5 Changes between 1.0.0g and 1.0.0h [xx XXX xxxx]
7 *) Fix CVE-2011-4619: make sure we really are receiving a
8 client hello before rejecting multiple SGC restarts. Thanks to
9 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
12 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
14 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
15 Thanks to Antonio Martin, Enterprise Secure Access Research and
16 Development, Cisco Systems, Inc. for discovering this bug and
17 preparing a fix. (CVE-2012-0050)
20 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
22 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
23 of the Vaudenay padding oracle attack on CBC mode encryption
24 which enables an efficient plaintext recovery attack against
25 the OpenSSL implementation of DTLS. Their attack exploits timing
26 differences arising during decryption processing. A research
27 paper describing this attack can be found at:
28 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
29 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
30 Security Group at Royal Holloway, University of London
31 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
32 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
33 for preparing the fix. (CVE-2011-4108)
34 [Robin Seggelmann, Michael Tuexen]
36 *) Clear bytes used for block padding of SSL 3.0 records.
38 [Adam Langley (Google)]
40 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
41 Kadianakis <desnacked@gmail.com> for discovering this issue and
42 Adam Langley for preparing the fix. (CVE-2011-4619)
43 [Adam Langley (Google)]
45 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
46 [Andrey Kulikov <amdeich@gmail.com>]
48 *) Prevent malformed RFC3779 data triggering an assertion failure.
49 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
50 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
51 [Rob Austein <sra@hactrn.net>]
53 *) Improved PRNG seeding for VOS.
54 [Paul Green <Paul.Green@stratus.com>]
56 *) Fix ssl_ciph.c set-up race.
57 [Adam Langley (Google)]
59 *) Fix spurious failures in ecdsatest.c.
60 [Emilia Käsper (Google)]
62 *) Fix the BIO_f_buffer() implementation (which was mixing different
63 interpretations of the '..._len' fields).
64 [Adam Langley (Google)]
66 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
67 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
68 threads won't reuse the same blinding coefficients.
70 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
71 lock to call BN_BLINDING_invert_ex, and avoids one use of
72 BN_BLINDING_update for each BN_BLINDING structure (previously,
73 the last update always remained unused).
74 [Emilia Käsper (Google)]
76 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
77 [Bob Buckholz (Google)]
79 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
81 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
82 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
83 [Kaspar Brand <ossl@velox.ch>]
85 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
86 for multi-threaded use of ECDH. (CVE-2011-3210)
87 [Adam Langley (Google)]
89 *) Fix x509_name_ex_d2i memory leak on bad inputs.
92 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
93 signature public key algorithm by using OID xref utilities instead.
94 Before this you could only use some ECC ciphersuites with SHA1 only.
97 *) Add protection against ECDSA timing attacks as mentioned in the paper
98 by Billy Bob Brumley and Nicola Tuveri, see:
100 http://eprint.iacr.org/2011/232.pdf
102 [Billy Bob Brumley and Nicola Tuveri]
104 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
106 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
107 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
109 *) Fix bug in string printing code: if *any* escaping is enabled we must
110 escape the escape character (backslash) or the resulting string is
114 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
116 *) Disable code workaround for ancient and obsolete Netscape browsers
117 and servers: an attacker can use it in a ciphersuite downgrade attack.
118 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
121 *) Fixed J-PAKE implementation error, originally discovered by
122 Sebastien Martini, further info and confirmation from Stefan
123 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
126 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
128 *) Fix extension code to avoid race conditions which can result in a buffer
129 overrun vulnerability: resumed sessions must not be modified as they can
130 be shared by multiple threads. CVE-2010-3864
133 *) Fix WIN32 build system to correctly link an ENGINE directory into
137 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
139 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
141 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
143 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
145 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
146 context. The operation can be customised via the ctrl mechanism in
147 case ENGINEs want to include additional functionality.
150 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
153 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
154 output hashes compatible with older versions of OpenSSL.
155 [Willy Weisz <weisz@vcpc.univie.ac.at>]
157 *) Fix compression algorithm handling: if resuming a session use the
158 compression algorithm of the resumed session instead of determining
159 it from client hello again. Don't allow server to change algorithm.
162 *) Add load_crls() function to apps tidying load_certs() too. Add option
163 to verify utility to allow additional CRLs to be included.
166 *) Update OCSP request code to permit adding custom headers to the request:
167 some responders need this.
170 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
172 [Julia Lawall <julia@diku.dk>]
174 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
175 needlessly dereferenced structures, used obsolete functions and
176 didn't handle all updated verify codes correctly.
179 *) Disable MD2 in the default configuration.
182 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
183 indicate the initial BIO being pushed or popped. This makes it possible
184 to determine whether the BIO is the one explicitly called or as a result
185 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
186 it handles reference counts correctly and doesn't zero out the I/O bio
187 when it is not being explicitly popped. WARNING: applications which
188 included workarounds for the old buggy behaviour will need to be modified
189 or they could free up already freed BIOs.
192 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
193 renaming to all platforms (within the 0.9.8 branch, this was
194 done conditionally on Netware platforms to avoid a name clash).
195 [Guenter <lists@gknw.net>]
197 *) Add ECDHE and PSK support to DTLS.
198 [Michael Tuexen <tuexen@fh-muenster.de>]
200 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
204 *) Add "missing" function EVP_MD_flags() (without this the only way to
205 retrieve a digest flags is by accessing the structure directly. Update
206 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
207 or cipher is registered as in the "from" argument. Print out all
208 registered digests in the dgst usage message instead of manually
209 attempting to work them out.
212 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
213 this allows the use of compression and extensions. Change default cipher
214 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
215 by default unless an application cipher string requests it.
218 *) Alter match criteria in PKCS12_parse(). It used to try to use local
219 key ids to find matching certificates and keys but some PKCS#12 files
220 don't follow the (somewhat unwritten) rules and this strategy fails.
221 Now just gather all certificates together and the first private key
222 then look for the first certificate that matches the key.
225 *) Support use of registered digest and cipher names for dgst and cipher
226 commands instead of having to add each one as a special case. So now
233 openssl dgst -sha256 foo
235 and this works for ENGINE based algorithms too.
239 *) Update Gost ENGINE to support parameter files.
240 [Victor B. Wagner <vitus@cryptocom.ru>]
242 *) Support GeneralizedTime in ca utility.
243 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
245 *) Enhance the hash format used for certificate directory links. The new
246 form uses the canonical encoding (meaning equivalent names will work
247 even if they aren't identical) and uses SHA1 instead of MD5. This form
248 is incompatible with the older format and as a result c_rehash should
249 be used to rebuild symbolic links.
252 *) Make PKCS#8 the default write format for private keys, replacing the
253 traditional format. This form is standardised, more secure and doesn't
254 include an implicit MD5 dependency.
257 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
258 committed to OpenSSL should pass this lot as a minimum.
261 *) Add session ticket override functionality for use by EAP-FAST.
262 [Jouni Malinen <j@w1.fi>]
264 *) Modify HMAC functions to return a value. Since these can be implemented
265 in an ENGINE errors can occur.
268 *) Type-checked OBJ_bsearch_ex.
271 *) Type-checked OBJ_bsearch. Also some constification necessitated
272 by type-checking. Still to come: TXT_DB, bsearch(?),
273 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
277 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
278 seconds to a tm structure directly, instead of going through OS
279 specific date routines. This avoids any issues with OS routines such
280 as the year 2038 bug. New *_adj() functions for ASN1 time structures
281 and X509_time_adj_ex() to cover the extended range. The existing
282 X509_time_adj() is still usable and will no longer have any date issues.
285 *) Delta CRL support. New use deltas option which will attempt to locate
286 and search any appropriate delta CRLs available.
288 This work was sponsored by Google.
291 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
292 code and add additional score elements. Validate alternate CRL paths
293 as part of the CRL checking and indicate a new error "CRL path validation
294 error" in this case. Applications wanting additional details can use
295 the verify callback and check the new "parent" field. If this is not
296 NULL CRL path validation is taking place. Existing applications wont
297 see this because it requires extended CRL support which is off by
300 This work was sponsored by Google.
303 *) Support for freshest CRL extension.
