom@arcot.com>.
5454 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
5455 returning non-zero before the data has been completely received
5456 when using non-blocking I/O.
5457 [Bodo Moeller; problem pointed out by John Hughes]
5459 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
5460 [Ben Laurie, Lutz Jaenicke]
5462 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
5463 Yoram Zahavi <YoramZ@gilian.com>).
5466 *) Add information about CygWin 1.3 and on, and preserve proper
5467 configuration for the versions before that.
5468 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5470 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
5471 check whether we deal with a copy of a session and do not delete from
5472 the cache in this case. Problem reported by "Izhar Shoshani Levi"
5473 <izhar@checkpoint.com>.
5476 *) Do not store session data into the internal session cache, if it
5477 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
5478 flag is set). Proposed by Aslam <aslam@funk.com>.
5481 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
5485 *) [In 0.9.6d-engine release:]
5486 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5487 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5489 *) Add the configuration target linux-s390x.
5490 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
5492 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
5493 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
5494 variable as an indication that a ClientHello message has been
5495 received. As the flag value will be lost between multiple
5496 invocations of ssl3_accept when using non-blocking I/O, the
5497 function may not be aware that a handshake has actually taken
5498 place, thus preventing a new session from being added to the
5501 To avoid this problem, we now set s->new_session to 2 instead of
5502 using a local variable.
5503 [Lutz Jaenicke, Bodo Moeller]
5505 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
5506 if the SSL_R_LENGTH_MISMATCH error is detected.
5507 [Geoff Thorpe, Bodo Moeller]
5509 *) New 'shared_ldflag' column in Configure platform table.
5512 *) Fix EVP_CIPHER_mode macro.
5513 ["Dan S. Camper" <dan@bti.net>]
5515 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
5516 type, we must throw them away by setting rr->length to 0.
5517 [D P Chang <dpc@qualys.com>]
5519 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
5521 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
5522 <Dominikus.Scherkl@biodata.com>. (The previous implementation
5523 worked incorrectly for those cases where range = 10..._2 and
5524 3*range is two bits longer than range.)
5527 *) Only add signing time to PKCS7 structures if it is not already
5531 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
5532 OBJ_ld_ce should be OBJ_id_ce.
5533 Also some ip-pda OIDs in crypto/objects/objects.txt were
5534 incorrect (cf. RFC 3039).
5535 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
5537 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
5538 returns early because it has nothing to do.
5539 [Andy Schneider <andy.schneider@bjss.co.uk>]
5541 *) [In 0.9.6c-engine release:]
5542 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
5543 [Andy Schneider <andy.schneider@bjss.co.uk>]
5545 *) [In 0.9.6c-engine release:]
5546 Add support for Cryptographic Appliance's keyserver technology.
5547 (Use engine 'keyclient')
5548 [Cryptographic Appliances and Geoff Thorpe]
5550 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
5551 is called via tools/c89.sh because arguments have to be
5552 rearranged (all '-L' options must appear before the first object
5554 [Richard Shapiro <rshapiro@abinitio.com>]
5556 *) [In 0.9.6c-engine release:]
5557 Add support for Broadcom crypto accelerator cards, backported
5559 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
5561 *) [In 0.9.6c-engine release:]
5562 Add support for SureWare crypto accelerator cards from
5563 Baltimore Technologies. (Use engine 'sureware')
5564 [Baltimore Technologies and Mark Cox]
5566 *) [In 0.9.6c-engine release:]
5567 Add support for crypto accelerator cards from Accelerated
5568 Encryption Processing, www.aep.ie. (Use engine 'aep')
5569 [AEP Inc. and Mark Cox]
5571 *) Add a configuration entry for gcc on UnixWare.
5572 [Gary Benson <gbenson@redhat.com>]
5574 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
5575 messages are stored in a single piece (fixed-length part and
5576 variable-length part combined) and fix various bugs found on the way.
5579 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
5580 instead. BIO_gethostbyname() does not know what timeouts are
5581 appropriate, so entries would stay in cache even when they have
5583 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
5585 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
5586 faced with a pathologically small ClientHello fragment that does
5587 not contain client_version: Instead of aborting with an error,
5588 simply choose the highest available protocol version (i.e.,
5589 TLS 1.0 unless it is disabled). In practice, ClientHello
5590 messages are never sent like this, but this change gives us
5591 strictly correct behaviour at least for TLS.
5594 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
5595 never resets s->method to s->ctx->method when called from within
5596 one of the SSL handshake functions.
5597 [Bodo Moeller; problem pointed out by Niko Baric]
5599 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
5600 (sent using the client's version number) if client_version is
5601 smaller than the protocol version in use. Also change
5602 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
5603 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
5604 the client will at least see that alert.
5607 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
5611 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
5612 client receives HelloRequest while in a handshake.
5613 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
5615 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
5616 should end in 'break', not 'goto end' which circuments various
5617 cleanups done in state SSL_ST_OK. But session related stuff
5618 must be disabled for SSL_ST_OK in the case that we just sent a
5621 Also avoid some overhead by not calling ssl_init_wbio_buffer()
5622 before just sending a HelloRequest.
5623 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
5625 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
5626 reveal whether illegal block cipher padding was found or a MAC
5627 verification error occured. (Neither SSLerr() codes nor alerts
5628 are directly visible to potential attackers, but the information
5629 may leak via logfiles.)
5631 Similar changes are not required for the SSL 2.0 implementation
5632 because the number of padding bytes is sent in clear for SSL 2.0,
5633 and the extra bytes are just ignored. However ssl/s2_pkt.c
5634 failed to verify that the purported number of padding bytes is in
5638 *) Add OpenUNIX-8 support including shared libraries
5639 (Boyd Lynn Gerber <gerberb@zenez.com>).
5642 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
5643 'wristwatch attack' using huge encoding parameters (cf.
5644 James H. Manger's CRYPTO 2001 paper). Note that the
5645 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
5646 encoding parameters and hence was not vulnerable.
5649 *) BN_sqr() bug fix.
5650 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
5652 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
5653 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
5654 followed by modular reduction.
5655 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
5657 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
5658 equivalent based on BN_pseudo_rand() instead of BN_rand().
5661 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
5662 This function was broken, as the check for a new client hello message
5663 to handle SGC did not allow these large messages.
5664 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
5667 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
5670 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
5671 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
5674 *) Rework the configuration and shared library support for Tru64 Unix.
5675 The configuration part makes use of modern compiler features and
5676 still retains old compiler behavior for those that run older versions
5677 of the OS. The shared library support part includes a variant that
5678 uses the RPATH feature, and is available through the special
5679 configuration target "alpha-cc-rpath", which will never be selected
5681 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
5683 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
5684 with the same message size as in ssl3_get_certificate_request().
5685 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
5686 messages might inadvertently be reject as too long.
5687 [Petr Lampa <lampa@fee.vutbr.cz>]
5689 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
5692 *) Modified SSL library such that the verify_callback that has been set
5693 specificly for an SSL object with SSL_set_verify() is actually being
5694 used. Before the change, a verify_callback set with this function was
5695 ignored and the verify_callback() set in the SSL_CTX at the time of
5696 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
5697 to allow the necessary settings.
5700 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
5701 explicitly to NULL, as at least on Solaris 8 this seems not always to be
5702 done automatically (in contradiction to the requirements of the C
5703 standard). This made problems when used from OpenSSH.
5706 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
5707 dh->length and always used
5709 BN_rand_range(priv_key, dh->p).
5711 BN_rand_range() is not necessary for Diffie-Hellman, and this
5712 specific range makes Diffie-Hellman unnecessarily inefficient if
5713 dh->length (recommended exponent length) is much smaller than the
5714 length of dh->p. We could use BN_rand_range() if the order of
5715 the subgroup was stored in the DH structure, but we only have
5720 BN_rand(priv_key, l, ...)
5722 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
5728 RSA_eay_public_encrypt
5729 RSA_eay_private_decrypt
5730 RSA_eay_private_encrypt (signing)
5731 RSA_eay_public_decrypt (signature verification)
5733 (default implementations for RSA_public_encrypt,
5734 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
5735 always reject numbers >= n.
5738 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
5739 to synchronize access to 'locking_thread'. This is necessary on
5740 systems where access to 'locking_thread' (an 'unsigned long'
5741 variable) is not atomic.
5744 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
5745 *before* setting the 'crypto_lock_rand' flag. The previous code had
5746 a race condition if 0 is a valid thread ID.
5747 [Travis Vitek <vitek@roguewave.com>]
5749 *) Add support for shared libraries under Irix.
5750 [Albert Chin-A-Young <china@thewrittenword.com>]
5752 *) Add configuration option to build on Linux on both big-endian and
5754 [Ralf Baechle <ralf@uni-koblenz.de>]
5756 *) Add the possibility to create shared libraries on HP-UX.
5759 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
5761 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
5762 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
5763 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
5764 PRNG state recovery was possible based on the output of
5765 one PRNG request appropriately sized to gain knowledge on
5766 'md' followed by enough consecutive 1-byte PRNG requests
5767 to traverse all of 'state'.
5769 1. When updating 'md_local' (the current thread's copy of 'md')
5770 during PRNG output generation, hash all of the previous
5771 'md_local' value, not just the half used for PRNG output.
5773 2. Make the number of bytes from 'state' included into the hash
5774 independent from the number of PRNG bytes requested.
5776 The first measure alone would be sufficient to avoid
5777 Markku-Juhani's attack. (Actually it had never occurred
5778 to me that the half of 'md_local' used for chaining was the
5779 half from which PRNG output bytes were taken -- I had always
5780 assumed that the secret half would be used.) The second
5781 measure makes sure that additional data from 'state' is never
5782 mixed into 'md_local' in small portions; this heuristically
5783 further strengthens the PRNG.
5786 *) Fix crypto/bn/asm/mips3.s.
5789 *) When only the key is given to "enc", the IV is undefined. Print out
5790 an error message in this case.
5793 *) Handle special case when X509_NAME is empty in X509 printing routines.
5796 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
5797 positive and less than q.
5800 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
5801 used: it isn't thread safe and the add_lock_callback should handle
5803 [Paul Rose <Paul.Rose@bridge.com>]
5805 *) Verify that incoming data obeys the block size in
5806 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
5810 [Ulf Möller, Bodo Möller]
5812 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
5813 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
5814 when fixing the server behaviour for backwards-compatible 'client
5815 hello' messages. (Note that the attack is impractical against
5816 SSL 3.0 and TLS 1.0 anyway because length and version checking
5817 means that the probability of guessing a valid ciphertext is
5818 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
5821 Before 0.9.5, the countermeasure (hide the error by generating a
5822 random 'decryption result') did not work properly because
5823 ERR_clear_error() was missing, meaning that SSL_get_error() would
5824 detect the supposedly ignored error.
5826 Both problems are now fixed.
5829 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
5830 (previously it was 1024).
5833 *) Fix for compatibility mode trust settings: ignore trust settings
5834 unless some valid trust or reject settings are present.
5837 *) Fix for blowfish EVP: its a variable length cipher.
5840 *) Fix various bugs related to DSA S/MIME verification. Handle missing
5841 parameters in DSA public key structures and return an error in the
5842 DSA routines if parameters are absent.
5845 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
5846 in the current directory if neither $RANDFILE nor $HOME was set.
5847 RAND_file_name() in 0.9.6a returned NULL in this case. This has
5848 caused some confusion to Windows users who haven't defined $HOME.
5849 Thus RAND_file_name() is changed again: e_os.h can define a
5850 DEFAULT_HOME, which will be used if $HOME is not set.
5851 For Windows, we use "C:"; on other platforms, we still require
5852 environment variables.
5854 *) Move 'if (!initialized) RAND_poll()' into regions protected by
5855 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
5856 having multiple threads call RAND_poll() concurrently.
5859 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
5860 combination of a flag and a thread ID variable.
5861 Otherwise while one thread is in ssleay_rand_bytes (which sets the
5862 flag), *other* threads can enter ssleay_add_bytes without obeying
5863 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
5864 that they do not hold after the first thread unsets add_do_not_lock).
5867 *) Change bctest again: '-x' expressions are not available in all
5871 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
5873 *) Fix a couple of memory leaks in PKCS7_dataDecode()
5874 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
5876 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
5877 the default extension for executables, if any. Also, make the perl
5878 scripts that use symlink() to test if it really exists and use "cp"
5879 if it doesn't. All this made OpenSSL compilable and installable in
5883 *) Fix for asn1_GetSequence() for indefinite length constructed data.
5884 If SEQUENCE is length is indefinite just set c->slen to the total
5885 amount of data available.
5886 [Steve Henson, reported by shige@FreeBSD.org]
5887 [This change does not apply to 0.9.7.]
5889 *) Change bctest to avoid here-documents inside command substitution
5890 (workaround for FreeBSD /bin/sh bug).
5891 For compatibility with Ultrix, avoid shell functions (introduced
5892 in the bctest version that searches along $PATH).
5895 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
5896 with des_encrypt() defined on some operating systems, like Solaris
5900 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
5901 On the Importance of Eliminating Errors in Cryptographic
5902 Computations, J. Cryptology 14 (2001) 2, 101-119,
5903 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
5906 *) MIPS assembler BIGNUM division bug fix.
5909 *) Disabled incorrect Alpha assembler code.
5912 *) Fix PKCS#7 decode routines so they correctly update the length
5913 after reading an EOC for the EXPLICIT tag.
5915 [This change does not apply to 0.9.7.]
5917 *) Fix bug in PKCS#12 key generation routines. This was triggered
5918 if a 3DES key was generated with a 0 initial byte. Include
5919 PKCS12_BROKEN_KEYGEN compilation option to retain the old
5920 (but broken) behaviour.
5923 *) Enhance bctest to search for a working bc along $PATH and print
5925 [Tim Rice <tim@multitalents.net> via Richard Levitte]
5927 *) Fix memory leaks in err.c: free err_data string if necessary;
5928 don't write to the wrong index in ERR_set_error_data.
5931 *) Implement ssl23_peek (analogous to ssl23_read), which previously
5935 *) Replace rdtsc with _emit statements for VC++ version 5.
5936 [Jeremy Cooper <jeremy@baymoo.org>]
5938 *) Make it possible to reuse SSLv2 sessions.
5941 *) In copy_email() check for >= 0 as a return value for
5942 X509_NAME_get_index_by_NID() since 0 is a valid index.
5943 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
5945 *) Avoid coredump with unsupported or invalid public keys by checking if
5946 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
5947 PKCS7_verify() fails with non detached data.
5950 *) Don't use getenv in library functions when run as setuid/setgid.
5951 New function OPENSSL_issetugid().
5954 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
5955 due to incorrect handling of multi-threading:
5957 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
5959 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
5961 3. Count how many times MemCheck_off() has been called so that
5962 nested use can be treated correctly. This also avoids
5963 inband-signalling in the previous code (which relied on the
5964 assumption that thread ID 0 is impossible).
5967 *) Add "-rand" option also to s_client and s_server.
5970 *) Fix CPU detection on Irix 6.x.
5971 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
5972 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
5974 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
5977 [This change does not apply to 0.9.7.]
5979 *) Use the cached encoding of an X509_NAME structure rather than
5980 copying it. This is apparently the reason for the libsafe "errors"
5981 but the code is actually correct.
5984 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
5985 Bleichenbacher's DSA attack.
5986 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
5987 to be set and top=0 forces the highest bit to be set; top=-1 is new
5988 and leaves the highest bit random.
5989 [Ulf Moeller, Bodo Moeller]
5991 *) In the NCONF_...-based implementations for CONF_... queries
5992 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
5993 a temporary CONF structure with the data component set to NULL
5994 (which gives segmentation faults in lh_retrieve).
5995 Instead, use NULL for the CONF pointer in CONF_get_string and
5996 CONF_get_number (which may use environment variables) and directly
5997 return NULL from CONF_get_section.
6000 *) Fix potential buffer overrun for EBCDIC.
6003 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6004 keyUsage if basicConstraints absent for a CA.
6007 *) Make SMIME_write_PKCS7() write mail header values with a format that
6008 is more generally accepted (no spaces before the semicolon), since
6009 some programs can't parse those values properly otherwise. Also make
6010 sure BIO's that break lines after each write do not create invalid
6014 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6015 macros previously used would not encode an empty SEQUENCE OF
6016 and break the signature.
6018 [This change does not apply to 0.9.7.]
6020 *) Zero the premaster secret after deriving the master secret in
6024 *) Add some EVP_add_digest_alias registrations (as found in
6025 OpenSSL_add_all_digests()) to SSL_library_init()
6026 aka OpenSSL_add_ssl_algorithms(). This provides improved
6027 compatibility with peers using X.509 certificates
6028 with unconventional AlgorithmIdentifier OIDs.
6031 *) Fix for Irix with NO_ASM.
6032 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6034 *) ./config script fixes.
6035 [Ulf Moeller, Richard Levitte]
6037 *) Fix 'openssl passwd -1'.
6040 *) Change PKCS12_key_gen_asc() so it can cope with non null
6041 terminated strings whose length is passed in the passlen
6042 parameter, for example from PEM callbacks. This was done
6043 by adding an extra length parameter to asc2uni().
6044 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6046 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6047 call failed, free the DSA structure.
6050 *) Fix to uni2asc() to cope with zero length Unicode strings.
6051 These are present in some PKCS#12 files.
6054 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6055 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6056 when writing a 32767 byte record.
6057 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6059 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6060 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6062 (RSA objects have a reference count access to which is protected
6063 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6064 so they are meant to be shared between threads.)
6065 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6066 "Reddie, Steven" <Steven.Reddie@ca.com>]
6068 *) Fix a deadlock in CRYPTO_mem_leaks().
6071 *) Use better test patterns in bntest.
6074 *) rand_win.c fix for Borland C.
6077 *) BN_rshift bugfix for n == 0.
6080 *) Add a 'bctest' script that checks for some known 'bc' bugs
6081 so that 'make test' does not abort just because 'bc' is broken.
6084 *) Store verify_result within SSL_SESSION also for client side to
6085 avoid potential security hole. (Re-used sessions on the client side
6086 always resulted in verify_result==X509_V_OK, not using the original
6087 result of the server certificate verification.)
6090 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6091 SSL3_RT_APPLICATION_DATA, return 0.
6092 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6096 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6097 releases, have been re-implemented by renaming the previous
6098 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6099 and ssl3_read_internal, respectively, and adding 'peek' parameters
6100 to them. The new ssl[23]_{read,peek} functions are calls to
6101 ssl[23]_read_internal with the 'peek' flag set appropriately.
6102 A 'peek' parameter has also been added to ssl3_read_bytes, which
6103 does the actual work for ssl3_read_internal.
6106 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6107 the method-specific "init()" handler. Also clean up ex_data after
6108 calling the method-specific "finish()" handler. Previously, this was
6109 happening the other way round.
6112 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6113 The previous value, 12, was not always sufficient for BN_mod_exp().
6116 *) Make sure that shared libraries get the internal name engine with
6117 the full version number and not just 0. This should mark the
6118 shared libraries as not backward compatible. Of course, this should
6119 be changed again when we can guarantee backward binary compatibility.
6122 *) Fix typo in get_cert_by_subject() in by_dir.c
6123 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6125 *) Rework the system to generate shared libraries:
6127 - Make note of the expected extension for the shared libraries and
6128 if there is a need for symbolic links from for example libcrypto.so.0
6129 to libcrypto.so.0.9.7. There is extended info in Configure for
6132 - Make as few rebuilds of the shared libraries as possible.
6134 - Still avoid linking the OpenSSL programs with the shared libraries.
6136 - When installing, install the shared libraries separately from the
6140 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6142 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6143 and not in SSL_clear because the latter is also used by the
6144 accept/connect functions; previously, the settings made by
6145 SSL_set_read_ahead would be lost during the handshake.
6146 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6148 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6149 Previously, it would create entries for disableed algorithms no
6153 *) Added several new manual pages for SSL_* function.
6156 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6158 *) In ssl23_get_client_hello, generate an error message when faced
6159 with an initial SSL 3.0/TLS record that is too small to contain the
6160 first two bytes of the ClientHello message, i.e. client_version.
6161 (Note that this is a pathologic case that probably has never happened
6162 in real life.) The previous approach was to use the version number
6163 from the record header as a substitute; but our protocol choice
6164 should not depend on that one because it is not authenticated
6165 by the Finished messages.
6168 *) More robust randomness gathering functions for Windows.
6169 [Jeffrey Altman <jaltman@columbia.edu>]
6171 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
6172 not set then we don't setup the error code for issuer check errors
6173 to avoid possibly overwriting other errors which the callback does
6174 handle. If an application does set the flag then we assume it knows
6175 what it is doing and can handle the new informational codes
6179 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
6180 a general "ANY" type, as such it should be able to decode anything
6181 including tagged types. However it didn't check the class so it would
6182 wrongly interpret tagged types in the same way as their universal
6183 counterpart and unknown types were just rejected. Changed so that the
6184 tagged and unknown types are handled in the same way as a SEQUENCE:
6185 that is the encoding is stored intact. There is also a new type
6186 "V_ASN1_OTHER" which is used when the class is not universal, in this
6187 case we have no idea what the actual type is so we just lump them all
6191 *) On VMS, stdout may very well lead to a file that is written to
6192 in a record-oriented fashion. That means that every write() will
6193 write a separate record, which will be read separately by the
6194 programs trying to read from it. This can be very confusing.
6196 The solution is to put a BIO filter in the way that will buffer
6197 text until a linefeed is reached, and then write everything a
6198 line at a time, so every record written will be an actual line,
6199 not chunks of lines and not (usually doesn't happen, but I've
6200 seen it once) several lines in one record. BIO_f_linebuffer() is
6203 Currently, it's a VMS-only method, because that's where it has
6204 been tested well enough.
6207 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
6208 it can return incorrect results.
6209 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
6210 but it was in 0.9.6-beta[12].)
6213 *) Disable the check for content being present when verifying detached
6214 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
6215 include zero length content when signing messages.
6218 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
6219 BIO_ctrl (for BIO pairs).
6222 *) Add DSO method for VMS.
6225 *) Bug fix: Montgomery multiplication could produce results with the
6229 *) Add RPM specification openssl.spec and modify it to build three
6230 packages. The default package contains applications, application
6231 documentation and run-time libraries. The devel package contains
6232 include files, static libraries and function documentation. The
6233 doc package contains the contents of the doc directory. The original
6234 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
6237 *) Add a large number of documentation files for many SSL routines.
6238 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6240 *) Add a configuration entry for Sony News 4.
6241 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
6243 *) Don't set the two most significant bits to one when generating a
6244 random number < q in the DSA library.
6247 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
6248 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
6249 the underlying transport is blocking) if a handshake took place.
6250 (The default behaviour is needed by applications such as s_client
6251 and s_server that use select() to determine when to use SSL_read;
6252 but for applications that know in advance when to expect data, it
6253 just makes things more complicated.)
6256 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
6260 *) Add a few more EBCDIC conditionals that make `req' and `x509'
6261 work better on such systems.
6262 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6264 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
6265 Update PKCS12_parse() so it copies the friendlyName and the
6266 keyid to the certificates aux info.
6269 *) Fix bug in PKCS7_verify() which caused an infinite loop
6270 if there was more than one signature.
6271 [Sven Uszpelkat <su@celocom.de>]
6273 *) Major change in util/mkdef.pl to include extra information
6274 about each symbol, as well as presentig variables as well
6275 as functions. This change means that there's n more need
6276 to rebuild the .num files when some algorithms are excluded.
6279 *) Allow the verify time to be set by an application,
6280 rather than always using the current time.
6283 *) Phase 2 verify code reorganisation. The certificate
6284 verify code now looks up an issuer certificate by a
6285 number of criteria: subject name, authority key id
6286 and key usage. It also verifies self signed certificates
6287 by the same criteria. The main comparison function is
6288 X509_check_issued() which performs these checks.
6290 Lot of changes were necessary in order to support this
6291 without completely rewriting the lookup code.
6293 Authority and subject key identifier are now cached.
6295 The LHASH 'certs' is X509_STORE has now been replaced
6296 by a STACK_OF(X509_OBJECT). This is mainly because an
6297 LHASH can't store or retrieve multiple objects with
6298 the same hash value.
6300 As a result various functions (which were all internal
6301 use only) have changed to handle the new X509_STORE
6302 structure. This will break anything that messed round
6303 with X509_STORE internally.
6305 The functions X509_STORE_add_cert() now checks for an
6306 exact match, rather than just subject name.
6308 The X509_STORE API doesn't directly support the retrieval
6309 of multiple certificates matching a given criteria, however
6310 this can be worked round by performing a lookup first
6311 (which will fill the cache with candidate certificates)
6312 and then examining the cache for matches. This is probably
6313 the best we can do without throwing out X509_LOOKUP
6314 entirely (maybe later...).
6316 The X509_VERIFY_CTX structure has been enhanced considerably.
6318 All certificate lookup operations now go via a get_issuer()
6319 callback. Although this currently uses an X509_STORE it
6320 can be replaced by custom lookups. This is a simple way
6321 to bypass the X509_STORE hackery necessary to make this
6322 work and makes it possible to use more efficient techniques
6323 in future. A very simple version which uses a simple
6324 STACK for its trusted certificate store is also provided
6325 using X509_STORE_CTX_trusted_stack().
6327 The verify_cb() and verify() callbacks now have equivalents
6328 in the X509_STORE_CTX structure.
6330 X509_STORE_CTX also has a 'flags' field which can be used
6331 to customise the verify behaviour.
6334 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
6335 excludes S/MIME capabilities.
6338 *) When a certificate request is read in keep a copy of the
6339 original encoding of the signed data and use it when outputing
6340 again. Signatures then use the original encoding rather than
6341 a decoded, encoded version which may cause problems if the
6342 request is improperly encoded.
6345 *) For consistency with other BIO_puts implementations, call
6346 buffer_write(b, ...) directly in buffer_puts instead of calling
6349 In BIO_puts, increment b->num_write as in BIO_write.
6350 [Peter.Sylvester@EdelWeb.fr]
6352 *) Fix BN_mul_word for the case where the word is 0. (We have to use
6353 BN_zero, we may not return a BIGNUM with an array consisting of
6357 *) Avoid calling abort() from within the library when problems are
6358 detected, except if preprocessor symbols have been defined
6359 (such as REF_CHECK, BN_DEBUG etc.).
6362 *) New openssl application 'rsautl'. This utility can be
6363 used for low level RSA operations. DER public key
6364 BIO/fp routines also added.
6367 *) New Configure entry and patches for compiling on QNX 4.
6368 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
6370 *) A demo state-machine implementation was sponsored by
6371 Nuron (http://www.nuron.com/) and is now available in
6372 demos/state_machine.
6375 *) New options added to the 'dgst' utility for signature
6376 generation and verification.
6379 *) Unrecognized PKCS#7 content types are now handled via a
6380 catch all ASN1_TYPE structure. This allows unsupported
6381 types to be stored as a "blob" and an application can
6382 encode and decode it manually.
6385 *) Fix various signed/unsigned issues to make a_strex.c
6387 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
6389 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
6390 length if passed a buffer. ASN1_INTEGER_to_BN failed
6391 if passed a NULL BN and its argument was negative.
6392 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
6394 *) Modification to PKCS#7 encoding routines to output definite
6395 length encoding. Since currently the whole structures are in
6396 memory there's not real point in using indefinite length
6397 constructed encoding. However if OpenSSL is compiled with
6398 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
6401 *) Added BIO_vprintf() and BIO_vsnprintf().
6404 *) Added more prefixes to parse for in the the strings written
6405 through a logging bio, to cover all the levels that are available
6406 through syslog. The prefixes are now:
6408 PANIC, EMERG, EMR => LOG_EMERG
6409 ALERT, ALR => LOG_ALERT
6410 CRIT, CRI => LOG_CRIT
6411 ERROR, ERR => LOG_ERR
6412 WARNING, WARN, WAR => LOG_WARNING
6413 NOTICE, NOTE, NOT => LOG_NOTICE
6414 INFO, INF => LOG_INFO
6415 DEBUG, DBG => LOG_DEBUG
6417 and as before, if none of those prefixes are present at the
6418 beginning of the string, LOG_ERR is chosen.
6420 On Win32, the LOG_* levels are mapped according to this:
6422 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
6423 LOG_WARNING => EVENTLOG_WARNING_TYPE
6424 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
6428 *) Made it possible to reconfigure with just the configuration
6429 argument "reconf" or "reconfigure". The command line arguments
6430 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
6431 and are retrieved from there when reconfiguring.
6435 [Assar Westerlund <assar@sics.se>, Richard Levitte]
6437 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
6440 *) The obj_dat.pl script was messing up the sorting of object
6441 names. The reason was that it compared the quoted version
6442 of strings as a result "OCSP" > "OCSP Signing" because
6443 " > SPACE. Changed script to store unquoted versions of
6444 names and add quotes on output. It was also omitting some
6445 names from the lookup table if they were given a default
6446 value (that is if SN is missing it is given the same
6447 value as LN and vice versa), these are now added on the
6448 grounds that if an object has a name we should be able to
6449 look it up. Finally added warning output when duplicate
6450 short or long names are found.
6453 *) Changes needed for Tandem NSK.
6454 [Scott Uroff <scott@xypro.com>]
6456 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
6457 RSA_padding_check_SSLv23(), special padding was never detected
6458 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
6459 version rollback attacks was not effective.
6461 In s23_clnt.c, don't use special rollback-attack detection padding
6462 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
6463 client; similarly, in s23_srvr.c, don't do the rollback check if
6464 SSL 2.0 is the only protocol enabled in the server.
6467 *) Make it possible to get hexdumps of unprintable data with 'openssl
6468 asn1parse'. By implication, the functions ASN1_parse_dump() and
6469 BIO_dump_indent() are added.
6472 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
6473 these print out strings and name structures based on various
6474 flags including RFC2253 support and proper handling of
6475 multibyte characters. Added options to the 'x509' utility
6476 to allow the various flags to be set.
6479 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
6480 Also change the functions X509_cmp_current_time() and
6481 X509_gmtime_adj() work with an ASN1_TIME structure,
6482 this will enable certificates using GeneralizedTime in validity
6483 dates to be checked.
6486 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
6487 negative public key encodings) on by default,
6488 NO_NEG_PUBKEY_BUG can be set to disable it.
6491 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
6492 content octets. An i2c_ASN1_OBJECT is unnecessary because
6493 the encoding can be trivially obtained from the structure.
6496 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
6497 not read locks (CRYPTO_r_[un]lock).
6500 *) A first attempt at creating official support for shared
6501 libraries through configuration. I've kept it so the
6502 default is static libraries only, and the OpenSSL programs
6503 are always statically linked for now, but there are
6504 preparations for dynamic linking in place.
6505 This has been tested on Linux and Tru64.
6508 *) Randomness polling function for Win9x, as described in:
6509 Peter Gutmann, Software Generation of Practically Strong
6513 *) Fix so PRNG is seeded in req if using an already existing
6517 *) New options to smime application. -inform and -outform
6518 allow alternative formats for the S/MIME message including
6519 PEM and DER. The -content option allows the content to be
6520 specified separately. This should allow things like Netscape
6521 form signing output easier to verify.