305 This work was sponsored by Google.
308 *) Initial indirect CRL support. Currently only supported in the CRLs
309 passed directly and not via lookup. Process certificate issuer
310 CRL entry extension and lookup CRL entries by bother issuer name
311 and serial number. Check and process CRL issuer entry in IDP extension.
313 This work was sponsored by Google.
316 *) Add support for distinct certificate and CRL paths. The CRL issuer
317 certificate is validated separately in this case. Only enabled if
318 an extended CRL support flag is set: this flag will enable additional
319 CRL functionality in future.
321 This work was sponsored by Google.
324 *) Add support for policy mappings extension.
326 This work was sponsored by Google.
329 *) Fixes to pathlength constraint, self issued certificate handling,
330 policy processing to align with RFC3280 and PKITS tests.
332 This work was sponsored by Google.
335 *) Support for name constraints certificate extension. DN, email, DNS
336 and URI types are currently supported.
338 This work was sponsored by Google.
341 *) To cater for systems that provide a pointer-based thread ID rather
342 than numeric, deprecate the current numeric thread ID mechanism and
343 replace it with a structure and associated callback type. This
344 mechanism allows a numeric "hash" to be extracted from a thread ID in
345 either case, and on platforms where pointers are larger than 'long',
346 mixing is done to help ensure the numeric 'hash' is usable even if it
347 can't be guaranteed unique. The default mechanism is to use "&errno"
348 as a pointer-based thread ID to distinguish between threads.
350 Applications that want to provide their own thread IDs should now use
351 CRYPTO_THREADID_set_callback() to register a callback that will call
352 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
354 Note that ERR_remove_state() is now deprecated, because it is tied
355 to the assumption that thread IDs are numeric. ERR_remove_state(0)
356 to free the current thread's error state should be replaced by
357 ERR_remove_thread_state(NULL).
359 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
360 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
361 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
362 application was previously providing a numeric thread callback that
363 was inappropriate for distinguishing threads, then uniqueness might
364 have been obtained with &errno that happened immediately in the
365 intermediate development versions of OpenSSL; this is no longer the
366 case, the numeric thread callback will now override the automatic use
368 [Geoff Thorpe, with help from Bodo Moeller]
370 *) Initial support for different CRL issuing certificates. This covers a
371 simple case where the self issued certificates in the chain exist and
372 the real CRL issuer is higher in the existing chain.
374 This work was sponsored by Google.
377 *) Removed effectively defunct crypto/store from the build.
380 *) Revamp of STACK to provide stronger type-checking. Still to come:
381 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
382 ASN1_STRING, CONF_VALUE.
385 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
386 RAM on SSL connections. This option can save about 34k per idle SSL.
389 *) Revamp of LHASH to provide stronger type-checking. Still to come:
390 STACK, TXT_DB, bsearch, qsort.
393 *) Initial support for Cryptographic Message Syntax (aka CMS) based
394 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
395 support for data, signedData, compressedData, digestedData and
396 encryptedData, envelopedData types included. Scripts to check against
397 RFC4134 examples draft and interop and consistency checks of many
398 content types and variants.
401 *) Add options to enc utility to support use of zlib compression BIO.
404 *) Extend mk1mf to support importing of options and assembly language
405 files from Configure script, currently only included in VC-WIN32.
406 The assembly language rules can now optionally generate the source
407 files from the associated perl scripts.
410 *) Implement remaining functionality needed to support GOST ciphersuites.
411 Interop testing has been performed using CryptoPro implementations.
412 [Victor B. Wagner <vitus@cryptocom.ru>]
414 *) s390x assembler pack.
417 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
421 *) Implement Opaque PRF Input TLS extension as specified in
422 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
423 official specification yet and no extension type assignment by
424 IANA exists, this extension (for now) will have to be explicitly
425 enabled when building OpenSSL by providing the extension number
426 to use. For example, specify an option
428 -DTLSEXT_TYPE_opaque_prf_input=0x9527
430 to the "config" or "Configure" script to enable the extension,
431 assuming extension number 0x9527 (which is a completely arbitrary
432 and unofficial assignment based on the MD5 hash of the Internet
433 Draft). Note that by doing so, you potentially lose
434 interoperability with other TLS implementations since these might
435 be using the same extension number for other purposes.
437 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
438 opaque PRF input value to use in the handshake. This will create
439 an interal copy of the length-'len' string at 'src', and will
440 return non-zero for success.
442 To get more control and flexibility, provide a callback function
445 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
446 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
450 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
453 Callback function 'cb' will be called in handshakes, and is
454 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
455 Argument 'arg' is for application purposes (the value as given to
456 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
457 be provided to the callback function). The callback function
458 has to return non-zero to report success: usually 1 to use opaque
459 PRF input just if possible, or 2 to enforce use of the opaque PRF
460 input. In the latter case, the library will abort the handshake
461 if opaque PRF input is not successfully negotiated.
463 Arguments 'peerinput' and 'len' given to the callback function
464 will always be NULL and 0 in the case of a client. A server will
465 see the client's opaque PRF input through these variables if
466 available (NULL and 0 otherwise). Note that if the server
467 provides an opaque PRF input, the length must be the same as the
468 length of the client's opaque PRF input.
470 Note that the callback function will only be called when creating
471 a new session (session resumption can resume whatever was
472 previously negotiated), and will not be called in SSL 2.0
473 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
474 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
475 for applications that need to enforce opaque PRF input.
479 *) Update ssl code to support digests other than SHA1+MD5 for handshake
482 [Victor B. Wagner <vitus@cryptocom.ru>]
484 *) Add RFC4507 support to OpenSSL. This includes the corrections in
485 RFC4507bis. The encrypted ticket format is an encrypted encoded
486 SSL_SESSION structure, that way new session features are automatically
489 If a client application caches session in an SSL_SESSION structure
490 support is transparent because tickets are now stored in the encoded
493 The SSL_CTX structure automatically generates keys for ticket
494 protection in servers so again support should be possible
495 with no application modification.
497 If a client or server wishes to disable RFC4507 support then the option
498 SSL_OP_NO_TICKET can be set.
500 Add a TLS extension debugging callback to allow the contents of any client
501 or server extensions to be examined.
503 This work was sponsored by Google.
506 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
507 OpenSSL should now compile cleanly on gcc 4.2
508 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
510 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
511 support including streaming MAC support: this is required for GOST
513 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
515 *) Add option -stream to use PKCS#7 streaming in smime utility. New
516 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
517 to output in BER and PEM format.
520 *) Experimental support for use of HMAC via EVP_PKEY interface. This
521 allows HMAC to be handled via the EVP_DigestSign*() interface. The
522 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
523 ENGINE support for HMAC keys which are unextractable. New -mac and
524 -macopt options to dgst utility.
527 *) New option -sigopt to dgst utility. Update dgst to use
528 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
529 alternative signing paramaters such as X9.31 or PSS in the dgst
533 *) Change ssl_cipher_apply_rule(), the internal function that does
534 the work each time a ciphersuite string requests enabling
535 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
536 removing ("!foo+bar") a class of ciphersuites: Now it maintains
537 the order of disabled ciphersuites such that those ciphersuites
538 that most recently went from enabled to disabled not only stay
539 in order with respect to each other, but also have higher priority
540 than other disabled ciphersuites the next time ciphersuites are
543 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
544 the same ciphersuites as with "HIGH" alone, but in a specific
545 order where the PSK ciphersuites come first (since they are the
546 most recently disabled ciphersuites when "HIGH" is parsed).
548 Also, change ssl_create_cipher_list() (using this new
549 funcionality) such that between otherwise identical
550 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
554 *) Change ssl_create_cipher_list() so that it automatically
555 arranges the ciphersuites in reasonable order before starting
556 to process the rule string. Thus, the definition for "DEFAULT"
557 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
558 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
559 This makes it much easier to arrive at a reasonable default order
560 in applications for which anonymous ciphers are OK (meaning
561 that you can't actually use DEFAULT).