6524 *) Fix the ASN1 encoding of tags using the 'long form'.
6527 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
6528 STRING types. These convert content octets to and from the
6529 underlying type. The actual tag and length octets are
6530 already assumed to have been read in and checked. These
6531 are needed because all other string types have virtually
6532 identical handling apart from the tag. By having versions
6533 of the ASN1 functions that just operate on content octets
6534 IMPLICIT tagging can be handled properly. It also allows
6535 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
6536 and ASN1_INTEGER are identical apart from the tag.
6539 *) Change the handling of OID objects as follows:
6541 - New object identifiers are inserted in objects.txt, following
6542 the syntax given in objects.README.
6543 - objects.pl is used to process obj_mac.num and create a new
6545 - obj_dat.pl is used to create a new obj_dat.h, using the data in
6548 This is currently kind of a hack, and the perl code in objects.pl
6549 isn't very elegant, but it works as I intended. The simplest way
6550 to check that it worked correctly is to look in obj_dat.h and
6551 check the array nid_objs and make sure the objects haven't moved
6552 around (this is important!). Additions are OK, as well as
6553 consistent name changes.
6556 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
6559 *) Addition of the command line parameter '-rand file' to 'openssl req'.
6560 The given file adds to whatever has already been seeded into the
6561 random pool through the RANDFILE configuration file option or
6562 environment variable, or the default random state file.
6565 *) mkstack.pl now sorts each macro group into lexical order.
6566 Previously the output order depended on the order the files
6567 appeared in the directory, resulting in needless rewriting
6571 *) Patches to make OpenSSL compile under Win32 again. Mostly
6572 work arounds for the VC++ problem that it treats func() as
6573 func(void). Also stripped out the parts of mkdef.pl that
6574 added extra typesafe functions: these no longer exist.
6577 *) Reorganisation of the stack code. The macros are now all
6578 collected in safestack.h . Each macro is defined in terms of
6579 a "stack macro" of the form SKM_<name>(type, a, b). The
6580 DEBUG_SAFESTACK is now handled in terms of function casts,
6581 this has the advantage of retaining type safety without the
6582 use of additional functions. If DEBUG_SAFESTACK is not defined
6583 then the non typesafe macros are used instead. Also modified the
6584 mkstack.pl script to handle the new form. Needs testing to see
6585 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
6586 the default if no major problems. Similar behaviour for ASN1_SET_OF
6587 and PKCS12_STACK_OF.
6590 *) When some versions of IIS use the 'NET' form of private key the
6591 key derivation algorithm is different. Normally MD5(password) is
6592 used as a 128 bit RC4 key. In the modified case
6593 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
6594 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
6595 as the old Netscape_RSA functions except they have an additional
6596 'sgckey' parameter which uses the modified algorithm. Also added
6597 an -sgckey command line option to the rsa utility. Thanks to
6598 Adrian Peck <bertie@ncipher.com> for posting details of the modified
6599 algorithm to openssl-dev.
6602 *) The evp_local.h macros were using 'c.##kname' which resulted in
6603 invalid expansion on some systems (SCO 5.0.5 for example).
6604 Corrected to 'c.kname'.
6605 [Phillip Porch <root@theporch.com>]
6607 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
6608 a STACK of email addresses from a certificate or request, these look
6609 in the subject name and the subject alternative name extensions and
6610 omit any duplicate addresses.
6613 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
6614 This makes DSA verification about 2 % faster.
6617 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
6618 (meaning that now 2^5 values will be precomputed, which is only 4 KB
6619 plus overhead for 1024 bit moduli).
6620 This makes exponentiations about 0.5 % faster for 1024 bit
6621 exponents (as measured by "openssl speed rsa2048").
6624 *) Rename memory handling macros to avoid conflicts with other
6626 Malloc => OPENSSL_malloc
6627 Malloc_locked => OPENSSL_malloc_locked
6628 Realloc => OPENSSL_realloc
6629 Free => OPENSSL_free
6632 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
6633 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
6636 *) CygWin32 support.
6637 [John Jarvie <jjarvie@newsguy.com>]
6639 *) The type-safe stack code has been rejigged. It is now only compiled
6640 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
6641 by default all type-specific stack functions are "#define"d back to
6642 standard stack functions. This results in more streamlined output
6643 but retains the type-safety checking possibilities of the original
6647 *) The STACK code has been cleaned up, and certain type declarations
6648 that didn't make a lot of sense have been brought in line. This has
6649 also involved a cleanup of sorts in safestack.h to more correctly
6650 map type-safe stack functions onto their plain stack counterparts.
6651 This work has also resulted in a variety of "const"ifications of
6652 lots of the code, especially "_cmp" operations which should normally
6653 be prototyped with "const" parameters anyway.
6656 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
6657 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
6658 (The PRNG state consists of two parts, the large pool 'state' and 'md',
6659 where all of 'md' is used each time the PRNG is used, but 'state'
6660 is used only indexed by a cyclic counter. As entropy may not be
6661 well distributed from the beginning, 'md' is important as a
6662 chaining variable. However, the output function chains only half
6663 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
6664 all of 'md', and seeding with STATE_SIZE dummy bytes will result
6665 in all of 'state' being rewritten, with the new values depending
6666 on virtually all of 'md'. This overcomes the 80 bit limitation.)
6669 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
6670 the handshake is continued after ssl_verify_cert_chain();
6671 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
6672 can lead to 'unexplainable' connection aborts later.
6673 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
6675 *) Major EVP API cipher revision.
6676 Add hooks for extra EVP features. This allows various cipher
6677 parameters to be set in the EVP interface. Support added for variable
6678 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
6679 setting of RC2 and RC5 parameters.
6681 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
6684 Remove lots of duplicated code from the EVP library. For example *every*
6685 cipher init() function handles the 'iv' in the same way according to the
6686 cipher mode. They also all do nothing if the 'key' parameter is NULL and
6687 for CFB and OFB modes they zero ctx->num.
6689 New functionality allows removal of S/MIME code RC2 hack.
6691 Most of the routines have the same form and so can be declared in terms
6694 By shifting this to the top level EVP_CipherInit() it can be removed from
6695 all individual ciphers. If the cipher wants to handle IVs or keys
6696 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
6699 Change lots of functions like EVP_EncryptUpdate() to now return a
6700 value: although software versions of the algorithms cannot fail
6701 any installed hardware versions can.
6704 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
6705 this option is set, tolerate broken clients that send the negotiated
6706 protocol version number instead of the requested protocol version
6710 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
6711 i.e. non-zero for export ciphersuites, zero otherwise.
6712 Previous versions had this flag inverted, inconsistent with
6713 rsa_tmp_cb (..._TMP_RSA_CB).
6714 [Bodo Moeller; problem reported by Amit Chopra]
6716 *) Add missing DSA library text string. Work around for some IIS
6717 key files with invalid SEQUENCE encoding.
6720 *) Add a document (doc/standards.txt) that list all kinds of standards
6721 and so on that are implemented in OpenSSL.
6724 *) Enhance c_rehash script. Old version would mishandle certificates
6725 with the same subject name hash and wouldn't handle CRLs at all.
6726 Added -fingerprint option to crl utility, to support new c_rehash
6730 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
6733 *) Fix for SSL server purpose checking. Server checking was
6734 rejecting certificates which had extended key usage present
6735 but no ssl client purpose.
6736 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
6738 *) Make PKCS#12 code work with no password. The PKCS#12 spec
6739 is a little unclear about how a blank password is handled.
6740 Since the password in encoded as a BMPString with terminating
6741 double NULL a zero length password would end up as just the
6742 double NULL. However no password at all is different and is
6743 handled differently in the PKCS#12 key generation code. NS
6744 treats a blank password as zero length. MSIE treats it as no
6745 password on export: but it will try both on import. We now do
6746 the same: PKCS12_parse() tries zero length and no password if
6747 the password is set to "" or NULL (NULL is now a valid password:
6748 it wasn't before) as does the pkcs12 application.
6751 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
6752 perror when PEM_read_bio_X509_REQ fails, the error message must
6753 be obtained from the error queue.
6756 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
6757 it in ERR_remove_state if appropriate, and change ERR_get_state
6758 accordingly to avoid race conditions (this is necessary because
6759 thread_hash is no longer constant once set).
6762 *) Bugfix for linux-elf makefile.one.
6765 *) RSA_get_default_method() will now cause a default
6766 RSA_METHOD to be chosen if one doesn't exist already.
6767 Previously this was only set during a call to RSA_new()
6768 or RSA_new_method(NULL) meaning it was possible for
6769 RSA_get_default_method() to return NULL.
6772 *) Added native name translation to the existing DSO code
6773 that will convert (if the flag to do so is set) filenames
6774 that are sufficiently small and have no path information
6775 into a canonical native form. Eg. "blah" converted to
6776 "libblah.so" or "blah.dll" etc.
6779 *) New function ERR_error_string_n(e, buf, len) which is like
6780 ERR_error_string(e, buf), but writes at most 'len' bytes
6781 including the 0 terminator. For ERR_error_string_n, 'buf'
6783 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
6785 *) CONF library reworked to become more general. A new CONF
6786 configuration file reader "class" is implemented as well as a
6787 new functions (NCONF_*, for "New CONF") to handle it. The now
6788 old CONF_* functions are still there, but are reimplemented to
6789 work in terms of the new functions. Also, a set of functions
6790 to handle the internal storage of the configuration data is
6791 provided to make it easier to write new configuration file
6792 reader "classes" (I can definitely see something reading a
6793 configuration file in XML format, for example), called _CONF_*,
6794 or "the configuration storage API"...
6796 The new configuration file reading functions are:
6798 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
6799 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
6801 NCONF_default, NCONF_WIN32
6803 NCONF_dump_fp, NCONF_dump_bio
6805 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
6806 NCONF_new creates a new CONF object. This works in the same way
6807 as other interfaces in OpenSSL, like the BIO interface.
6808 NCONF_dump_* dump the internal storage of the configuration file,
6809 which is useful for debugging. All other functions take the same
6810 arguments as the old CONF_* functions wth the exception of the
6811 first that must be a `CONF *' instead of a `LHASH *'.
6813 To make it easer to use the new classes with the old CONF_* functions,
6814 the function CONF_set_default_method is provided.
6817 *) Add '-tls1' option to 'openssl ciphers', which was already
6818 mentioned in the documentation but had not been implemented.
6819 (This option is not yet really useful because even the additional
6820 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
6823 *) Initial DSO code added into libcrypto for letting OpenSSL (and
6824 OpenSSL-based applications) load shared libraries and bind to
6825 them in a portable way.
6826 [Geoff Thorpe, with contributions from Richard Levitte]
6828 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
6830 *) Make sure _lrotl and _lrotr are only used with MSVC.
6832 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
6833 (the default implementation of RAND_status).
6835 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
6836 to '-clrext' (= clear extensions), as intended and documented.
6837 [Bodo Moeller; inconsistency pointed out by Michael Attili
6840 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
6841 was larger than the MD block size.
6842 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
6844 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
6845 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
6846 using the passed key: if the passed key was a private key the result
6847 of X509_print(), for example, would be to print out all the private key
6851 *) des_quad_cksum() byte order bug fix.
6852 [Ulf Möller, using the problem description in krb4-0.9.7, where
6853 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
6855 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
6857 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
6859 *) For easily testing in shell scripts whether some command
6860 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
6861 returns with exit code 0 iff no command of the given name is available.
6862 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
6863 the output goes to stdout and nothing is printed to stderr.
6864 Additional arguments are always ignored.
6866 Since for each cipher there is a command of the same name,
6867 the 'no-cipher' compilation switches can be tested this way.
6869 ('openssl no-XXX' is not able to detect pseudo-commands such
6870 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
6873 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
6876 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
6877 is set; it will be thrown away anyway because each handshake creates
6879 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
6880 to parameters -- in previous versions (since OpenSSL 0.9.3) the
6881 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
6882 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
6885 *) New s_client option -ign_eof: EOF at stdin is ignored, and
6886 'Q' and 'R' lose their special meanings (quit/renegotiate).
6887 This is part of what -quiet does; unlike -quiet, -ign_eof
6888 does not suppress any output.
6891 *) Add compatibility options to the purpose and trust code. The
6892 purpose X509_PURPOSE_ANY is "any purpose" which automatically
6893 accepts a certificate or CA, this was the previous behaviour,
6894 with all the associated security issues.
6896 X509_TRUST_COMPAT is the old trust behaviour: only and
6897 automatically trust self signed roots in certificate store. A
6898 new trust setting X509_TRUST_DEFAULT is used to specify that
6899 a purpose has no associated trust setting and it should instead
6900 use the value in the default purpose.
6903 *) Fix the PKCS#8 DSA private key code so it decodes keys again
6904 and fix a memory leak.
6907 *) In util/mkerr.pl (which implements 'make errors'), preserve
6908 reason strings from the previous version of the .c file, as
6909 the default to have only downcase letters (and digits) in
6910 automatically generated reasons codes is not always appropriate.
6913 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
6914 using strerror. Previously, ERR_reason_error_string() returned
6915 library names as reason strings for SYSerr; but SYSerr is a special
6916 case where small numbers are errno values, not library numbers.
6919 *) Add '-dsaparam' option to 'openssl dhparam' application. This
6920 converts DSA parameters into DH parameters. (When creating parameters,
6921 DSA_generate_parameters is used.)
6924 *) Include 'length' (recommended exponent length) in C code generated
6925 by 'openssl dhparam -C'.
6928 *) The second argument to set_label in perlasm was already being used
6929 so couldn't be used as a "file scope" flag. Moved to third argument
6933 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
6934 instead of RAND_bytes for encryption IVs and salts.
6937 *) Include RAND_status() into RAND_METHOD instead of implementing
6938 it only for md_rand.c Otherwise replacing the PRNG by calling
6939 RAND_set_rand_method would be impossible.
6942 *) Don't let DSA_generate_key() enter an infinite loop if the random
6943 number generation fails.
6946 *) New 'rand' application for creating pseudo-random output.
6949 *) Added configuration support for Linux/IA64
6950 [Rolf Haberrecker <rolf@suse.de>]
6952 *) Assembler module support for Mingw32.
6955 *) Shared library support for HPUX (in shlib/).
6956 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
6958 *) Shared library support for Solaris gcc.
6959 [Lutz Behnke <behnke@trustcenter.de>]
6961 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
6963 *) PKCS7_encrypt() was adding text MIME headers twice because they
6964 were added manually and by SMIME_crlf_copy().
6967 *) In bntest.c don't call BN_rand with zero bits argument.
6968 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
6970 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
6971 case was implemented. This caused BN_div_recp() to fail occasionally.
6974 *) Add an optional second argument to the set_label() in the perl
6975 assembly language builder. If this argument exists and is set
6976 to 1 it signals that the assembler should use a symbol whose
6977 scope is the entire file, not just the current function. This
6978 is needed with MASM which uses the format label:: for this scope.
6979 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
6981 *) Change the ASN1 types so they are typedefs by default. Before
6982 almost all types were #define'd to ASN1_STRING which was causing
6983 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
6987 *) Change names of new functions to the new get1/get0 naming
6988 convention: After 'get1', the caller owns a reference count
6989 and has to call ..._free; 'get0' returns a pointer to some
6990 data structure without incrementing reference counters.
6991 (Some of the existing 'get' functions increment a reference
6992 counter, some don't.)
6993 Similarly, 'set1' and 'add1' functions increase reference
6994 counters or duplicate objects.
6997 *) Allow for the possibility of temp RSA key generation failure:
6998 the code used to assume it always worked and crashed on failure.
7001 *) Fix potential buffer overrun problem in BIO_printf().
7002 [Ulf Möller, using public domain code by Patrick Powell; problem
7003 pointed out by David Sacerdote <das33@cornell.edu>]
7005 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7006 RAND_egd() and RAND_status(). In the command line application,
7007 the EGD socket can be specified like a seed file using RANDFILE
7011 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7012 Some CAs (e.g. Verisign) distribute certificates in this form.
7015 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7016 list to exclude them. This means that no special compilation option
7017 is needed to use anonymous DH: it just needs to be included in the
7021 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7022 EVP_MD_type. The old functionality is available in a new macro called
7023 EVP_MD_md(). Change code that uses it and update docs.
7026 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7027 where the 'void *' argument is replaced by a function pointer argument.
7028 Previously 'void *' was abused to point to functions, which works on
7029 many platforms, but is not correct. As these functions are usually
7030 called by macros defined in OpenSSL header files, most source code
7031 should work without changes.
7034 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7035 sections with information on -D... compiler switches used for
7036 compiling the library so that applications can see them. To enable
7037 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7038 must be defined. E.g.,
7039 #define OPENSSL_ALGORITHM_DEFINES
7040 #include <openssl/opensslconf.h>
7041 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7042 [Richard Levitte, Ulf and Bodo Möller]
7044 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7048 *) Change the 'other' type in certificate aux info to a STACK_OF
7049 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7050 the required ASN1 format: arbitrary types determined by an OID.
7053 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7054 argument to 'req'. This is not because the function is newer or
7055 better than others it just uses the work 'NEW' in the certificate
7056 request header lines. Some software needs this.
7059 *) Reorganise password command line arguments: now passwords can be
7060 obtained from various sources. Delete the PEM_cb function and make
7061 it the default behaviour: i.e. if the callback is NULL and the
7062 usrdata argument is not NULL interpret it as a null terminated pass
7063 phrase. If usrdata and the callback are NULL then the pass phrase
7064 is prompted for as usual.
7067 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7068 the support is automatically enabled. The resulting binaries will
7069 autodetect the card and use it if present.
7070 [Ben Laurie and Compaq Inc.]
7072 *) Work around for Netscape hang bug. This sends certificate request
7073 and server done in one record. Since this is perfectly legal in the
7074 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7075 the bugs/SSLv3 entry for more info.
7078 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7081 *) Add -rand argument to smime and pkcs12 applications and read/write
7085 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7088 *) Add command line password options to the remaining applications.
7091 *) Bug fix for BN_div_recp() for numerators with an even number of
7095 *) More tests in bntest.c, and changed test_bn output.
7098 *) ./config recognizes MacOS X now.
7101 *) Bug fix for BN_div() when the first words of num and divsor are
7102 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7105 *) Add support for various broken PKCS#8 formats, and command line
7106 options to produce them.
7109 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7110 get temporary BIGNUMs from a BN_CTX.
7113 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7117 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7118 include a #define from the old name to the new. The original intent
7119 was that statically linked binaries could for example just call
7120 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7121 link with digests. This never worked becayse SSLeay_add_all_digests()
7122 and SSLeay_add_all_ciphers() were in the same source file so calling
7123 one would link with the other. They are now in separate source files.
7126 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7129 *) Use a less unusual form of the Miller-Rabin primality test (it used
7130 a binary algorithm for exponentiation integrated into the Miller-Rabin
7131 loop, our standard modexp algorithms are faster).
7134 *) Support for the EBCDIC character set completed.
7135 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7137 *) Source code cleanups: use const where appropriate, eliminate casts,
7138 use void * instead of char * in lhash.
7141 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7142 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7143 this the server could overwrite ephemeral keys that the client
7147 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7148 using 50 iterations of the Rabin-Miller test.
7150 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7151 iterations of the Rabin-Miller test as required by the appendix
7152 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7153 As BN_is_prime_fasttest includes trial division, DSA parameter
7154 generation becomes much faster.
7156 This implies a change for the callback functions in DSA_is_prime
7157 and DSA_generate_parameters: The callback function is called once
7158 for each positive witness in the Rabin-Miller test, not just
7159 occasionally in the inner loop; and the parameters to the
7160 callback function now provide an iteration count for the outer
7161 loop rather than for the current invocation of the inner loop.
7162 DSA_generate_parameters additionally can call the callback
7163 function with an 'iteration count' of -1, meaning that a
7164 candidate has passed the trial division test (when q is generated
7165 from an application-provided seed, trial division is skipped).
7168 *) New function BN_is_prime_fasttest that optionally does trial
7169 division before starting the Rabin-Miller test and has
7170 an additional BN_CTX * argument (whereas BN_is_prime always
7171 has to allocate at least one BN_CTX).
7172 'callback(1, -1, cb_arg)' is called when a number has passed the
7173 trial division stage.
7176 *) Fix for bug in CRL encoding. The validity dates weren't being handled
7180 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
7183 *) New function BN_pseudo_rand().
7186 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
7187 bignum version of BN_from_montgomery() with the working code from
7188 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
7192 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
7193 made it impossible to use the same SSL_SESSION data structure in
7194 SSL2 clients in multiple threads.
7197 *) The return value of RAND_load_file() no longer counts bytes obtained
7198 by stat(). RAND_load_file(..., -1) is new and uses the complete file
7199 to seed the PRNG (previously an explicit byte count was required).
7200 [Ulf Möller, Bodo Möller]
7202 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
7203 used (char *) instead of (void *) and had casts all over the place.
7206 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
7209 *) Retain source code compatibility for BN_prime_checks macro:
7210 BN_is_prime(..., BN_prime_checks, ...) now uses
7211 BN_prime_checks_for_size to determine the appropriate number of
7212 Rabin-Miller iterations.
7215 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
7216 DH_CHECK_P_NOT_SAFE_PRIME.
7217 (Check if this is true? OpenPGP calls them "strong".)
7220 *) Merge the functionality of "dh" and "gendh" programs into a new program
7221 "dhparam". The old programs are retained for now but will handle DH keys
7222 (instead of parameters) in future.
7225 *) Make the ciphers, s_server and s_client programs check the return values
7226 when a new cipher list is set.
7229 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
7230 ciphers. Before when the 56bit ciphers were enabled the sorting was
7233 The syntax for the cipher sorting has been extended to support sorting by
7234 cipher-strength (using the strength_bits hard coded in the tables).
7235 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
7237 Fix a bug in the cipher-command parser: when supplying a cipher command
7238 string with an "undefined" symbol (neither command nor alphanumeric
7239 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
7240 an error is flagged.
7242 Due to the strength-sorting extension, the code of the
7243 ssl_create_cipher_list() function was completely rearranged. I hope that
7244 the readability was also increased :-)
7245 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7247 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
7248 for the first serial number and places 2 in the serial number file. This
7249 avoids problems when the root CA is created with serial number zero and
7250 the first user certificate has the same issuer name and serial number
7254 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
7255 the new code. Add documentation for this stuff.
7258 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
7259 X509_*() to X509at_*() on the grounds that they don't handle X509
7260 structures and behave in an analagous way to the X509v3 functions:
7261 they shouldn't be called directly but wrapper functions should be used
7264 So we also now have some wrapper functions that call the X509at functions
7265 when passed certificate requests. (TO DO: similar things can be done with
7266 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
7267 things. Some of these need some d2i or i2d and print functionality
7268 because they handle more complex structures.)
7271 *) Add missing #ifndefs that caused missing symbols when building libssl
7272 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
7273 NO_RSA in ssl/s2*.c.
7274 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
7276 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
7277 has a return value which indicates the quality of the random data
7278 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
7279 error queue. New function RAND_pseudo_bytes() generates output that is
7280 guaranteed to be unique but not unpredictable. RAND_add is like
7281 RAND_seed, but takes an extra argument for an entropy estimate
7282 (RAND_seed always assumes full entropy).
7285 *) Do more iterations of Rabin-Miller probable prime test (specifically,
7286 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
7287 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
7288 in crypto/bn/bn_prime.c for the complete table). This guarantees a
7289 false-positive rate of at most 2^-80 for random input.
7292 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
7295 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
7296 in the 0.9.5 release), this returns the chain
7297 from an X509_CTX structure with a dup of the stack and all
7298 the X509 reference counts upped: so the stack will exist
7299 after X509_CTX_cleanup() has been called. Modify pkcs12.c
7302 Also make SSL_SESSION_print() print out the verify return
7306 *) Add manpage for the pkcs12 command. Also change the default
7307 behaviour so MAC iteration counts are used unless the new
7308 -nomaciter option is used. This improves file security and
7309 only older versions of MSIE (4.0 for example) need it.
7312 *) Honor the no-xxx Configure options when creating .DEF files.
7315 *) Add PKCS#10 attributes to field table: challengePassword,
7316 unstructuredName and unstructuredAddress. These are taken from
7317 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
7318 international characters are used.
7320 More changes to X509_ATTRIBUTE code: allow the setting of types
7321 based on strings. Remove the 'loc' parameter when adding
7322 attributes because these will be a SET OF encoding which is sorted
7326 *) Initial changes to the 'req' utility to allow request generation
7327 automation. This will allow an application to just generate a template
7328 file containing all the field values and have req construct the
7331 Initial support for X509_ATTRIBUTE handling. Stacks of these are
7332 used all over the place including certificate requests and PKCS#7
7333 structures. They are currently handled manually where necessary with
7334 some primitive wrappers for PKCS#7. The new functions behave in a
7335 manner analogous to the X509 extension functions: they allow
7336 attributes to be looked up by NID and added.
7338 Later something similar to the X509V3 code would be desirable to
7339 automatically handle the encoding, decoding and printing of the
7340 more complex types. The string types like challengePassword can
7341 be handled by the string table functions.
7343 Also modified the multi byte string table handling. Now there is
7344 a 'global mask' which masks out certain types. The table itself
7345 can use the flag STABLE_NO_MASK to ignore the mask setting: this
7346 is useful when for example there is only one permissible type
7347 (as in countryName) and using the mask might result in no valid
7351 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
7352 SSL_get_peer_finished to allow applications to obtain the latest
7353 Finished messages sent to the peer or expected from the peer,
7354 respectively. (SSL_get_peer_finished is usually the Finished message
7355 actually received from the peer, otherwise the protocol will be aborted.)
7357 As the Finished message are message digests of the complete handshake
7358 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
7359 be used for external authentication procedures when the authentication
7360 provided by SSL/TLS is not desired or is not enough.
7363 *) Enhanced support for Alpha Linux is added. Now ./config checks if
7364 the host supports BWX extension and if Compaq C is present on the
7365 $PATH. Just exploiting of the BWX extension results in 20-30%
7366 performance kick for some algorithms, e.g. DES and RC4 to mention
7367 a couple. Compaq C in turn generates ~20% faster code for MD5 and
7371 *) Add support for MS "fast SGC". This is arguably a violation of the
7372 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
7373 weak crypto and after checking the certificate is SGC a second one
7374 with strong crypto. MS SGC stops the first handshake after receiving
7375 the server certificate message and sends a second client hello. Since
7376 a server will typically do all the time consuming operations before
7377 expecting any further messages from the client (server key exchange
7378 is the most expensive) there is little difference between the two.
7380 To get OpenSSL to support MS SGC we have to permit a second client
7381 hello message after we have sent server done. In addition we have to
7382 reset the MAC if we do get this second client hello.
7385 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
7386 if a DER encoded private key is RSA or DSA traditional format. Changed
7387 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
7388 format DER encoded private key. Newer code should use PKCS#8 format which
7389 has the key type encoded in the ASN1 structure. Added DER private key
7390 support to pkcs8 application.
7393 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
7394 ciphersuites has been selected (as required by the SSL 3/TLS 1
7395 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
7396 is set, we interpret this as a request to violate the specification
7397 (the worst that can happen is a handshake failure, and 'correct'
7398 behaviour would result in a handshake failure anyway).
7401 *) In SSL_CTX_add_session, take into account that there might be multiple
7402 SSL_SESSION structures with the same session ID (e.g. when two threads
7403 concurrently obtain them from an external cache).
7404 The internal cache can handle only one SSL_SESSION with a given ID,
7405 so if there's a conflict, we now throw out the old one to achieve
7409 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
7410 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
7411 some routines that use cipher OIDs: some ciphers do not have OIDs
7412 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
7416 *) Simplify the trust setting structure and code. Now we just have
7417 two sequences of OIDs for trusted and rejected settings. These will
7418 typically have values the same as the extended key usage extension
7419 and any application specific purposes.
7421 The trust checking code now has a default behaviour: it will just
7422 check for an object with the same NID as the passed id. Functions can
7423 be provided to override either the default behaviour or the behaviour
7424 for a given id. SSL client, server and email already have functions
7425 in place for compatibility: they check the NID and also return "trusted"
7426 if the certificate is self signed.
7429 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
7430 traditional format into an EVP_PKEY structure.
7433 *) Add a password callback function PEM_cb() which either prompts for
7434 a password if usr_data is NULL or otherwise assumes it is a null
7435 terminated password. Allow passwords to be passed on command line
7436 environment or config files in a few more utilities.
7439 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
7440 keys. Add some short names for PKCS#8 PBE algorithms and allow them
7441 to be specified on the command line for the pkcs8 and pkcs12 utilities.
7442 Update documentation.
7445 *) Support for ASN1 "NULL" type. This could be handled before by using
7446 ASN1_TYPE but there wasn't any function that would try to read a NULL
7447 and produce an error if it couldn't. For compatibility we also have
7448 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
7449 don't allocate anything because they don't need to.
7452 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
7454 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
7456 *) Rebuild of the memory allocation routines used by OpenSSL code and
7457 possibly others as well. The purpose is to make an interface that
7458 provide hooks so anyone can build a separate set of allocation and
7459 deallocation routines to be used by OpenSSL, for example memory
7460 pool implementations, or something else, which was previously hard
7461 since Malloc(), Realloc() and Free() were defined as macros having
7462 the values malloc, realloc and free, respectively (except for Win32
7463 compilations). The same is provided for memory debugging code.
7464 OpenSSL already comes with functionality to find memory leaks, but
7465 this gives people a chance to debug other memory problems.
7467 With these changes, a new set of functions and macros have appeared:
7469 CRYPTO_set_mem_debug_functions() [F]
7470 CRYPTO_get_mem_debug_functions() [F]
7471 CRYPTO_dbg_set_options() [F]
7472 CRYPTO_dbg_get_options() [F]
7473 CRYPTO_malloc_debug_init() [M]
7475 The memory debug functions are NULL by default, unless the library
7476 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
7477 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
7478 gives the standard debugging functions that come with OpenSSL) or
7479 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
7480 provided by the library user) must be used. When the standard
7481 debugging functions are used, CRYPTO_dbg_set_options can be used to
7482 request additional information:
7483 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
7484 the CRYPTO_MDEBUG_xxx macro when compiling the library.
7486 Also, things like CRYPTO_set_mem_functions will always give the
7487 expected result (the new set of functions is used for allocation
7488 and deallocation) at all times, regardless of platform and compiler
7491 To finish it up, some functions that were never use in any other
7492 way than through macros have a new API and new semantic:
7495 CRYPTO_dbg_realloc()
7498 All macros of value have retained their old syntax.
7499 [Richard Levitte and Bodo Moeller]
7501 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
7502 ordering of SMIMECapabilities wasn't in "strength order" and there
7503 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
7507 *) Some ASN1 types with illegal zero length encoding (INTEGER,
7508 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
7509 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
7511 *) Merge in my S/MIME library for OpenSSL. This provides a simple
7512 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
7513 functionality to handle multipart/signed properly) and a utility
7514 called 'smime' to call all this stuff. This is based on code I
7515 originally wrote for Celo who have kindly allowed it to be
7516 included in OpenSSL.