562 [Bodo Moeller; suggested by Victor Duchovni]
564 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
565 processing) into multiple integers instead of setting
566 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
567 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
568 (These masks as well as the individual bit definitions are hidden
569 away into the non-exported interface ssl/ssl_locl.h, so this
570 change to the definition of the SSL_CIPHER structure shouldn't
571 affect applications.) This give us more bits for each of these
572 categories, so there is no longer a need to coagulate AES128 and
573 AES256 into a single algorithm bit, and to coagulate Camellia128
574 and Camellia256 into a single algorithm bit, which has led to all
577 Thus, among other things, the kludge introduced in 0.9.7m and
578 0.9.8e for masking out AES256 independently of AES128 or masking
579 out Camellia256 independently of AES256 is not needed here in 0.9.9.
581 With the change, we also introduce new ciphersuite aliases that
582 so far were missing: "AES128", "AES256", "CAMELLIA128", and
586 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
587 Use the leftmost N bytes of the signature input if the input is
588 larger than the prime q (with N being the size in bytes of q).
591 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
592 it yet and it is largely untested.
595 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
598 *) Initial incomplete changes to avoid need for function casts in OpenSSL
599 some compilers (gcc 4.2 and later) reject their use. Safestack is
600 reimplemented. Update ASN1 to avoid use of legacy functions.
603 *) Win32/64 targets are linked with Winsock2.
606 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
607 to external functions. This can be used to increase CRL handling
608 efficiency especially when CRLs are very large by (for example) storing
609 the CRL revoked certificates in a database.
612 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
613 new CRLs added to a directory can be used. New command line option
614 -verify_return_error to s_client and s_server. This causes real errors
615 to be returned by the verify callback instead of carrying on no matter
616 what. This reflects the way a "real world" verify callback would behave.
619 *) GOST engine, supporting several GOST algorithms and public key formats.
620 Kindly donated by Cryptocom.
623 *) Partial support for Issuing Distribution Point CRL extension. CRLs
624 partitioned by DP are handled but no indirect CRL or reason partitioning
625 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
626 selected via a scoring technique which handles IDP and AKID in CRLs.
629 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
630 will ultimately be used for all verify operations: this will remove the
631 X509_STORE dependency on certificate verification and allow alternative
632 lookup methods. X509_STORE based implementations of these two callbacks.
635 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
636 Modify get_crl() to find a valid (unexpired) CRL if possible.
639 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
640 this would be called X509_CRL_cmp() but that name is already used by
641 a function that just compares CRL issuer names. Cache several CRL
642 extensions in X509_CRL structure and cache CRLDP in X509.
645 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
646 this maps equivalent X509_NAME structures into a consistent structure.
647 Name comparison can then be performed rapidly using memcmp().
650 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
654 *) Allow digests to supply their own micalg string for S/MIME type using
655 the ctrl EVP_MD_CTRL_MICALG.
658 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
659 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
660 ctrl. It can then customise the structure before and/or after signing
664 *) New function OBJ_add_sigid() to allow application defined signature OIDs
665 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
666 to free up any added signature OIDs.
669 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
670 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
671 digest and cipher tables. New options added to openssl utility:
672 list-message-digest-algorithms and list-cipher-algorithms.
675 *) Change the array representation of binary polynomials: the list
676 of degrees of non-zero coefficients is now terminated with -1.
677 Previously it was terminated with 0, which was also part of the
678 value; thus, the array representation was not applicable to
679 polynomials where t^0 has coefficient zero. This change makes
680 the array representation useful in a more general context.
683 *) Various modifications and fixes to SSL/TLS cipher string
684 handling. For ECC, the code now distinguishes between fixed ECDH
685 with RSA certificates on the one hand and with ECDSA certificates
686 on the other hand, since these are separate ciphersuites. The
687 unused code for Fortezza ciphersuites has been removed.
689 For consistency with EDH, ephemeral ECDH is now called "EECDH"
690 (not "ECDHE"). For consistency with the code for DH
691 certificates, use of ECDH certificates is now considered ECDH
692 authentication, not RSA or ECDSA authentication (the latter is
693 merely the CA's signing algorithm and not actively used in the
696 The temporary ciphersuite alias "ECCdraft" is no longer
697 available, and ECC ciphersuites are no longer excluded from "ALL"
698 and "DEFAULT". The following aliases now exist for RFC 4492
699 ciphersuites, most of these by analogy with the DH case:
701 kECDHr - ECDH cert, signed with RSA
702 kECDHe - ECDH cert, signed with ECDSA
703 kECDH - ECDH cert (signed with either RSA or ECDSA)
704 kEECDH - ephemeral ECDH
705 ECDH - ECDH cert or ephemeral ECDH
711 AECDH - anonymous ECDH
712 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
716 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
717 Use correct micalg parameters depending on digest(s) in signed message.
720 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
721 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
724 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
725 an engine to register a method. Add ENGINE lookups for methods and
726 functional reference processing.
729 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
730 EVP_{Sign,Verify}* which allow an application to customise the signature
734 *) New -resign option to smime utility. This adds one or more signers
735 to an existing PKCS#7 signedData structure. Also -md option to use an
736 alternative message digest algorithm for signing.
739 *) Tidy up PKCS#7 routines and add new functions to make it easier to
740 create PKCS7 structures containing multiple signers. Update smime
741 application to support multiple signers.
744 *) New -macalg option to pkcs12 utility to allow setting of an alternative
748 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
749 Reorganize PBE internals to lookup from a static table using NIDs,
750 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
751 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
752 PRF which will be automatically used with PBES2.
755 *) Replace the algorithm specific calls to generate keys in "req" with the
759 *) Update PKCS#7 enveloped data routines to use new API. This is now
760 supported by any public key method supporting the encrypt operation. A
761 ctrl is added to allow the public key algorithm to examine or modify
762 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
766 *) Add a ctrl to asn1 method to allow a public key algorithm to express
767 a default digest type to use. In most cases this will be SHA1 but some
768 algorithms (such as GOST) need to specify an alternative digest. The
769 return value indicates how strong the prefernce is 1 means optional and
770 2 is mandatory (that is it is the only supported type). Modify
771 ASN1_item_sign() to accept a NULL digest argument to indicate it should
772 use the default md. Update openssl utilities to use the default digest
773 type for signing if it is not explicitly indicated.
776 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
777 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
778 signing method from the key type. This effectively removes the link
779 between digests and public key types.
782 *) Add an OID cross reference table and utility functions. Its purpose is to
783 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
784 rsaEncryption. This will allow some of the algorithm specific hackery
785 needed to use the correct OID to be removed.
788 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
789 structures for PKCS7_sign(). They are now set up by the relevant public
793 *) Add provisional EC pkey method with support for ECDSA and ECDH.
796 *) Add support for key derivation (agreement) in the API, DH method and
800 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
801 public and private key formats. As a side effect these add additional
802 command line functionality not previously available: DSA signatures can be
803 generated and verified using pkeyutl and DH key support and generation in
808 [Oliver Tappe <zooey@hirschkaefer.de>]
810 *) New make target "install_html_docs" installs HTML renditions of the
812 [Oliver Tappe <zooey@hirschkaefer.de>]
814 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
815 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
816 support key and parameter generation and add initial key generation
817 functionality for RSA.
820 *) Add functions for main EVP_PKEY_method operations. The undocumented
821 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
822 EVP_PKEY_{encrypt,decrypt}_old.
825 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
826 key API, doesn't do much yet.
829 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
830 public key algorithms. New option to openssl utility:
831 "list-public-key-algorithms" to print out info.
834 *) Implement the Supported Elliptic Curves Extension for
835 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
838 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
839 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
842 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
843 utilities such as rsa, dsa, dsaparam etc except they process any key
847 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
848 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
849 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
853 *) Initial support for pluggable public key ASN1.
854 De-spaghettify the public key ASN1 handling. Move public and private
855 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
856 algorithm specific handling to a single module within the relevant
857 algorithm directory. Add functions to allow (near) opaque processing
858 of public and private key structures.
861 *) Implement the Supported Point Formats Extension for
862 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
865 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
866 for the psk identity [hint] and the psk callback functions to the
867 SSL_SESSION, SSL and SSL_CTX structure.
870 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
874 SSL_CTX_use_psk_identity_hint
875 SSL_get_psk_identity_hint
877 SSL_use_psk_identity_hint
879 [Mika Kousa and Pasi Eronen of Nokia Corporation]
881 *) Add RFC 3161 compliant time stamp request creation, response generation
882 and response verification functionality.