7519 *) Add variants des_set_key_checked and des_set_key_unchecked of
7520 des_set_key (aka des_key_sched). Global variable des_check_key
7521 decides which of these is called by des_set_key; this way
7522 des_check_key behaves as it always did, but applications and
7523 the library itself, which was buggy for des_check_key == 1,
7524 have a cleaner way to pick the version they need.
7527 *) New function PKCS12_newpass() which changes the password of a
7531 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
7532 dynamic mix. In both cases the ids can be used as an index into the
7533 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
7534 functions so they accept a list of the field values and the
7535 application doesn't need to directly manipulate the X509_TRUST
7539 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
7543 *) Modify the way the V3 extension code looks up extensions. This now
7544 works in a similar way to the object code: we have some "standard"
7545 extensions in a static table which is searched with OBJ_bsearch()
7546 and the application can add dynamic ones if needed. The file
7547 crypto/x509v3/ext_dat.h now has the info: this file needs to be
7548 updated whenever a new extension is added to the core code and kept
7549 in ext_nid order. There is a simple program 'tabtest.c' which checks
7550 this. New extensions are not added too often so this file can readily
7551 be maintained manually.
7553 There are two big advantages in doing things this way. The extensions
7554 can be looked up immediately and no longer need to be "added" using
7555 X509V3_add_standard_extensions(): this function now does nothing.
7556 [Side note: I get *lots* of email saying the extension code doesn't
7557 work because people forget to call this function]
7558 Also no dynamic allocation is done unless new extensions are added:
7559 so if we don't add custom extensions there is no need to call
7560 X509V3_EXT_cleanup().
7563 *) Modify enc utility's salting as follows: make salting the default. Add a
7564 magic header, so unsalted files fail gracefully instead of just decrypting
7565 to garbage. This is because not salting is a big security hole, so people
7566 should be discouraged from doing it.
7569 *) Fixes and enhancements to the 'x509' utility. It allowed a message
7570 digest to be passed on the command line but it only used this
7571 parameter when signing a certificate. Modified so all relevant
7572 operations are affected by the digest parameter including the
7573 -fingerprint and -x509toreq options. Also -x509toreq choked if a
7574 DSA key was used because it didn't fix the digest.
7577 *) Initial certificate chain verify code. Currently tests the untrusted
7578 certificates for consistency with the verify purpose (which is set
7579 when the X509_STORE_CTX structure is set up) and checks the pathlength.
7581 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
7582 this is because it will reject chains with invalid extensions whereas
7583 every previous version of OpenSSL and SSLeay made no checks at all.
7585 Trust code: checks the root CA for the relevant trust settings. Trust
7586 settings have an initial value consistent with the verify purpose: e.g.
7587 if the verify purpose is for SSL client use it expects the CA to be
7588 trusted for SSL client use. However the default value can be changed to
7589 permit custom trust settings: one example of this would be to only trust
7590 certificates from a specific "secure" set of CAs.
7592 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
7593 which should be used for version portability: especially since the
7594 verify structure is likely to change more often now.
7596 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
7597 to set them. If not set then assume SSL clients will verify SSL servers
7600 Two new options to the verify program: -untrusted allows a set of
7601 untrusted certificates to be passed in and -purpose which sets the
7602 intended purpose of the certificate. If a purpose is set then the
7603 new chain verify code is used to check extension consistency.
7606 *) Support for the authority information access extension.
7609 *) Modify RSA and DSA PEM read routines to transparently handle
7610 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
7611 public keys in a format compatible with certificate
7612 SubjectPublicKeyInfo structures. Unfortunately there were already
7613 functions called *_PublicKey_* which used various odd formats so
7614 these are retained for compatibility: however the DSA variants were
7615 never in a public release so they have been deleted. Changed dsa/rsa
7616 utilities to handle the new format: note no releases ever handled public
7617 keys so we should be OK.
7619 The primary motivation for this change is to avoid the same fiasco
7620 that dogs private keys: there are several incompatible private key
7621 formats some of which are standard and some OpenSSL specific and
7622 require various evil hacks to allow partial transparent handling and
7623 even then it doesn't work with DER formats. Given the option anything
7624 other than PKCS#8 should be dumped: but the other formats have to
7625 stay in the name of compatibility.
7627 With public keys and the benefit of hindsight one standard format
7628 is used which works with EVP_PKEY, RSA or DSA structures: though
7629 it clearly returns an error if you try to read the wrong kind of key.
7631 Added a -pubkey option to the 'x509' utility to output the public key.
7632 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
7633 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
7634 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
7635 that do the same as the EVP_PKEY_assign_*() except they up the
7636 reference count of the added key (they don't "swallow" the
7640 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
7641 CRLs would fail if the file contained no certificates or no CRLs:
7642 added a new function to read in both types and return the number
7643 read: this means that if none are read it will be an error. The
7644 DER versions of the certificate and CRL reader would always fail
7645 because it isn't possible to mix certificates and CRLs in DER format
7646 without choking one or the other routine. Changed this to just read
7647 a certificate: this is the best we can do. Also modified the code
7648 in apps/verify.c to take notice of return codes: it was previously
7649 attempting to read in certificates from NULL pointers and ignoring
7650 any errors: this is one reason why the cert and CRL reader seemed
7651 to work. It doesn't check return codes from the default certificate
7652 routines: these may well fail if the certificates aren't installed.
7655 *) Code to support otherName option in GeneralName.
7658 *) First update to verify code. Change the verify utility
7659 so it warns if it is passed a self signed certificate:
7660 for consistency with the normal behaviour. X509_verify
7661 has been modified to it will now verify a self signed
7662 certificate if *exactly* the same certificate appears
7663 in the store: it was previously impossible to trust a
7664 single self signed certificate. This means that:
7665 openssl verify ss.pem
7666 now gives a warning about a self signed certificate but
7667 openssl verify -CAfile ss.pem ss.pem
7671 *) For servers, store verify_result in SSL_SESSION data structure
7672 (and add it to external session representation).
7673 This is needed when client certificate verifications fails,
7674 but an application-provided verification callback (set by
7675 SSL_CTX_set_cert_verify_callback) allows accepting the session
7676 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
7677 but returns 1): When the session is reused, we have to set
7678 ssl->verify_result to the appropriate error code to avoid
7680 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
7682 *) Fix a bug in the new PKCS#7 code: it didn't consider the
7683 case in PKCS7_dataInit() where the signed PKCS7 structure
7684 didn't contain any existing data because it was being created.
7685 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
7687 *) Add a salt to the key derivation routines in enc.c. This
7688 forms the first 8 bytes of the encrypted file. Also add a
7689 -S option to allow a salt to be input on the command line.
7692 *) New function X509_cmp(). Oddly enough there wasn't a function
7693 to compare two certificates. We do this by working out the SHA1
7694 hash and comparing that. X509_cmp() will be needed by the trust
7698 *) SSL_get1_session() is like SSL_get_session(), but increments
7699 the reference count in the SSL_SESSION returned.
7700 [Geoff Thorpe <geoff@eu.c2.net>]
7702 *) Fix for 'req': it was adding a null to request attributes.
7703 Also change the X509_LOOKUP and X509_INFO code to handle
7704 certificate auxiliary information.
7707 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
7711 *) Add the possibility to add extra information to the memory leak
7712 detecting output, to form tracebacks, showing from where each
7713 allocation was originated: CRYPTO_push_info("constant string") adds
7714 the string plus current file name and line number to a per-thread
7715 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
7716 is like calling CYRPTO_pop_info() until the stack is empty.
7717 Also updated memory leak detection code to be multi-thread-safe.
7720 *) Add options -text and -noout to pkcs7 utility and delete the
7721 encryption options which never did anything. Update docs.
7724 *) Add options to some of the utilities to allow the pass phrase
7725 to be included on either the command line (not recommended on
7726 OSes like Unix) or read from the environment. Update the
7727 manpages and fix a few bugs.
7730 *) Add a few manpages for some of the openssl commands.
7733 *) Fix the -revoke option in ca. It was freeing up memory twice,
7734 leaking and not finding already revoked certificates.
7737 *) Extensive changes to support certificate auxiliary information.
7738 This involves the use of X509_CERT_AUX structure and X509_AUX
7739 functions. An X509_AUX function such as PEM_read_X509_AUX()
7740 can still read in a certificate file in the usual way but it
7741 will also read in any additional "auxiliary information". By
7742 doing things this way a fair degree of compatibility can be
7743 retained: existing certificates can have this information added
7744 using the new 'x509' options.
7746 Current auxiliary information includes an "alias" and some trust
7747 settings. The trust settings will ultimately be used in enhanced
7748 certificate chain verification routines: currently a certificate
7749 can only be trusted if it is self signed and then it is trusted
7753 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
7754 The problem was that one of the replacement routines had not been working
7755 since SSLeay releases. For now the offending routine has been replaced
7756 with non-optimised assembler. Even so, this now gives around 95%
7757 performance improvement for 1024 bit RSA signs.
7760 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
7761 handling. Most clients have the effective key size in bits equal to
7762 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
7763 A few however don't do this and instead use the size of the decrypted key
7764 to determine the RC2 key length and the AlgorithmIdentifier to determine
7765 the effective key length. In this case the effective key length can still
7766 be 40 bits but the key length can be 168 bits for example. This is fixed
7767 by manually forcing an RC2 key into the EVP_PKEY structure because the
7768 EVP code can't currently handle unusual RC2 key sizes: it always assumes
7769 the key length and effective key length are equal.
7772 *) Add a bunch of functions that should simplify the creation of
7773 X509_NAME structures. Now you should be able to do:
7774 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
7775 and have it automatically work out the correct field type and fill in
7776 the structures. The more adventurous can try:
7777 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
7778 and it will (hopefully) work out the correct multibyte encoding.
7781 *) Change the 'req' utility to use the new field handling and multibyte
7782 copy routines. Before the DN field creation was handled in an ad hoc
7783 way in req, ca, and x509 which was rather broken and didn't support
7784 BMPStrings or UTF8Strings. Since some software doesn't implement
7785 BMPStrings or UTF8Strings yet, they can be enabled using the config file
7786 using the dirstring_type option. See the new comment in the default
7787 openssl.cnf for more info.
7790 *) Make crypto/rand/md_rand.c more robust:
7791 - Assure unique random numbers after fork().
7792 - Make sure that concurrent threads access the global counter and
7793 md serializably so that we never lose entropy in them
7794 or use exactly the same state in multiple threads.
7795 Access to the large state is not always serializable because
7796 the additional locking could be a performance killer, and
7797 md should be large enough anyway.
7800 *) New file apps/app_rand.c with commonly needed functionality
7801 for handling the random seed file.
7803 Use the random seed file in some applications that previously did not:
7805 dsaparam -genkey (which also ignored its '-rand' option),
7808 x509 (when signing).
7809 Except on systems with /dev/urandom, it is crucial to have a random
7810 seed file at least for key creation, DSA signing, and for DH exchanges;
7811 for RSA signatures we could do without one.
7813 gendh and gendsa (unlike genrsa) used to read only the first byte
7814 of each file listed in the '-rand' option. The function as previously
7815 found in genrsa is now in app_rand.c and is used by all programs
7816 that support '-rand'.
7819 *) In RAND_write_file, use mode 0600 for creating files;
7820 don't just chmod when it may be too late.
7823 *) Report an error from X509_STORE_load_locations
7824 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
7827 *) New function ASN1_mbstring_copy() this copies a string in either
7828 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
7829 into an ASN1_STRING type. A mask of permissible types is passed
7830 and it chooses the "minimal" type to use or an error if not type
7834 *) Add function equivalents to the various macros in asn1.h. The old
7835 macros are retained with an M_ prefix. Code inside the library can
7836 use the M_ macros. External code (including the openssl utility)
7837 should *NOT* in order to be "shared library friendly".
7840 *) Add various functions that can check a certificate's extensions
7841 to see if it usable for various purposes such as SSL client,
7842 server or S/MIME and CAs of these types. This is currently
7843 VERY EXPERIMENTAL but will ultimately be used for certificate chain
7844 verification. Also added a -purpose flag to x509 utility to
7845 print out all the purposes.
7848 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
7852 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
7853 for, obtain and decode and extension and obtain its critical flag.
7854 This allows all the necessary extension code to be handled in a
7855 single function call.
7858 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
7859 platforms. See crypto/rc4/rc4_enc.c for further details.
7862 *) New -noout option to asn1parse. This causes no output to be produced
7863 its main use is when combined with -strparse and -out to extract data
7864 from a file (which may not be in ASN.1 format).
7867 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
7868 when producing the local key id.
7869 [Richard Levitte <levitte@stacken.kth.se>]
7871 *) New option -dhparam in s_server. This allows a DH parameter file to be
7872 stated explicitly. If it is not stated then it tries the first server
7873 certificate file. The previous behaviour hard coded the filename
7877 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
7878 a public key to be input or output. For example:
7879 openssl rsa -in key.pem -pubout -out pubkey.pem
7880 Also added necessary DSA public key functions to handle this.
7883 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
7884 in the message. This was handled by allowing
7885 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
7886 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
7888 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
7889 to the end of the strings whereas this didn't. This would cause problems
7890 if strings read with d2i_ASN1_bytes() were later modified.
7891 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
7893 *) Fix for base64 decode bug. When a base64 bio reads only one line of
7894 data and it contains EOF it will end up returning an error. This is
7895 caused by input 46 bytes long. The cause is due to the way base64
7896 BIOs find the start of base64 encoded data. They do this by trying a
7897 trial decode on each line until they find one that works. When they
7898 do a flag is set and it starts again knowing it can pass all the
7899 data directly through the decoder. Unfortunately it doesn't reset
7900 the context it uses. This means that if EOF is reached an attempt
7901 is made to pass two EOFs through the context and this causes the
7902 resulting error. This can also cause other problems as well. As is
7903 usual with these problems it takes *ages* to find and the fix is
7904 trivial: move one line.
7905 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
7907 *) Ugly workaround to get s_client and s_server working under Windows. The
7908 old code wouldn't work because it needed to select() on sockets and the
7909 tty (for keypresses and to see if data could be written). Win32 only
7910 supports select() on sockets so we select() with a 1s timeout on the
7911 sockets and then see if any characters are waiting to be read, if none
7912 are present then we retry, we also assume we can always write data to
7913 the tty. This isn't nice because the code then blocks until we've
7914 received a complete line of data and it is effectively polling the
7915 keyboard at 1s intervals: however it's quite a bit better than not
7916 working at all :-) A dedicated Windows application might handle this
7917 with an event loop for example.
7920 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
7921 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
7922 will be called when RSA_sign() and RSA_verify() are used. This is useful
7923 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
7924 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
7925 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
7926 This necessitated the support of an extra signature type NID_md5_sha1
7927 for SSL signatures and modifications to the SSL library to use it instead
7928 of calling RSA_public_decrypt() and RSA_private_encrypt().
7931 *) Add new -verify -CAfile and -CApath options to the crl program, these
7932 will lookup a CRL issuers certificate and verify the signature in a
7933 similar way to the verify program. Tidy up the crl program so it
7934 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
7935 less strict. It will now permit CRL extensions even if it is not
7936 a V2 CRL: this will allow it to tolerate some broken CRLs.