883 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
885 *) Add initial support for TLS extensions, specifically for the server_name
886 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
887 have new members for a host name. The SSL data structure has an
888 additional member SSL_CTX *initial_ctx so that new sessions can be
889 stored in that context to allow for session resumption, even after the
890 SSL has been switched to a new SSL_CTX in reaction to a client's
891 server_name extension.
893 New functions (subject to change):
896 SSL_get_servername_type()
899 New CTRL codes and macros (subject to change):
901 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
902 - SSL_CTX_set_tlsext_servername_callback()
903 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
904 - SSL_CTX_set_tlsext_servername_arg()
905 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
907 openssl s_client has a new '-servername ...' option.
909 openssl s_server has new options '-servername_host ...', '-cert2 ...',
910 '-key2 ...', '-servername_fatal' (subject to change). This allows
911 testing the HostName extension for a specific single host name ('-cert'
912 and '-key' remain fallbacks for handshakes without HostName
913 negotiation). If the unrecogninzed_name alert has to be sent, this by
914 default is a warning; it becomes fatal with the '-servername_fatal'
917 [Peter Sylvester, Remy Allais, Christophe Renou]
919 *) Whirlpool hash implementation is added.
922 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
923 bn(64,32). Because of instruction set limitations it doesn't have
924 any negative impact on performance. This was done mostly in order
925 to make it possible to share assembler modules, such as bn_mul_mont
926 implementations, between 32- and 64-bit builds without hassle.
929 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
930 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
934 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
935 dedicated Montgomery multiplication procedure, is introduced.
936 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
937 "64-bit" performance on certain 32-bit targets.
940 *) New option SSL_OP_NO_COMP to disable use of compression selectively
941 in SSL structures. New SSL ctrl to set maximum send fragment size.
942 Save memory by seeting the I/O buffer sizes dynamically instead of
943 using the maximum available value.
946 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
947 in addition to the text details.
950 *) Very, very preliminary EXPERIMENTAL support for printing of general
951 ASN1 structures. This currently produces rather ugly output and doesn't
952 handle several customised structures at all.
955 *) Integrated support for PVK file format and some related formats such
956 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
957 these in the 'rsa' and 'dsa' utilities.
960 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
963 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
964 place for the (very old) "NETSCAPE" format certificates which are now
965 handled using new ASN1 code equivalents.
968 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
969 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
970 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
973 *) Modify CRL distribution points extension code to print out previously
974 unsupported fields. Enhance extension setting code to allow setting of
978 *) Add print and set support for Issuing Distribution Point CRL extension.
981 *) Change 'Configure' script to enable Camellia by default.
984 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
986 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
987 Thanks to Antonio Martin, Enterprise Secure Access Research and
988 Development, Cisco Systems, Inc. for discovering this bug and
989 preparing a fix. (CVE-2012-0050)
992 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
994 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
995 of the Vaudenay padding oracle attack on CBC mode encryption
996 which enables an efficient plaintext recovery attack against
997 the OpenSSL implementation of DTLS. Their attack exploits timing
998 differences arising during decryption processing. A research
999 paper describing this attack can be found at:
1000 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1001 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1002 Security Group at Royal Holloway, University of London
1003 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1004 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1005 for preparing the fix. (CVE-2011-4108)
1006 [Robin Seggelmann, Michael Tuexen]
1008 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
1009 [Ben Laurie, Kasper <ekasper@google.com>]
1011 *) Clear bytes used for block padding of SSL 3.0 records.
1013 [Adam Langley (Google)]
1015 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1016 Kadianakis <desnacked@gmail.com> for discovering this issue and
1017 Adam Langley for preparing the fix. (CVE-2011-4619)
1018 [Adam Langley (Google)]
1020 *) Prevent malformed RFC3779 data triggering an assertion failure.
1021 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1022 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1023 [Rob Austein <sra@hactrn.net>]
1025 *) Fix ssl_ciph.c set-up race.
1026 [Adam Langley (Google)]
1028 *) Fix spurious failures in ecdsatest.c.
1029 [Emilia Käsper (Google)]
1031 *) Fix the BIO_f_buffer() implementation (which was mixing different
1032 interpretations of the '..._len' fields).
1033 [Adam Langley (Google)]
1035 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1036 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1037 threads won't reuse the same blinding coefficients.
1039 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1040 lock to call BN_BLINDING_invert_ex, and avoids one use of
1041 BN_BLINDING_update for each BN_BLINDING structure (previously,
1042 the last update always remained unused).
1043 [Emilia Käsper (Google)]
1045 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1046 for multi-threaded use of ECDH.
1047 [Adam Langley (Google)]
1049 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1052 *) Add protection against ECDSA timing attacks as mentioned in the paper
1053 by Billy Bob Brumley and Nicola Tuveri, see:
1055 http://eprint.iacr.org/2011/232.pdf
1057 [Billy Bob Brumley and Nicola Tuveri]
1059 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
1061 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1062 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1064 *) Fix bug in string printing code: if *any* escaping is enabled we must
1065 escape the escape character (backslash) or the resulting string is
1069 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
1071 *) Disable code workaround for ancient and obsolete Netscape browsers
1072 and servers: an attacker can use it in a ciphersuite downgrade attack.
1073 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1076 *) Fixed J-PAKE implementation error, originally discovered by
1077 Sebastien Martini, further info and confirmation from Stefan
1078 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1081 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
1083 *) Fix extension code to avoid race conditions which can result in a buffer
1084 overrun vulnerability: resumed sessions must not be modified as they can
1085 be shared by multiple threads. CVE-2010-3864
1088 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
1091 *) Don't reencode certificate when calculating signature: cache and use
1092 the original encoding instead. This makes signature verification of
1093 some broken encodings work correctly.
1096 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
1097 is also one of the inputs.
1098 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1100 *) Don't repeatedly append PBE algorithms to table if they already exist.
1101 Sort table on each new add. This effectively makes the table read only
1102 after all algorithms are added and subsequent calls to PKCS12_pbe_add
1106 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
1108 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
1111 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
1112 access or freeing data twice (CVE-2010-0742)
1113 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
1115 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
1116 common in certificates and some applications which only call
1117 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
1121 Reduce copying into .apps and .test in makevms.com
1122 Don't try to use blank CA certificate in CA.com
1123 Allow use of C files from original directories in maketests.com
1124 [Steven M. Schweda" <sms@antinode.info>]
1126 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1128 *) When rejecting SSL/TLS records due to an incorrect version number, never
1129 update s->server with a new major version number. As of
1130 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1131 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1132 the previous behavior could result in a read attempt at NULL when
1133 receiving specific incorrect SSL/TLS records once record payload
1134 protection is active. (CVE-2010-0740)
1135 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1137 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1138 could be crashed if the relevant tables were not present (e.g. chrooted).
1139 [Tomas Hoger <thoger@redhat.com>]
1141 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1143 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1144 [Martin Olsson, Neel Mehta]
1146 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1147 accommodate for stack sorting, always a write lock!).
1150 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1151 excessive delays in the RAND_poll(): over a minute. As a workaround
1152 include a time check in the inner Heap32Next loop too.
1155 *) The code that handled flushing of data in SSL/TLS originally used the
1156 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1157 the problem outlined in PR#1949. The fix suggested there however can
1158 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1159 of Apache). So instead simplify the code to flush unconditionally.
1160 This should be fine since flushing with no data to flush is a no op.
1163 *) Handle TLS versions 2.0 and later properly and correctly use the
1164 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1165 off ancient servers have a habit of sticking around for a while...
1168 *) Modify compression code so it frees up structures without using the
1169 ex_data callbacks. This works around a problem where some applications
1170 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1171 restarting) then use compression (e.g. SSL with compression) later.
1172 This results in significant per-connection memory leaks and
1173 has caused some security issues including CVE-2008-1678 and
1177 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1178 change when encrypting or decrypting.
1181 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1182 connect and renegotiate with servers which do not support RI.
1183 Until RI is more widely deployed this option is enabled by default.
1186 *) Add "missing" ssl ctrls to clear options and mode.
1189 *) If client attempts to renegotiate and doesn't support RI respond with
1190 a no_renegotiation alert as required by RFC5746. Some renegotiating
1191 TLS clients will continue a connection gracefully when they receive
1192 the alert. Unfortunately OpenSSL mishandled this alert and would hang
1193 waiting for a server hello which it will never receive. Now we treat a
1194 received no_renegotiation alert as a fatal error. This is because
1195 applications requesting a renegotiation might well expect it to succeed
1196 and would have no code in place to handle the server denying it so the
1197 only safe thing to do is to terminate the connection.
1200 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
1201 peer supports secure renegotiation and 0 otherwise. Print out peer
1202 renegotiation support in s_client/s_server.
1205 *) Replace the highly broken and deprecated SPKAC certification method with
1206 the updated NID creation version. This should correctly handle UTF8.
1209 *) Implement RFC5746. Re-enable renegotiation but require the extension
1210 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
1211 turns out to be a bad idea. It has been replaced by
1212 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
1213 SSL_CTX_set_options(). This is really not recommended unless you
1214 know what you are doing.
1215 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
1217 *) Fixes to stateless session resumption handling. Use initial_ctx when
1218 issuing and attempting to decrypt tickets in case it has changed during
1219 servername handling. Use a non-zero length session ID when attempting
1220 stateless session resumption: this makes it possible to determine if
1221 a resumption has occurred immediately after receiving server hello
1222 (several places in OpenSSL subtly assume this) instead of later in
1226 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
1227 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
1228 fixes for a few places where the return code is not checked
1230 [Julia Lawall <julia@diku.dk>]
1232 *) Add --strict-warnings option to Configure script to include devteam
1233 warnings in other configurations.
1236 *) Add support for --libdir option and LIBDIR variable in makefiles. This
1237 makes it possible to install openssl libraries in locations which
1238 have names other than "lib", for example "/usr/lib64" which some
1240 [Steve Henson, based on patch from Jeremy Utley]
1242 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
1243 X690 8.9.12 and can produce some misleading textual output of OIDs.
1244 [Steve Henson, reported by Dan Kaminsky]
1246 *) Delete MD2 from algorithm tables. This follows the recommendation in
1247 several standards that it is not used in new applications due to
1248 several cryptographic weaknesses. For binary compatibility reasons
1249 the MD2 API is still compiled in by default.
1252 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
1256 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
1257 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
1259 [Guenter <lists@gknw.net>]
1261 *) Fix the server certificate chain building code to use X509_verify_cert(),
1262 it used to have an ad-hoc builder which was unable to cope with anything
1263 other than a simple chain.
1264 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
1266 *) Don't check self signed certificate signatures in X509_verify_cert()
1267 by default (a flag can override this): it just wastes time without
1268 adding any security. As a useful side effect self signed root CAs
1269 with non-FIPS digests are now usable in FIPS mode.
1272 *) In dtls1_process_out_of_seq_message() the check if the current message
1273 is already buffered was missing. For every new message was memory
1274 allocated, allowing an attacker to perform an denial of service attack
1275 with sending out of seq handshake messages until there is no memory
1276 left. Additionally every future messege was buffered, even if the
1277 sequence number made no sense and would be part of another handshake.
1278 So only messages with sequence numbers less than 10 in advance will be
1279 buffered. (CVE-2009-1378)
1280 [Robin Seggelmann, discovered by Daniel Mentz]
1282 *) Records are buffered if they arrive with a future epoch to be
1283 processed after finishing the corresponding handshake. There is
1284 currently no limitation to this buffer allowing an attacker to perform
1285 a DOS attack with sending records with future epochs until there is no
1286 memory left. This patch adds the pqueue_size() function to detemine
1287 the size of a buffer and limits the record buffer to 100 entries.
1289 [Robin Seggelmann, discovered by Daniel Mentz]
1291 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
1292 parent structure is freed. (CVE-2009-1379)
1295 *) Handle non-blocking I/O properly in SSL_shutdown() call.
1296 [Darryl Miles <darryl-mailinglists@netbauds.net>]
1299 [Ilya O. <vrghost@gmail.com>]
1301 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
1303 *) Disable renegotiation completely - this fixes a severe security
1304 problem (CVE-2009-3555) at the cost of breaking all
1305 renegotiation. Renegotiation can be re-enabled by setting
1306 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
1307 run-time. This is really not recommended unless you know what
1311 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
1313 *) Don't set val to NULL when freeing up structures, it is freed up by
1314 underlying code. If sizeof(void *) > sizeof(long) this can result in
1315 zeroing past the valid field. (CVE-2009-0789)
1316 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
1318 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
1319 checked correctly. This would allow some invalid signed attributes to
1320 appear to verify correctly. (CVE-2009-0591)
1321 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1323 *) Reject UniversalString and BMPString types with invalid lengths. This
1324 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
1325 a legal length. (CVE-2009-0590)
1328 *) Set S/MIME signing as the default purpose rather than setting it
1329 unconditionally. This allows applications to override it at the store
1333 *) Permit restricted recursion of ASN1 strings. This is needed in practice
1334 to handle some structures.
1337 *) Improve efficiency of mem_gets: don't search whole buffer each time
1339 [Jeremy Shapiro <jnshapir@us.ibm.com>]
1341 *) New -hex option for openssl rand.
1344 *) Print out UTF8String and NumericString when parsing ASN1.
1347 *) Support NumericString type for name components.
1350 *) Allow CC in the environment to override the automatically chosen
1351 compiler. Note that nothing is done to ensure flags work with the
1355 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
1357 *) Properly check EVP_VerifyFinal() and similar return values
1359 [Ben Laurie, Bodo Moeller, Google Security Team]
1361 *) Enable TLS extensions by default.
1364 *) Allow the CHIL engine to be loaded, whether the application is
1365 multithreaded or not. (This does not release the developer from the
1366 obligation to set up the dynamic locking callbacks.)
1367 [Sander Temme <sander@temme.net>]
1369 *) Use correct exit code if there is an error in dgst command.
1370 [Steve Henson; problem pointed out by Roland Dirlewanger]
1372 *) Tweak Configure so that you need to say "experimental-jpake" to enable
1373 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
1376 *) Add experimental JPAKE support, including demo authentication in
1377 s_client and s_server.
1380 *) Set the comparison function in v3_addr_canonize().
1381 [Rob Austein <sra@hactrn.net>]
1383 *) Add support for XMPP STARTTLS in s_client.
1384 [Philip Paeps <philip@freebsd.org>]
1386 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
1387 to ensure that even with this option, only ciphersuites in the
1388 server's preference list will be accepted. (Note that the option
1389 applies only when resuming a session, so the earlier behavior was
1390 just about the algorithm choice for symmetric cryptography.)
1393 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
1395 *) Fix NULL pointer dereference if a DTLS server received
1396 ChangeCipherSpec as first record (CVE-2009-1386).
1399 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
1400 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
1403 *) The fix in 0.9.8c that supposedly got rid of unsafe
1404 double-checked locking was incomplete for RSA blinding,
1405 addressing just one layer of what turns out to have been
1406 doubly unsafe triple-checked locking.
1408 So now fix this for real by retiring the MONT_HELPER macro
1409 in crypto/rsa/rsa_eay.c.
1411 [Bodo Moeller; problem pointed out by Marius Schilder]
1413 *) Various precautionary measures:
1415 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
1417 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
1418 (NB: This would require knowledge of the secret session ticket key
1419 to exploit, in which case you'd be SOL either way.)
1421 - Change bn_nist.c so that it will properly handle input BIGNUMs
1422 outside the expected range.
1424 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
1427 [Neel Mehta, Bodo Moeller]
1429 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
1430 the load fails. Useful for distros.
1431 [Ben Laurie and the FreeBSD team]
1433 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
1436 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
1439 *) Expand ENGINE to support engine supplied SSL client certificate functions.
1441 This work was sponsored by Logica.
1444 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
1445 keystores. Support for SSL/TLS client authentication too.