7939 *) Initialize all non-automatic variables each time one of the openssl
7940 sub-programs is started (this is necessary as they may be started
7941 multiple times from the "OpenSSL>" prompt).
7942 [Lennart Bang, Bodo Moeller]
7944 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
7945 removing all other RSA functionality (this is what NO_RSA does). This
7946 is so (for example) those in the US can disable those operations covered
7947 by the RSA patent while allowing storage and parsing of RSA keys and RSA
7951 *) Non-copying interface to BIO pairs.
7952 (still largely untested)
7955 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
7956 ASCII string. This was handled independently in various places before.
7959 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
7960 UTF8 strings a character at a time.
7963 *) Use client_version from client hello to select the protocol
7964 (s23_srvr.c) and for RSA client key exchange verification
7965 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
7968 *) Add various utility functions to handle SPKACs, these were previously
7969 handled by poking round in the structure internals. Added new function
7970 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
7971 print, verify and generate SPKACs. Based on an original idea from
7972 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
7975 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
7978 *) Allow the config file extension section to be overwritten on the
7979 command line. Based on an original idea from Massimiliano Pala
7980 <madwolf@comune.modena.it>. The new option is called -extensions
7981 and can be applied to ca, req and x509. Also -reqexts to override
7982 the request extensions in req and -crlexts to override the crl extensions
7986 *) Add new feature to the SPKAC handling in ca. Now you can include
7987 the same field multiple times by preceding it by "XXXX." for example:
7990 this is the same syntax as used in the req config file.
7993 *) Allow certificate extensions to be added to certificate requests. These
7994 are specified in a 'req_extensions' option of the req section of the
7995 config file. They can be printed out with the -text option to req but
7996 are otherwise ignored at present.
7999 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8000 data read consists of only the final block it would not decrypted because
8001 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8002 A misplaced 'break' also meant the decrypted final block might not be
8003 copied until the next read.
8006 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8007 a few extra parameters to the DH structure: these will be useful if
8008 for example we want the value of 'q' or implement X9.42 DH.
8011 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8012 provides hooks that allow the default DSA functions or functions on a
8013 "per key" basis to be replaced. This allows hardware acceleration and
8014 hardware key storage to be handled without major modification to the
8015 library. Also added low level modexp hooks and CRYPTO_EX structure and
8016 associated functions.
8019 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8020 as "read only": it can't be written to and the buffer it points to will
8021 not be freed. Reading from a read only BIO is much more efficient than
8022 a normal memory BIO. This was added because there are several times when
8023 an area of memory needs to be read from a BIO. The previous method was
8024 to create a memory BIO and write the data to it, this results in two
8025 copies of the data and an O(n^2) reading algorithm. There is a new
8026 function BIO_new_mem_buf() which creates a read only memory BIO from
8027 an area of memory. Also modified the PKCS#7 routines to use read only
8031 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8032 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8033 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8034 but a retry condition occured while trying to read the rest.
8037 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8038 NID_pkcs7_encrypted by default: this was wrong since this should almost
8039 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8040 the encrypted data type: this is a more sensible place to put it and it
8041 allows the PKCS#12 code to be tidied up that duplicated this
8045 *) Changed obj_dat.pl script so it takes its input and output files on
8046 the command line. This should avoid shell escape redirection problems
8050 *) Initial support for certificate extension requests, these are included
8051 in things like Xenroll certificate requests. Included functions to allow
8052 extensions to be obtained and added.
8055 *) -crlf option to s_client and s_server for sending newlines as
8056 CRLF (as required by many protocols).
8059 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8061 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8062 [Ralf S. Engelschall]
8064 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8065 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8067 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8071 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8072 DH parameters/keys (q is lost during that conversion, but the resulting
8073 DH parameters contain its length).
8075 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8076 much faster than DH_generate_parameters (which creates parameters
8077 where p = 2*q + 1), and also the smaller q makes DH computations
8078 much more efficient (160-bit exponentiation instead of 1024-bit
8079 exponentiation); so this provides a convenient way to support DHE
8080 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8081 utter importance to use
8082 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8084 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8085 when such DH parameters are used, because otherwise small subgroup
8086 attacks may become possible!
8089 *) Avoid memory leak in i2d_DHparams.
8092 *) Allow the -k option to be used more than once in the enc program:
8093 this allows the same encrypted message to be read by multiple recipients.
8096 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8097 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8098 it will always use the numerical form of the OID, even if it has a short
8102 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8103 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8104 otherwise bn_mod_exp was called. In the case of hardware keys for example
8105 no private key components need be present and it might store extra data
8106 in the RSA structure, which cannot be accessed from bn_mod_exp.
8107 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8108 private key operations.
8111 *) Added support for SPARC Linux.
8114 *) pem_password_cb function type incompatibly changed from
8115 typedef int pem_password_cb(char *buf, int size, int rwflag);
8117 ....(char *buf, int size, int rwflag, void *userdata);
8118 so that applications can pass data to their callbacks:
8119 The PEM[_ASN1]_{read,write}... functions and macros now take an
8120 additional void * argument, which is just handed through whenever
8121 the password callback is called.
8122 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8124 New function SSL_CTX_set_default_passwd_cb_userdata.
8126 Compatibility note: As many C implementations push function arguments
8127 onto the stack in reverse order, the new library version is likely to
8128 interoperate with programs that have been compiled with the old
8129 pem_password_cb definition (PEM_whatever takes some data that
8130 happens to be on the stack as its last argument, and the callback
8131 just ignores this garbage); but there is no guarantee whatsoever that
8134 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8135 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8136 problems not only on Windows, but also on some Unix platforms.
8137 To avoid problematic command lines, these definitions are now in an
8138 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8139 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8142 *) MIPS III/IV assembler module is reimplemented.
8145 *) More DES library cleanups: remove references to srand/rand and
8146 delete an unused file.
8149 *) Add support for the the free Netwide assembler (NASM) under Win32,
8150 since not many people have MASM (ml) and it can be hard to obtain.
8151 This is currently experimental but it seems to work OK and pass all
8152 the tests. Check out INSTALL.W32 for info.
8155 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8156 without temporary keys kept an extra copy of the server key,
8157 and connections with temporary keys did not free everything in case
8161 *) New function RSA_check_key and new openssl rsa option -check
8162 for verifying the consistency of RSA keys.
8163 [Ulf Moeller, Bodo Moeller]
8165 *) Various changes to make Win32 compile work:
8166 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
8167 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
8168 comparison" warnings.
8169 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
8172 *) Add a debugging option to PKCS#5 v2 key generation function: when
8173 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
8174 derived keys are printed to stderr.
8177 *) Copy the flags in ASN1_STRING_dup().
8178 [Roman E. Pavlov <pre@mo.msk.ru>]
8180 *) The x509 application mishandled signing requests containing DSA
8181 keys when the signing key was also DSA and the parameters didn't match.
8183 It was supposed to omit the parameters when they matched the signing key:
8184 the verifying software was then supposed to automatically use the CA's
8185 parameters if they were absent from the end user certificate.
8187 Omitting parameters is no longer recommended. The test was also
8188 the wrong way round! This was probably due to unusual behaviour in
8189 EVP_cmp_parameters() which returns 1 if the parameters match.
8190 This meant that parameters were omitted when they *didn't* match and
8191 the certificate was useless. Certificates signed with 'ca' didn't have
8193 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
8195 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
8196 The interface is as follows:
8197 Applications can use
8198 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
8199 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
8200 "off" is now the default.
8201 The library internally uses
8202 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
8203 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
8204 to disable memory-checking temporarily.
8206 Some inconsistent states that previously were possible (and were
8207 even the default) are now avoided.
8209 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
8210 with each memory chunk allocated; this is occasionally more helpful
8211 than just having a counter.
8213 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
8215 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
8219 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
8220 which largely parallels "options", but is for changing API behaviour,
8221 whereas "options" are about protocol behaviour.
8222 Initial "mode" flags are:
8224 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
8225 a single record has been written.
8226 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
8227 retries use the same buffer location.
8228 (But all of the contents must be
8232 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
8235 *) Fix problems with no-hmac etc.
8236 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
8238 *) New functions RSA_get_default_method(), RSA_set_method() and
8239 RSA_get_method(). These allows replacement of RSA_METHODs without having
8240 to mess around with the internals of an RSA structure.
8243 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
8244 Also really enable memory leak checks in openssl.c and in some
8246 [Chad C. Mulligan, Bodo Moeller]
8248 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
8249 up the length of negative integers. This has now been simplified to just
8250 store the length when it is first determined and use it later, rather
8251 than trying to keep track of where data is copied and updating it to
8253 [Steve Henson, reported by Brien Wheeler
8254 <bwheeler@authentica-security.com>]
8256 *) Add a new function PKCS7_signatureVerify. This allows the verification
8257 of a PKCS#7 signature but with the signing certificate passed to the
8258 function itself. This contrasts with PKCS7_dataVerify which assumes the
8259 certificate is present in the PKCS#7 structure. This isn't always the
8260 case: certificates can be omitted from a PKCS#7 structure and be
8261 distributed by "out of band" means (such as a certificate database).
8264 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
8265 function prototypes in pem.h, also change util/mkdef.pl to add the
8266 necessary function names.
8269 *) mk1mf.pl (used by Windows builds) did not properly read the
8270 options set by Configure in the top level Makefile, and Configure
8271 was not even able to write more than one option correctly.
8272 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
8275 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
8276 file to be loaded from a BIO or FILE pointer. The BIO version will
8277 for example allow memory BIOs to contain config info.
8280 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
8281 Whoever hopes to achieve shared-library compatibility across versions
8282 must use this, not the compile-time macro.
8283 (Exercise 0.9.4: Which is the minimum library version required by
8285 Note: All this applies only to multi-threaded programs, others don't
8289 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
8290 through a BIO pair triggered the default case, i.e.
8291 SSLerr(...,SSL_R_UNKNOWN_STATE).
8294 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
8295 can use the SSL library even if none of the specific BIOs is
8299 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
8300 for the encoded length.
8301 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
8303 *) Add initial documentation of the X509V3 functions.
8306 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
8307 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
8308 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
8309 secure PKCS#8 private key format with a high iteration count.
8312 *) Fix determination of Perl interpreter: A perl or perl5
8313 _directory_ in $PATH was also accepted as the interpreter.
8314 [Ralf S. Engelschall]
8316 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
8317 wrong with it but it was very old and did things like calling
8318 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
8322 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
8323 to use the new extension code.
8326 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
8327 with macros. This should make it easier to change their form, add extra
8328 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
8332 *) Add to configuration table a new entry that can specify an alternative
8333 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
8334 according to Mark Crispin <MRC@Panda.COM>.
8338 *) DES CBC did not update the IV. Weird.
8341 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
8342 Changing the behaviour of the former might break existing programs --
8343 where IV updating is needed, des_ncbc_encrypt can be used.
8346 *) When bntest is run from "make test" it drives bc to check its
8347 calculations, as well as internally checking them. If an internal check
8348 fails, it needs to cause bc to give a non-zero result or make test carries
8349 on without noticing the failure. Fixed.
8352 *) DES library cleanups.
8355 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
8356 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
8357 ciphers. NOTE: although the key derivation function has been verified
8358 against some published test vectors it has not been extensively tested
8359 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
8363 *) Instead of "mkdir -p", which is not fully portable, use new
8364 Perl script "util/mkdir-p.pl".
8367 *) Rewrite the way password based encryption (PBE) is handled. It used to
8368 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
8369 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
8370 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
8371 the 'parameter' field of the AlgorithmIdentifier is passed to the
8372 underlying key generation function so it must do its own ASN1 parsing.
8373 This has also changed the EVP_PBE_CipherInit() function which now has a
8374 'parameter' argument instead of literal salt and iteration count values
8375 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
8378 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
8379 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
8380 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
8381 KEY" because this clashed with PKCS#8 unencrypted string. Since this
8382 value was just used as a "magic string" and not used directly its
8383 value doesn't matter.
8386 *) Introduce some semblance of const correctness to BN. Shame C doesn't
8390 *) "linux-sparc64" configuration (ultrapenguin).
8391 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
8392 "linux-sparc" configuration.
8393 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
8395 *) config now generates no-xxx options for missing ciphers.
8398 *) Support the EBCDIC character set (work in progress).
8399 File ebcdic.c not yet included because it has a different license.
8400 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8402 *) Support BS2000/OSD-POSIX.
8403 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8405 *) Make callbacks for key generation use void * instead of char *.
8408 *) Make S/MIME samples compile (not yet tested).
8411 *) Additional typesafe stacks.
8414 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
8418 Changes between 0.9.3 and 0.9.3a [29 May 1999]
8420 *) New configuration variant "sco5-gcc".
8422 *) Updated some demos.
8423 [Sean O Riordain, Wade Scholine]
8425 *) Add missing BIO_free at exit of pkcs12 application.
8428 *) Fix memory leak in conf.c.
8431 *) Updates for Win32 to assembler version of MD5.
8434 *) Set #! path to perl in apps/der_chop to where we found it
8435 instead of using a fixed path.
8438 *) SHA library changes for irix64-mips4-cc.
8441 *) Improvements for VMS support.
8445 Changes between 0.9.2b and 0.9.3 [24 May 1999]
8447 *) Bignum library bug fix. IRIX 6 passes "make test" now!
8448 This also avoids the problems with SC4.2 and unpatched SC5.
8449 [Andy Polyakov <appro@fy.chalmers.se>]
8451 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
8452 These are required because of the typesafe stack would otherwise break
8453 existing code. If old code used a structure member which used to be STACK
8454 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
8455 sk_num or sk_value it would produce an error because the num, data members
8456 are not present in STACK_OF. Now it just produces a warning. sk_set
8457 replaces the old method of assigning a value to sk_value
8458 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
8459 that does this will no longer work (and should use sk_set instead) but
8460 this could be regarded as a "questionable" behaviour anyway.
8463 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
8464 correctly handle encrypted S/MIME data.
8467 *) Change type of various DES function arguments from des_cblock
8468 (which means, in function argument declarations, pointer to char)
8469 to des_cblock * (meaning pointer to array with 8 char elements),
8470 which allows the compiler to do more typechecking; it was like
8471 that back in SSLeay, but with lots of ugly casts.
8473 Introduce new type const_des_cblock.
8476 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
8477 problems: find RecipientInfo structure that matches recipient certificate
8478 and initialise the ASN1 structures properly based on passed cipher.
8481 *) Belatedly make the BN tests actually check the results.
8484 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
8485 to and from BNs: it was completely broken. New compilation option
8486 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
8487 key elements as negative integers.
8490 *) Reorganize and speed up MD5.
8491 [Andy Polyakov <appro@fy.chalmers.se>]
8494 [Richard Levitte <richard@levitte.org>]
8496 *) New option -out to asn1parse to allow the parsed structure to be
8497 output to a file. This is most useful when combined with the -strparse
8498 option to examine the output of things like OCTET STRINGS.
8501 *) Make SSL library a little more fool-proof by not requiring any longer
8502 that SSL_set_{accept,connect}_state be called before
8503 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
8504 in many applications because usually everything *appeared* to work as
8505 intended anyway -- now it really works as intended).
8508 *) Move openssl.cnf out of lib/.
8511 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
8512 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
8513 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
8514 [Ralf S. Engelschall]
8516 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
8517 handle PKCS#7 enveloped data properly.
8518 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
8520 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
8521 copying pointers. The cert_st handling is changed by this in
8522 various ways (and thus what used to be known as ctx->default_cert
8523 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
8524 any longer when s->cert does not give us what we need).
8525 ssl_cert_instantiate becomes obsolete by this change.
8526 As soon as we've got the new code right (possibly it already is?),
8527 we have solved a couple of bugs of the earlier code where s->cert
8528 was used as if it could not have been shared with other SSL structures.
8530 Note that using the SSL API in certain dirty ways now will result
8531 in different behaviour than observed with earlier library versions:
8532 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
8533 does not influence s as it used to.
8535 In order to clean up things more thoroughly, inside SSL_SESSION
8536 we don't use CERT any longer, but a new structure SESS_CERT
8537 that holds per-session data (if available); currently, this is
8538 the peer's certificate chain and, for clients, the server's certificate
8539 and temporary key. CERT holds only those values that can have
8540 meaningful defaults in an SSL_CTX.
8543 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
8544 from the internal representation. Various PKCS#7 fixes: remove some
8545 evil casts and set the enc_dig_alg field properly based on the signing
8549 *) Allow PKCS#12 password to be set from the command line or the
8550 environment. Let 'ca' get its config file name from the environment
8551 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
8555 *) Allow certificate policies extension to use an IA5STRING for the
8556 organization field. This is contrary to the PKIX definition but
8557 VeriSign uses it and IE5 only recognises this form. Document 'x509'
8561 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
8562 without disallowing inline assembler and the like for non-pedantic builds.
8565 *) Support Borland C++ builder.
8566 [Janez Jere <jj@void.si>, modified by Ulf Möller]
8571 *) SHA-1 cleanups and performance enhancements.
8572 [Andy Polyakov <appro@fy.chalmers.se>]
8574 *) Sparc v8plus assembler for the bignum library.
8575 [Andy Polyakov <appro@fy.chalmers.se>]
8577 *) Accept any -xxx and +xxx compiler options in Configure.
8580 *) Update HPUX configuration.
8583 *) Add missing sk_<type>_unshift() function to safestack.h
8584 [Ralf S. Engelschall]
8586 *) New function SSL_CTX_use_certificate_chain_file that sets the
8587 "extra_cert"s in addition to the certificate. (This makes sense
8588 only for "PEM" format files, as chains as a whole are not
8592 *) Support verify_depth from the SSL API.
8593 x509_vfy.c had what can be considered an off-by-one-error:
8594 Its depth (which was not part of the external interface)
8595 was actually counting the number of certificates in a chain;
8596 now it really counts the depth.
8599 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
8600 instead of X509err, which often resulted in confusing error
8601 messages since the error codes are not globally unique
8602 (e.g. an alleged error in ssl3_accept when a certificate
8603 didn't match the private key).
8605 *) New function SSL_CTX_set_session_id_context that allows to set a default
8606 value (so that you don't need SSL_set_session_id_context for each
8607 connection using the SSL_CTX).
8610 *) OAEP decoding bug fix.
8613 *) Support INSTALL_PREFIX for package builders, as proposed by
8617 *) New Configure options "threads" and "no-threads". For systems
8618 where the proper compiler options are known (currently Solaris
8619 and Linux), "threads" is the default.
8622 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
8625 *) Install various scripts to $(OPENSSLDIR)/misc, not to
8626 $(INSTALLTOP)/bin -- they shouldn't clutter directories
8627 such as /usr/local/bin.
8630 *) "make linux-shared" to build shared libraries.
8631 [Niels Poppe <niels@netbox.org>]
8633 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
8636 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
8637 extension adding in x509 utility.
8640 *) Remove NOPROTO sections and error code comments.
8643 *) Partial rewrite of the DEF file generator to now parse the ANSI
8647 *) New Configure options --prefix=DIR and --openssldir=DIR.
8650 *) Complete rewrite of the error code script(s). It is all now handled
8651 by one script at the top level which handles error code gathering,
8652 header rewriting and C source file generation. It should be much better
8653 than the old method: it now uses a modified version of Ulf's parser to
8654 read the ANSI prototypes in all header files (thus the old K&R definitions
8655 aren't needed for error creation any more) and do a better job of
8656 translating function codes into names. The old 'ASN1 error code imbedded
8657 in a comment' is no longer necessary and it doesn't use .err files which
8658 have now been deleted. Also the error code call doesn't have to appear all
8659 on one line (which resulted in some large lines...).
8662 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
8665 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
8666 0 (which usually indicates a closed connection), but continue reading.
8669 *) Fix some race conditions.
8672 *) Add support for CRL distribution points extension. Add Certificate
8673 Policies and CRL distribution points documentation.
8676 *) Move the autogenerated header file parts to crypto/opensslconf.h.
8679 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
8680 8 of keying material. Merlin has also confirmed interop with this fix
8681 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
8682 [Merlin Hughes <merlin@baltimore.ie>]
8684 *) Fix lots of warnings.
8685 [Richard Levitte <levitte@stacken.kth.se>]
8687 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
8688 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
8689 [Richard Levitte <levitte@stacken.kth.se>]
8691 *) Fix problems with sizeof(long) == 8.
8692 [Andy Polyakov <appro@fy.chalmers.se>]
8694 *) Change functions to ANSI C.
8697 *) Fix typos in error codes.
8698 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
8700 *) Remove defunct assembler files from Configure.
8703 *) SPARC v8 assembler BIGNUM implementation.
8704 [Andy Polyakov <appro@fy.chalmers.se>]
8706 *) Support for Certificate Policies extension: both print and set.
8707 Various additions to support the r2i method this uses.
8710 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
8711 return a const string when you are expecting an allocated buffer.
8714 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
8715 types DirectoryString and DisplayText.
8718 *) Add code to allow r2i extensions to access the configuration database,
8719 add an LHASH database driver and add several ctx helper functions.
8722 *) Fix an evil bug in bn_expand2() which caused various BN functions to
8723 fail when they extended the size of a BIGNUM.
8726 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
8727 support typesafe stack.
8730 *) Fix typo in SSL_[gs]et_options().
8731 [Nils Frostberg <nils@medcom.se>]
8733 *) Delete various functions and files that belonged to the (now obsolete)
8734 old X509V3 handling code.
8737 *) New Configure option "rsaref".
8740 *) Don't auto-generate pem.h.
8743 *) Introduce type-safe ASN.1 SETs.
8746 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
8747 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
8749 *) Introduce type-safe STACKs. This will almost certainly break lots of code
8750 that links with OpenSSL (well at least cause lots of warnings), but fear
8751 not: the conversion is trivial, and it eliminates loads of evil casts. A
8752 few STACKed things have been converted already. Feel free to convert more.
8753 In the fullness of time, I'll do away with the STACK type altogether.
8756 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
8757 specified in <certfile> by updating the entry in the index.txt file.
8758 This way one no longer has to edit the index.txt file manually for
8759 revoking a certificate. The -revoke option does the gory details now.
8760 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
8762 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
8763 `-text' option at all and this way the `-noout -text' combination was
8764 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
8765 [Ralf S. Engelschall]
8767 *) Make sure a corresponding plain text error message exists for the
8768 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
8769 verify callback function determined that a certificate was revoked.
8770 [Ralf S. Engelschall]
8772 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
8773 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
8774 all available cipers including rc5, which was forgotten until now.
8775 In order to let the testing shell script know which algorithms
8776 are available, a new (up to now undocumented) command
8777 "openssl list-cipher-commands" is used.
8780 *) Bugfix: s_client occasionally would sleep in select() when
8781 it should have checked SSL_pending() first.
8784 *) New functions DSA_do_sign and DSA_do_verify to provide access to
8785 the raw DSA values prior to ASN.1 encoding.
8788 *) Tweaks to Configure
8789 [Niels Poppe <niels@netbox.org>]
8791 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
8795 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
8798 *) New config option to avoid instructions that are illegal on the 80386.
8799 The default code is faster, but requires at least a 486.
8802 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
8803 SSL2_SERVER_VERSION (not used at all) macros, which are now the
8804 same as SSL2_VERSION anyway.
8807 *) New "-showcerts" option for s_client.
8810 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
8811 application. Various cleanups and fixes.
8814 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
8815 modify error routines to work internally. Add error codes and PBE init
8816 to library startup routines.
8819 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
8820 packing functions to asn1 and evp. Changed function names and error
8821 codes along the way.
8824 *) PKCS12 integration: and so it begins... First of several patches to
8825 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
8826 objects to objects.h
8829 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
8830 and display support for Thawte strong extranet extension.
8833 *) Add LinuxPPC support.
8834 [Jeff Dubrule <igor@pobox.org>]
8836 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
8837 bn_div_words in alpha.s.
8838 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
8840 *) Make sure the RSA OAEP test is skipped under -DRSAref because
8841 OAEP isn't supported when OpenSSL is built with RSAref.
8842 [Ulf Moeller <ulf@fitug.de>]
8844 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
8845 so they no longer are missing under -DNOPROTO.
8846 [Soren S. Jorvang <soren@t.dk>]
8849 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
8851 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
8852 doesn't work when the session is reused. Coming soon!
8855 *) Fix a security hole, that allows sessions to be reused in the wrong
8856 context thus bypassing client cert protection! All software that uses
8857 client certs and session caches in multiple contexts NEEDS PATCHING to
8858 allow session reuse! A fuller solution is in the works.
8859 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
8861 *) Some more source tree cleanups (removed obsolete files
8862 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
8863 permission on "config" script to be executable) and a fix for the INSTALL
8865 [Ulf Moeller <ulf@fitug.de>]
8867 *) Remove some legacy and erroneous uses of malloc, free instead of
8869 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
8871 *) Make rsa_oaep_test return non-zero on error.
8872 [Ulf Moeller <ulf@fitug.de>]
8874 *) Add support for native Solaris shared libraries. Configure
8875 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
8876 if someone would make that last step automatic.
8877 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
8879 *) ctx_size was not built with the right compiler during "make links". Fixed.
8882 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
8883 except NULL ciphers". This means the default cipher list will no longer
8884 enable NULL ciphers. They need to be specifically enabled e.g. with
8885 the string "DEFAULT:eNULL".
8888 *) Fix to RSA private encryption routines: if p < q then it would
8889 occasionally produce an invalid result. This will only happen with
8890 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
8893 *) Be less restrictive and allow also `perl util/perlpath.pl
8894 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
8895 because this way one can also use an interpreter named `perl5' (which is
8896 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
8897 installed as `perl').
8898 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8900 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
8901 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8903 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
8904 advapi32.lib to Win32 build and change the pem test comparision
8905 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
8906 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
8907 and crypto/des/ede_cbcm_enc.c.
8910 *) DES quad checksum was broken on big-endian architectures. Fixed.
8913 *) Comment out two functions in bio.h that aren't implemented. Fix up the
8914 Win32 test batch file so it (might) work again. The Win32 test batch file
8915 is horrible: I feel ill....
8918 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
8919 in e_os.h. Audit of header files to check ANSI and non ANSI
8920 sections: 10 functions were absent from non ANSI section and not exported
8921 from Windows DLLs. Fixed up libeay.num for new functions.
8924 *) Make `openssl version' output lines consistent.
8925 [Ralf S. Engelschall]
8927 *) Fix Win32 symbol export lists for BIO functions: Added
8928 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
8929 to ms/libeay{16,32}.def.
8930 [Ralf S. Engelschall]
8932 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
8933 fine under Unix and passes some trivial tests I've now added. But the
8934 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
8935 added to make sure no one expects that this stuff really works in the
8936 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
8937 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
8939 [Ralf S. Engelschall]
8941 *) Fix the generation of two part addresses in perl.
8942 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
8944 *) Add config entry for Linux on MIPS.
8945 [John Tobey <jtobey@channel1.com>]
8947 *) Make links whenever Configure is run, unless we are on Windoze.
8950 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
8951 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
8955 *) Add a useful kludge to allow package maintainers to specify compiler and
8956 other platforms details on the command line without having to patch the
8957 Configure script everytime: One now can use ``perl Configure
8958 <id>:<details>'', i.e. platform ids are allowed to have details appended
8959 to them (seperated by colons). This is treated as there would be a static
8960 pre-configured entry in Configure's %table under key <id> with value
8961 <details> and ``perl Configure <id>'' is called. So, when you want to
8962 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
8963 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
8964 now, which overrides the FreeBSD-elf entry on-the-fly.
8965 [Ralf S. Engelschall]
8967 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
8970 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
8971 on the `perl Configure ...' command line. This way one can compile
8972 OpenSSL libraries with Position Independent Code (PIC) which is needed
8973 for linking it into DSOs.
8974 [Ralf S. Engelschall]
8976 *) Remarkably, export ciphers were totally broken and no-one had noticed!
8980 *) Cleaned up the LICENSE document: The official contact for any license
8981 questions now is the OpenSSL core team under openssl-core@openssl.org.
8982 And add a paragraph about the dual-license situation to make sure people
8983 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
8984 to the OpenSSL toolkit.
8985 [Ralf S. Engelschall]
8987 *) General source tree makefile cleanups: Made `making xxx in yyy...'
8988 display consistent in the source tree and replaced `/bin/rm' by `rm'.
8989 Additonally cleaned up the `make links' target: Remove unnecessary
8990 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
8991 to speed processing and no longer clutter the display with confusing
8992 stuff. Instead only the actually done links are displayed.
8993 [Ralf S. Engelschall]
8995 *) Permit null encryption ciphersuites, used for authentication only. It used
8996 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
8997 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9001 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9002 signed attributes when verifying signatures (this would break them),
9003 the detached data encoding was wrong and public keys obtained using
9004 X509_get_pubkey() weren't freed.
9007 *) Add text documentation for the BUFFER functions. Also added a work around
9008 to a Win95 console bug. This was triggered by the password read stuff: the
9009 last character typed gets carried over to the next fread(). If you were
9010 generating a new cert request using 'req' for example then the last
9011 character of the passphrase would be CR which would then enter the first
9015 *) Added the new `Includes OpenSSL Cryptography Software' button as
9016 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9017 button and can be used by applications based on OpenSSL to show the
9018 relationship to the OpenSSL project.
9019 [Ralf S. Engelschall]
9021 *) Remove confusing variables in function signatures in files
9022 ssl/ssl_lib.c and ssl/ssl.h.
9023 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9025 *) Don't install bss_file.c under PREFIX/include/
9026 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9028 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9029 functions that return function pointers and has support for NT specific
9030 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9031 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9032 unsigned to signed types: this was killing the Win32 compile.
9035 *) Add new certificate file to stack functions,
9036 SSL_add_dir_cert_subjects_to_stack() and
9037 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9038 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9039 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9040 This means that Apache-SSL and similar packages don't have to mess around
9041 to add as many CAs as they want to the preferred list.
9044 *) Experiment with doxygen documentation. Currently only partially applied to
9046 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9047 openssl.doxy as the configuration file.
9050 *) Get rid of remaining C++-style comments which strict C compilers hate.
9051 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9053 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9054 compiled in by default: it has problems with large keys.
9057 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9058 DH private keys and/or callback functions which directly correspond to
9059 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9060 is needed for applications which have to configure certificates on a
9061 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9063 For the RSA certificate situation is makes no difference, but
9064 for the DSA certificate situation this fixes the "no shared cipher"
9065 problem where the OpenSSL cipher selection procedure failed because the
9066 temporary keys were not overtaken from the context and the API provided
9067 no way to reconfigure them.
9068 The new functions now let applications reconfigure the stuff and they
9069 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9070 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9071 non-public-API function ssl_cert_instantiate() is used as a helper
9072 function and also to reduce code redundancy inside ssl_rsa.c.
9073 [Ralf S. Engelschall]
9075 *) Move s_server -dcert and -dkey options out of the undocumented feature
9076 area because they are useful for the DSA situation and should be
9077 recognized by the users.
9078 [Ralf S. Engelschall]
9080 *) Fix the cipher decision scheme for export ciphers: the export bits are
9081 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9082 SSL_EXP_MASK. So, the original variable has to be used instead of the
9083 already masked variable.