1446 Not compiled unless enable-capieng specified to Configure.
1448 This work was sponsored by Logica.
1451 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
1452 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
1453 attribute creation routines such as certifcate requests and PKCS#12
1457 Changes between 0.9.8g and 0.9.8h [28 May 2008]
1459 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
1460 handshake which could lead to a cilent crash as found using the
1461 Codenomicon TLS test suite (CVE-2008-1672)
1462 [Steve Henson, Mark Cox]
1464 *) Fix double free in TLS server name extensions which could lead to
1465 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
1468 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
1470 Clear the error queue to ensure that error entries left from
1471 older function calls do not interfere with the correct operation.
1472 [Lutz Jaenicke, Erik de Castro Lopo]
1474 *) Remove root CA certificates of commercial CAs:
1476 The OpenSSL project does not recommend any specific CA and does not
1477 have any policy with respect to including or excluding any CA.
1478 Therefore it does not make any sense to ship an arbitrary selection
1479 of root CA certificates with the OpenSSL software.
1482 *) RSA OAEP patches to fix two separate invalid memory reads.
1483 The first one involves inputs when 'lzero' is greater than
1484 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
1485 before the beginning of from). The second one involves inputs where
1486 the 'db' section contains nothing but zeroes (there is a one-byte
1487 invalid read after the end of 'db').
1488 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1490 *) Partial backport from 0.9.9-dev:
1492 Introduce bn_mul_mont (dedicated Montgomery multiplication
1493 procedure) as a candidate for BIGNUM assembler implementation.
1494 While 0.9.9-dev uses assembler for various architectures, only
1495 x86_64 is available by default here in the 0.9.8 branch, and
1496 32-bit x86 is available through a compile-time setting.
1498 To try the 32-bit x86 assembler implementation, use Configure
1499 option "enable-montasm" (which exists only for this backport).
1501 As "enable-montasm" for 32-bit x86 disclaims code stability
1502 anyway, in this constellation we activate additional code
1503 backported from 0.9.9-dev for further performance improvements,
1504 namely BN_from_montgomery_word. (To enable this otherwise,
1505 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
1507 [Andy Polyakov (backport partially by Bodo Moeller)]
1509 *) Add TLS session ticket callback. This allows an application to set
1510 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
1511 values. This is useful for key rollover for example where several key
1512 sets may exist with different names.
1515 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
1516 This was broken until now in 0.9.8 releases, such that the only way
1517 a registered ENGINE could be used (assuming it initialises
1518 successfully on the host) was to explicitly set it as the default
1519 for the relevant algorithms. This is in contradiction with 0.9.7
1520 behaviour and the documentation. With this fix, when an ENGINE is
1521 registered into a given algorithm's table of implementations, the
1522 'uptodate' flag is reset so that auto-discovery will be used next
1523 time a new context for that algorithm attempts to select an
1525 [Ian Lister (tweaked by Geoff Thorpe)]
1527 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
1528 implemention in the following ways:
1530 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
1533 Lack of BER streaming support means one pass streaming processing is
1534 only supported if data is detached: setting the streaming flag is
1535 ignored for embedded content.
1537 CMS support is disabled by default and must be explicitly enabled
1538 with the enable-cms configuration option.
1541 *) Update the GMP engine glue to do direct copies between BIGNUM and
1542 mpz_t when openssl and GMP use the same limb size. Otherwise the
1543 existing "conversion via a text string export" trick is still used.
1544 [Paul Sheer <paulsheer@gmail.com>]
1546 *) Zlib compression BIO. This is a filter BIO which compressed and
1547 uncompresses any data passed through it.
1550 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
1551 RFC3394 compatible AES key wrapping.
1554 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
1555 sets string data without copying. X509_ALGOR_set0() and
1556 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
1557 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
1558 from an X509_ATTRIBUTE structure optionally checking it occurs only
1559 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
1563 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
1564 to get the expected BN_FLG_CONSTTIME behavior.
1565 [Bodo Moeller (Google)]
1569 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
1570 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
1571 - added some more tests to do_tests.pl
1572 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
1573 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
1574 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
1575 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
1576 - various changes to netware.pl to enable gcc-cross builds on Win32
1578 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
1579 - various changes to fix missing prototype warnings
1580 - fixed x86nasm.pl to create correct asm files for NASM COFF output
1581 - added AES, WHIRLPOOL and CPUID assembler code to build files
1582 - added missing AES assembler make rules to mk1mf.pl
1583 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
1584 [Guenter Knauf <eflash@gmx.net>]
1586 *) Implement certificate status request TLS extension defined in RFC3546.
1587 A client can set the appropriate parameters and receive the encoded
1588 OCSP response via a callback. A server can query the supplied parameters
1589 and set the encoded OCSP response in the callback. Add simplified examples
1590 to s_client and s_server.
1593 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
1595 *) Fix various bugs:
1596 + Binary incompatibility of ssl_ctx_st structure
1597 + DTLS interoperation with non-compliant servers
1598 + Don't call get_session_cb() without proposed session
1599 + Fix ia64 assembler code
1600 [Andy Polyakov, Steve Henson]
1602 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
1604 *) DTLS Handshake overhaul. There were longstanding issues with
1605 OpenSSL DTLS implementation, which were making it impossible for
1606 RFC 4347 compliant client to communicate with OpenSSL server.
1607 Unfortunately just fixing these incompatibilities would "cut off"
1608 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
1609 server keeps tolerating non RFC compliant syntax. The opposite is
1610 not true, 0.9.8f client can not communicate with earlier server.
1611 This update even addresses CVE-2007-4995.
1614 *) Changes to avoid need for function casts in OpenSSL: some compilers
1615 (gcc 4.2 and later) reject their use.
1616 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
1619 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1620 RFC4507bis. The encrypted ticket format is an encrypted encoded
1621 SSL_SESSION structure, that way new session features are automatically
1624 If a client application caches session in an SSL_SESSION structure
1625 support is transparent because tickets are now stored in the encoded
1628 The SSL_CTX structure automatically generates keys for ticket
1629 protection in servers so again support should be possible
1630 with no application modification.
1632 If a client or server wishes to disable RFC4507 support then the option
1633 SSL_OP_NO_TICKET can be set.
1635 Add a TLS extension debugging callback to allow the contents of any client
1636 or server extensions to be examined.
1638 This work was sponsored by Google.
1641 *) Add initial support for TLS extensions, specifically for the server_name
1642 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1643 have new members for a host name. The SSL data structure has an
1644 additional member SSL_CTX *initial_ctx so that new sessions can be
1645 stored in that context to allow for session resumption, even after the
1646 SSL has been switched to a new SSL_CTX in reaction to a client's
1647 server_name extension.
1649 New functions (subject to change):
1651 SSL_get_servername()
1652 SSL_get_servername_type()
1655 New CTRL codes and macros (subject to change):
1657 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1658 - SSL_CTX_set_tlsext_servername_callback()
1659 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1660 - SSL_CTX_set_tlsext_servername_arg()
1661 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1663 openssl s_client has a new '-servername ...' option.
1665 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1666 '-key2 ...', '-servername_fatal' (subject to change). This allows
1667 testing the HostName extension for a specific single host name ('-cert'
1668 and '-key' remain fallbacks for handshakes without HostName
1669 negotiation). If the unrecogninzed_name alert has to be sent, this by
1670 default is a warning; it becomes fatal with the '-servername_fatal'
1673 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
1675 *) Add AES and SSE2 assembly language support to VC++ build.
1678 *) Mitigate attack on final subtraction in Montgomery reduction.
1681 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
1682 (which previously caused an internal error).
1685 *) Squeeze another 10% out of IGE mode when in != out.
1688 *) AES IGE mode speedup.
1689 [Dean Gaudet (Google)]
1691 *) Add the Korean symmetric 128-bit cipher SEED (see
1692 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
1693 add SEED ciphersuites from RFC 4162:
1695 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
1696 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
1697 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
1698 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
1700 To minimize changes between patchlevels in the OpenSSL 0.9.8
1701 series, SEED remains excluded from compilation unless OpenSSL
1702 is configured with 'enable-seed'.