9084 [Richard Levitte <levitte@stacken.kth.se>]
9086 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9087 [Richard Levitte <levitte@stacken.kth.se>]
9089 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9090 from `int' to `unsigned int' because it's a length and initialized by
9091 EVP_DigestFinal() which expects an `unsigned int *'.
9092 [Richard Levitte <levitte@stacken.kth.se>]
9094 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9095 script. Instead use the usual Shell->Perl transition trick.
9096 [Ralf S. Engelschall]
9098 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9099 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9100 -noout -modulus' as it's already the case for `openssl rsa -noout
9101 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9102 currently the public key is printed (a decision which was already done by
9103 `openssl dsa -modulus' in the past) which serves a similar purpose.
9104 Additionally the NO_RSA no longer completely removes the whole -modulus
9105 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9107 [Ralf S. Engelschall]
9109 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9110 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9111 [Arne Ansper <arne@ats.cyber.ee>]
9113 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9114 to be added. Now both 'req' and 'ca' can use new objects defined in the
9118 *) Add cool BIO that does syslog (or event log on NT).
9119 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9121 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9122 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9123 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9124 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9127 *) Add preliminary config info for new extension code.
9130 *) Make RSA_NO_PADDING really use no padding.
9131 [Ulf Moeller <ulf@fitug.de>]
9133 *) Generate errors when private/public key check is done.
9136 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9137 for some CRL extensions and new objects added.
9140 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9141 key usage extension and fuller support for authority key id.
9144 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9145 padding method for RSA, which is recommended for new applications in PKCS
9146 #1 v2.0 (RFC 2437, October 1998).
9147 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9148 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9149 against Bleichbacher's attack on RSA.
9150 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9153 *) Updates to the new SSL compression code
9154 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9156 *) Fix so that the version number in the master secret, when passed
9157 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9158 (because the server will not accept higher), that the version number
9159 is 0x03,0x01, not 0x03,0x00
9160 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9162 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9163 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9164 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
9167 *) Support for RAW extensions where an arbitrary extension can be
9168 created by including its DER encoding. See apps/openssl.cnf for
9172 *) Make sure latest Perl versions don't interpret some generated C array
9173 code as Perl array code in the crypto/err/err_genc.pl script.
9174 [Lars Weber <3weber@informatik.uni-hamburg.de>]
9176 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
9177 not many people have the assembler. Various Win32 compilation fixes and
9178 update to the INSTALL.W32 file with (hopefully) more accurate Win32
9182 *) Modify configure script 'Configure' to automatically create crypto/date.h
9183 file under Win32 and also build pem.h from pem.org. New script
9184 util/mkfiles.pl to create the MINFO file on environments that can't do a
9185 'make files': perl util/mkfiles.pl >MINFO should work.
9188 *) Major rework of DES function declarations, in the pursuit of correctness
9189 and purity. As a result, many evil casts evaporated, and some weirdness,
9190 too. You may find this causes warnings in your code. Zapping your evil
9191 casts will probably fix them. Mostly.
9194 *) Fix for a typo in asn1.h. Bug fix to object creation script
9195 obj_dat.pl. It considered a zero in an object definition to mean
9196 "end of object": none of the objects in objects.h have any zeros
9197 so it wasn't spotted.
9198 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
9200 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
9201 Masking (CBCM). In the absence of test vectors, the best I have been able
9202 to do is check that the decrypt undoes the encrypt, so far. Send me test
9203 vectors if you have them.
9206 *) Correct calculation of key length for export ciphers (too much space was
9207 allocated for null ciphers). This has not been tested!
9210 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
9211 message is now correct (it understands "crypto" and "ssl" on its
9212 command line). There is also now an "update" option. This will update
9213 the util/ssleay.num and util/libeay.num files with any new functions.
9215 perl util/mkdef.pl crypto ssl update
9216 it will update them.
9219 *) Overhauled the Perl interface (perl/*):
9220 - ported BN stuff to OpenSSL's different BN library
9221 - made the perl/ source tree CVS-aware
9222 - renamed the package from SSLeay to OpenSSL (the files still contain
9223 their history because I've copied them in the repository)
9224 - removed obsolete files (the test scripts will be replaced
9225 by better Test::Harness variants in the future)
9226 [Ralf S. Engelschall]
9228 *) First cut for a very conservative source tree cleanup:
9229 1. merge various obsolete readme texts into doc/ssleay.txt
9230 where we collect the old documents and readme texts.
9231 2. remove the first part of files where I'm already sure that we no
9232 longer need them because of three reasons: either they are just temporary
9233 files which were left by Eric or they are preserved original files where
9234 I've verified that the diff is also available in the CVS via "cvs diff
9235 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
9236 the crypto/md/ stuff).
9237 [Ralf S. Engelschall]
9239 *) More extension code. Incomplete support for subject and issuer alt
9240 name, issuer and authority key id. Change the i2v function parameters
9241 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
9242 what that's for :-) Fix to ASN1 macro which messed up
9243 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
9246 *) Preliminary support for ENUMERATED type. This is largely copied from the
9250 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
9251 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9253 *) Make sure `make rehash' target really finds the `openssl' program.
9254 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9256 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
9257 like to hear about it if this slows down other processors.
9260 *) Add CygWin32 platform information to Configure script.
9261 [Alan Batie <batie@aahz.jf.intel.com>]
9263 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
9264 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
9266 *) New program nseq to manipulate netscape certificate sequences
9269 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
9273 *) Fixes to BN code. Previously the default was to define BN_RECURSION
9274 but the BN code had some problems that would cause failures when
9275 doing certificate verification and some other functions.
9276 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9278 *) Add ASN1 and PEM code to support netscape certificate sequences.
9281 *) Add ASN1 and PEM code to support netscape certificate sequences.
9284 *) Add several PKIX and private extended key usage OIDs.
9287 *) Modify the 'ca' program to handle the new extension code. Modify
9288 openssl.cnf for new extension format, add comments.
9291 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
9292 and add a sample to openssl.cnf so req -x509 now adds appropriate
9296 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
9297 error code, add initial support to X509_print() and x509 application.
9300 *) Takes a deep breath and start addding X509 V3 extension support code. Add
9301 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
9302 stuff is currently isolated and isn't even compiled yet.
9305 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
9306 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
9307 Removed the versions check from X509 routines when loading extensions:
9308 this allows certain broken certificates that don't set the version
9309 properly to be processed.
9312 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
9313 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
9314 can still be regenerated with "make depend".
9317 *) Spelling mistake in C version of CAST-128.
9318 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
9320 *) Changes to the error generation code. The perl script err-code.pl
9321 now reads in the old error codes and retains the old numbers, only
9322 adding new ones if necessary. It also only changes the .err files if new
9323 codes are added. The makefiles have been modified to only insert errors
9324 when needed (to avoid needlessly modifying header files). This is done
9325 by only inserting errors if the .err file is newer than the auto generated
9326 C file. To rebuild all the error codes from scratch (the old behaviour)
9327 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
9328 or delete all the .err files.
9331 *) CAST-128 was incorrectly implemented for short keys. The C version has
9332 been fixed, but is untested. The assembler versions are also fixed, but
9333 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
9334 to regenerate it if needed.
9335 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
9336 Hagino <itojun@kame.net>]
9338 *) File was opened incorrectly in randfile.c.
9339 [Ulf Möller <ulf@fitug.de>]
9341 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
9342 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
9343 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
9344 al: it's just almost always a UTCTime. Note this patch adds new error
9345 codes so do a "make errors" if there are problems.
9348 *) Correct Linux 1 recognition in config.
9349 [Ulf Möller <ulf@fitug.de>]
9351 *) Remove pointless MD5 hash when using DSA keys in ca.
9352 [Anonymous <nobody@replay.com>]
9354 *) Generate an error if given an empty string as a cert directory. Also
9355 generate an error if handed NULL (previously returned 0 to indicate an
9356 error, but didn't set one).
9357 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
9359 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
9362 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
9363 parameters. This was causing a warning which killed off the Win32 compile.
9366 *) Remove C++ style comments from crypto/bn/bn_local.h.
9367 [Neil Costigan <neil.costigan@celocom.com>]
9369 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
9370 based on a text string, looking up short and long names and finally
9371 "dot" format. The "dot" format stuff didn't work. Added new function
9372 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
9373 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
9374 OID is not part of the table.
9377 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
9378 X509_LOOKUP_by_alias().
9381 *) Sort openssl functions by name.
9384 *) Get the gendsa program working (hopefully) and add it to app list. Remove
9385 encryption from sample DSA keys (in case anyone is interested the password
9389 *) Make _all_ *_free functions accept a NULL pointer.
9390 [Frans Heymans <fheymans@isaserver.be>]
9392 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
9394 [Anonymous <nobody@replay.com>]
9396 *) s_server should send the CAfile as acceptable CAs, not its own cert.
9397 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9399 *) Don't blow it for numeric -newkey arguments to apps/req.
9400 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9402 *) Temp key "for export" tests were wrong in s3_srvr.c.
9403 [Anonymous <nobody@replay.com>]
9405 *) Add prototype for temp key callback functions
9406 SSL_CTX_set_tmp_{rsa,dh}_callback().
9409 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
9410 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
9413 *) X509_name_add_entry() freed the wrong thing after an error.
9414 [Arne Ansper <arne@ats.cyber.ee>]
9416 *) rsa_eay.c would attempt to free a NULL context.
9417 [Arne Ansper <arne@ats.cyber.ee>]
9419 *) BIO_s_socket() had a broken should_retry() on Windoze.
9420 [Arne Ansper <arne@ats.cyber.ee>]
9422 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
9423 [Arne Ansper <arne@ats.cyber.ee>]
9425 *) Make sure the already existing X509_STORE->depth variable is initialized
9426 in X509_STORE_new(), but document the fact that this variable is still
9427 unused in the certificate verification process.
9428 [Ralf S. Engelschall]
9430 *) Fix the various library and apps files to free up pkeys obtained from
9431 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
9434 *) Fix reference counting in X509_PUBKEY_get(). This makes
9435 demos/maurice/example2.c work, amongst others, probably.
9436 [Steve Henson and Ben Laurie]
9438 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
9439 `openssl' and second, the shortcut symlinks for the `openssl <command>'
9440 are no longer created. This way we have a single and consistent command
9441 line interface `openssl <command>', similar to `cvs <command>'.
9442 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
9444 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
9445 BIT STRING wrapper always have zero unused bits.
9448 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
9451 *) Make the top-level INSTALL documentation easier to understand.
9454 *) Makefiles updated to exit if an error occurs in a sub-directory
9455 make (including if user presses ^C) [Paul Sutton]
9457 *) Make Montgomery context stuff explicit in RSA data structure.
9460 *) Fix build order of pem and err to allow for generated pem.h.
9463 *) Fix renumbering bug in X509_NAME_delete_entry().
9466 *) Enhanced the err-ins.pl script so it makes the error library number
9467 global and can add a library name. This is needed for external ASN1 and
9468 other error libraries.
9471 *) Fixed sk_insert which never worked properly.
9474 *) Fix ASN1 macros so they can handle indefinite length construted
9475 EXPLICIT tags. Some non standard certificates use these: they can now
9479 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
9480 into a single doc/ssleay.txt bundle. This way the information is still
9481 preserved but no longer messes up this directory. Now it's new room for
9482 the new set of documenation files.
9483 [Ralf S. Engelschall]
9485 *) SETs were incorrectly DER encoded. This was a major pain, because they
9486 shared code with SEQUENCEs, which aren't coded the same. This means that
9487 almost everything to do with SETs or SEQUENCEs has either changed name or
9488 number of arguments.
9489 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
9491 *) Fix test data to work with the above.
9494 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
9495 was already fixed by Eric for 0.9.1 it seems.
9496 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
9498 *) Autodetect FreeBSD3.
9501 *) Fix various bugs in Configure. This affects the following platforms:
9505 unixware-2.0-pentium
9509 *) Eliminate generated files from CVS. Reorder tests to regenerate files
9510 before they are needed.
9513 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
9517 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
9519 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
9520 changed SSLeay to OpenSSL in version strings.
9521 [Ralf S. Engelschall]
9523 *) Some fixups to the top-level documents.
9526 *) Fixed the nasty bug where rsaref.h was not found under compile-time
9527 because the symlink to include/ was missing.
9528 [Ralf S. Engelschall]
9530 *) Incorporated the popular no-RSA/DSA-only patches
9531 which allow to compile a RSA-free SSLeay.
9532 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
9534 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
9535 when "ssleay" is still not found.
9536 [Ralf S. Engelschall]
9538 *) Added more platforms to Configure: Cray T3E, HPUX 11,
9539 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
9541 *) Updated the README file.
9542 [Ralf S. Engelschall]
9544 *) Added various .cvsignore files in the CVS repository subdirs
9545 to make a "cvs update" really silent.
9546 [Ralf S. Engelschall]
9548 *) Recompiled the error-definition header files and added
9549 missing symbols to the Win32 linker tables.
9550 [Ralf S. Engelschall]
9552 *) Cleaned up the top-level documents;
9553 o new files: CHANGES and LICENSE
9554 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
9555 o merged COPYRIGHT into LICENSE
9556 o removed obsolete TODO file
9557 o renamed MICROSOFT to INSTALL.W32
9558 [Ralf S. Engelschall]
9560 *) Removed dummy files from the 0.9.1b source tree:
9561 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
9562 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
9563 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
9564 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
9565 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
9566 [Ralf S. Engelschall]
9568 *) Added various platform portability fixes.
9571 *) The Genesis of the OpenSSL rpject:
9572 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
9573 Young and Tim J. Hudson created while they were working for C2Net until
9575 [The OpenSSL Project]
9578 Changes between 0.9.0b and 0.9.1b [not released]
9580 *) Updated a few CA certificates under certs/
9583 *) Changed some BIGNUM api stuff.
9586 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
9587 DGUX x86, Linux Alpha, etc.
9590 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
9591 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
9595 *) Add -strparse option to asn1pars program which parses nested
9597 [Dr Stephen Henson <shenson@bigfoot.com>]
9599 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
9602 *) DSA fix for "ca" program.
9605 *) Added "-genkey" option to "dsaparam" program.
9608 *) Added RIPE MD160 (rmd160) message digest.
9611 *) Added -a (all) option to "ssleay version" command.
9614 *) Added PLATFORM define which is the id given to Configure.
9617 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
9620 *) Extended the ASN.1 parser routines.
9623 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
9626 *) Added a BN_CTX to the BN library.
9629 *) Fixed the weak key values in DES library
9632 *) Changed API in EVP library for cipher aliases.
9635 *) Added support for RC2/64bit cipher.
9638 *) Converted the lhash library to the crypto/mem.c functions.
9641 *) Added more recognized ASN.1 object ids.
9644 *) Added more RSA padding checks for SSL/TLS.
9647 *) Added BIO proxy/filter functionality.
9650 *) Added extra_certs to SSL_CTX which can be used
9651 send extra CA certificates to the client in the CA cert chain sending
9652 process. It can be configured with SSL_CTX_add_extra_chain_cert().
9655 *) Now Fortezza is denied in the authentication phase because
9656 this is key exchange mechanism is not supported by SSLeay at all.
9659 *) Additional PKCS1 checks.
9662 *) Support the string "TLSv1" for all TLS v1 ciphers.
9665 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
9666 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
9669 *) Fixed a few memory leaks.
9672 *) Fixed various code and comment typos.
9675 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
9676 bytes sent in the client random.
9677 [Edward Bishop <ebishop@spyglass.com>]