1703 [KISA, Bodo Moeller]
1705 *) Mitigate branch prediction attacks, which can be practical if a
1706 single processor is shared, allowing a spy process to extract
1707 information. For detailed background information, see
1708 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
1709 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
1710 and Necessary Software Countermeasures"). The core of the change
1711 are new versions BN_div_no_branch() and
1712 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
1713 respectively, which are slower, but avoid the security-relevant
1714 conditional branches. These are automatically called by BN_div()
1715 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
1716 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
1717 remove a conditional branch.
1719 BN_FLG_CONSTTIME is the new name for the previous
1720 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
1721 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
1722 in the exponent causes BN_mod_exp_mont() to use the alternative
1723 implementation in BN_mod_exp_mont_consttime().) The old name
1724 remains as a deprecated alias.
1726 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
1727 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
1728 constant-time implementations for more than just exponentiation.
1729 Here too the old name is kept as a deprecated alias.
1731 BN_BLINDING_new() will now use BN_dup() for the modulus so that
1732 the BN_BLINDING structure gets an independent copy of the
1733 modulus. This means that the previous "BIGNUM *m" argument to
1734 BN_BLINDING_new() and to BN_BLINDING_create_param() now
1735 essentially becomes "const BIGNUM *m", although we can't actually
1736 change this in the header file before 0.9.9. It allows
1737 RSA_setup_blinding() to use BN_with_flags() on the modulus to
1738 enable BN_FLG_CONSTTIME.
1740 [Matthew D Wood (Intel Corp)]
1742 *) In the SSL/TLS server implementation, be strict about session ID
1743 context matching (which matters if an application uses a single
1744 external cache for different purposes). Previously,
1745 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
1746 set. This did ensure strict client verification, but meant that,
1747 with applications using a single external cache for quite
1748 different requirements, clients could circumvent ciphersuite
1749 restrictions for a given session ID context by starting a session
1750 in a different context.
1753 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
1754 a ciphersuite string such as "DEFAULT:RSA" cannot enable
1755 authentication-only ciphersuites.
1758 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
1759 not complete and could lead to a possible single byte overflow
1760 (CVE-2007-5135) [Ben Laurie]
1762 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
1764 *) Since AES128 and AES256 (and similarly Camellia128 and
1765 Camellia256) share a single mask bit in the logic of
1766 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
1767 kludge to work properly if AES128 is available and AES256 isn't
1768 (or if Camellia128 is available and Camellia256 isn't).
1771 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
1772 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
1773 When a point or a seed is encoded in a BIT STRING, we need to
1774 prevent the removal of trailing zero bits to get the proper DER
1775 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
1776 of a NamedBitList, for which trailing 0 bits need to be removed.)
1779 *) Have SSL/TLS server implementation tolerate "mismatched" record
1780 protocol version while receiving ClientHello even if the
1781 ClientHello is fragmented. (The server can't insist on the
1782 particular protocol version it has chosen before the ServerHello
1783 message has informed the client about his choice.)
1786 *) Add RFC 3779 support.
1787 [Rob Austein for ARIN, Ben Laurie]
1789 *) Load error codes if they are not already present instead of using a
1790 static variable. This allows them to be cleanly unloaded and reloaded.
1791 Improve header file function name parsing.
1794 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
1795 or CAPABILITY handshake as required by RFCs.
1798 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
1800 *) Introduce limits to prevent malicious keys being able to
1801 cause a denial of service. (CVE-2006-2940)
1802 [Steve Henson, Bodo Moeller]
1804 *) Fix ASN.1 parsing of certain invalid structures that can result
1805 in a denial of service. (CVE-2006-2937) [Steve Henson]
1807 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
1808 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
1810 *) Fix SSL client code which could crash if connecting to a
1811 malicious SSLv2 server. (CVE-2006-4343)
1812 [Tavis Ormandy and Will Drewry, Google Security Team]
1814 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
1815 match only those. Before that, "AES256-SHA" would be interpreted
1816 as a pattern and match "AES128-SHA" too (since AES128-SHA got
1817 the same strength classification in 0.9.7h) as we currently only
1818 have a single AES bit in the ciphersuite description bitmap.
1819 That change, however, also applied to ciphersuite strings such as
1820 "RC4-MD5" that intentionally matched multiple ciphersuites --
1821 namely, SSL 2.0 ciphersuites in addition to the more common ones
1822 from SSL 3.0/TLS 1.0.
1824 So we change the selection algorithm again: Naming an explicit
1825 ciphersuite selects this one ciphersuite, and any other similar
1826 ciphersuite (same bitmap) from *other* protocol versions.
1827 Thus, "RC4-MD5" again will properly select both the SSL 2.0
1828 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
1830 Since SSL 2.0 does not have any ciphersuites for which the
1831 128/256 bit distinction would be relevant, this works for now.
1832 The proper fix will be to use different bits for AES128 and
1833 AES256, which would have avoided the problems from the beginning;
1834 however, bits are scarce, so we can only do this in a new release
1835 (not just a patchlevel) when we can change the SSL_CIPHER
1836 definition to split the single 'unsigned long mask' bitmap into
1837 multiple values to extend the available space.
1841 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
1843 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
1844 (CVE-2006-4339) [Ben Laurie and Google Security Team]
1846 *) Add AES IGE and biIGE modes.
1849 *) Change the Unix randomness entropy gathering to use poll() when
1850 possible instead of select(), since the latter has some
1851 undesirable limitations.
1852 [Darryl Miles via Richard Levitte and Bodo Moeller]
1854 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
1855 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
1856 cannot be implicitly activated as part of, e.g., the "AES" alias.
1857 However, please upgrade to OpenSSL 0.9.9[-dev] for
1858 non-experimental use of the ECC ciphersuites to get TLS extension
1859 support, which is required for curve and point format negotiation
1860 to avoid potential handshake problems.
1863 *) Disable rogue ciphersuites:
1865 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
1866 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
1867 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
1869 The latter two were purportedly from
1870 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
1873 Also deactivate the remaining ciphersuites from
1874 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
1875 unofficial, and the ID has long expired.
1878 *) Fix RSA blinding Heisenbug (problems sometimes occured on
1879 dual-core machines) and other potential thread-safety issues.
1882 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
1883 versions), which is now available for royalty-free use
1884 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
1885 Also, add Camellia TLS ciphersuites from RFC 4132.
1887 To minimize changes between patchlevels in the OpenSSL 0.9.8
1888 series, Camellia remains excluded from compilation unless OpenSSL
1889 is configured with 'enable-camellia'.
1892 *) Disable the padding bug check when compression is in use. The padding
1893 bug check assumes the first packet is of even length, this is not
1894 necessarily true if compresssion is enabled and can result in false
1895 positives causing handshake failure. The actual bug test is ancient
1896 code so it is hoped that implementations will either have fixed it by
1897 now or any which still have the bug do not support compression.
1900 Changes between 0.9.8a and 0.9.8b [04 May 2006]
1902 *) When applying a cipher rule check to see if string match is an explicit
1903 cipher suite and only match that one cipher suite if it is.
1906 *) Link in manifests for VC++ if needed.
1907 [Austin Ziegler <halostatue@gmail.com>]
1909 *) Update support for ECC-based TLS ciphersuites according to
1910 draft-ietf-tls-ecc-12.txt with proposed changes (but without
1911 TLS extensions, which are supported starting with the 0.9.9
1912 branch, not in the OpenSSL 0.9.8 branch).
1915 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
1916 opaque EVP_CIPHER_CTX handling.
1919 *) Fixes and enhancements to zlib compression code. We now only use
1920 "zlib1.dll" and use the default __cdecl calling convention on Win32
1921 to conform with the standards mentioned here:
1922 http://www.zlib.net/DLL_FAQ.txt
1923 Static zlib linking now works on Windows and the new --with-zlib-include
1924 --with-zlib-lib options to Configure can be used to supply the location
1925 of the headers and library. Gracefully handle case where zlib library
1929 *) Several fixes and enhancements to the OID generation code. The old code
1930 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
1931 handle numbers larger than ULONG_MAX, truncated printing and had a
1932 non standard OBJ_obj2txt() behaviour.
1935 *) Add support for building of engines under engine/ as shared libraries
1936 under VC++ build system.
1939 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
1940 Hopefully, we will not see any false combination of paths any more.
1943 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
1945 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1946 (part of SSL_OP_ALL). This option used to disable the
1947 countermeasure against man-in-the-middle protocol-version
1948 rollback in the SSL 2.0 server implementation, which is a bad
1949 idea. (CVE-2005-2969)
1951 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
1952 for Information Security, National Institute of Advanced Industrial
1953 Science and Technology [AIST], Japan)]
1955 *) Add two function to clear and return the verify parameter flags.
1958 *) Keep cipherlists sorted in the source instead of sorting them at
1959 runtime, thus removing the need for a lock.
1962 *) Avoid some small subgroup attacks in Diffie-Hellman.
1963 [Nick Mathewson and Ben Laurie]
1965 *) Add functions for well-known primes.
1968 *) Extended Windows CE support.
1969 [Satoshi Nakamura and Andy Polyakov]
1971 *) Initialize SSL_METHOD structures at compile time instead of during
1972 runtime, thus removing the need for a lock.
1975 *) Make PKCS7_decrypt() work even if no certificate is supplied by
1976 attempting to decrypt each encrypted key in turn. Add support to
1980 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
1982 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
1985 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
1988 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
1989 key into the same file any more.
1992 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
1995 *) Add -utf8 command line and config file option to 'ca'.
1996 [Stefan <stf@udoma.org]
1998 *) Removed the macro des_crypt(), as it seems to conflict with some
1999 libraries. Use DES_crypt().
2002 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2003 involves renaming the source and generated shared-libs for
2004 both. The engines will accept the corrected or legacy ids
2005 ('ncipher' and '4758_cca' respectively) when binding. NB,
2006 this only applies when building 'shared'.
2007 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2009 *) Add attribute functions to EVP_PKEY structure. Modify
2010 PKCS12_create() to recognize a CSP name attribute and
2011 use it. Make -CSP option work again in pkcs12 utility.
2014 *) Add new functionality to the bn blinding code:
2015 - automatic re-creation of the BN_BLINDING parameters after
2016 a fixed number of uses (currently 32)
2017 - add new function for parameter creation
2018 - introduce flags to control the update behaviour of the
2019 BN_BLINDING parameters
2020 - hide BN_BLINDING structure
2021 Add a second BN_BLINDING slot to the RSA structure to improve
2022 performance when a single RSA object is shared among several
2026 *) Add support for DTLS.
2027 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2029 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2030 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2033 *) Remove buggy and incompletet DH cert support from
2034 ssl/ssl_rsa.c and ssl/s3_both.c
2037 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2038 the apps/openssl applications.
2041 *) Compile clean with "-Wall -Wmissing-prototypes
2042 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2043 DEBUG_SAFESTACK must also be set.
2046 *) Change ./Configure so that certain algorithms can be disabled by default.
2047 The new counterpiece to "no-xxx" is "enable-xxx".
2049 The patented RC5 and MDC2 algorithms will now be disabled unless
2050 "enable-rc5" and "enable-mdc2", respectively, are specified.
2052 (IDEA remains enabled despite being patented. This is because IDEA
2053 is frequently required for interoperability, and there is no license
2054 fee for non-commercial use. As before, "no-idea" can be used to
2055 avoid this algorithm.)
2059 *) Add processing of proxy certificates (see RFC 3820). This work was
2060 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2061 EGEE (Enabling Grids for E-science in Europe).
2064 *) RC4 performance overhaul on modern architectures/implementations, such
2065 as Intel P4, IA-64 and AMD64.
2068 *) New utility extract-section.pl. This can be used specify an alternative
2069 section number in a pod file instead of having to treat each file as
2070 a separate case in Makefile. This can be done by adding two lines to the
2073 =for comment openssl_section:XXX
2075 The blank line is mandatory.
2079 *) New arguments -certform, -keyform and -pass for s_client and s_server
2080 to allow alternative format key and certificate files and passphrase
2084 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2085 update associated structures and add various utility functions.
2087 Add new policy related verify parameters, include policy checking in
2088 standard verify code. Enhance 'smime' application with extra parameters
2089 to support policy checking and print out.
2092 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2093 Nehemiah processors. These extensions support AES encryption in hardware
2094 as well as RNG (though RNG support is currently disabled).
2095 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2097 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2100 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2101 [Andy Polyakov and a number of other people]
2103 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2104 implementation contributed by IBM.
2105 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2107 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2108 exponent rather than 'unsigned long'. There is a corresponding change to
2109 the new 'rsa_keygen' element of the RSA_METHOD structure.
2110 [Jelte Jansen, Geoff Thorpe]
2112 *) Functionality for creating the initial serial number file is now
2113 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2115 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2116 number file to 1, which is bound to cause problems. To avoid
2117 the problems while respecting compatibility between different 0.9.7
2118 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2119 CA.pl for serial number initialization. With the new release 0.9.8,
2120 we can fix the problem directly in the 'ca' utility.)
2123 *) Reduced header interdepencies by declaring more opaque objects in
2124 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2125 give fewer recursive includes, which could break lazy source code - so
2126 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2127 developers should define this symbol when building and using openssl to
2128 ensure they track the recommended behaviour, interfaces, [etc], but
2129 backwards-compatible behaviour prevails when this isn't defined.
2132 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2135 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2136 This will generate a random key of the appropriate length based on the
2137 cipher context. The EVP_CIPHER can provide its own random key generation
2138 routine to support keys of a specific form. This is used in the des and
2139 3des routines to generate a key of the correct parity. Update S/MIME
2140 code to use new functions and hence generate correct parity DES keys.
2141 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2142 valid (weak or incorrect parity).
2145 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2146 as looking them up. This is useful when the verified structure may contain
2147 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2148 present unless the new PKCS7_NO_CRL flag is asserted.
2151 *) Extend ASN1 oid configuration module. It now additionally accepts the
2154 shortName = some long name, 1.2.3.4
2157 *) Reimplemented the BN_CTX implementation. There is now no more static
2158 limitation on the number of variables it can handle nor the depth of the
2159 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2160 information can now expand as required, and rather than having a single
2161 static array of bignums, BN_CTX now uses a linked-list of such arrays
2162 allowing it to expand on demand whilst maintaining the usefulness of
2163 BN_CTX's "bundling".
2166 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2167 to allow all RSA operations to function using a single BN_CTX.
2170 *) Preliminary support for certificate policy evaluation and checking. This
2171 is initially intended to pass the tests outlined in "Conformance Testing
2172 of Relying Party Client Certificate Path Processing Logic" v1.07.
2175 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2176 remained unused and not that useful. A variety of other little bignum
2177 tweaks and fixes have also been made continuing on from the audit (see
2181 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2182 associated ASN1, EVP and SSL functions and old ASN1 macros.
2185 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2186 and this should never fail. So the return value from the use of
2187 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2188 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2191 *) BN_CTX_get() should return zero-valued bignums, providing the same
2192 initialised value as BN_new().
2193 [Geoff Thorpe, suggested by Ulf Möller]
2195 *) Support for inhibitAnyPolicy certificate extension.
2198 *) An audit of the BIGNUM code is underway, for which debugging code is
2199 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
2200 is considered valid when processing BIGNUMs, and causes execution to
2201 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
2202 further steps are taken to deliberately pollute unused data in BIGNUM
2203 structures to try and expose faulty code further on. For now, openssl will
2204 (in its default mode of operation) continue to tolerate the inconsistent
2205 forms that it has tolerated in the past, but authors and packagers should
2206 consider trying openssl and their own applications when compiled with
2207 these debugging symbols defined. It will help highlight potential bugs in
2208 their own code, and will improve the test coverage for OpenSSL itself. At
2209 some point, these tighter rules will become openssl's default to improve
2210 maintainability, though the assert()s and other overheads will remain only
2211 in debugging configurations. See bn.h for more details.
2212 [Geoff Thorpe, Nils Larsch, Ulf Möller]
2214 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
2215 that can only be obtained through BN_CTX_new() (which implicitly
2216 initialises it). The presence of this function only made it possible
2217 to overwrite an existing structure (and cause memory leaks).
2220 *) Because of the callback-based approach for implementing LHASH as a
2221 template type, lh_insert() adds opaque objects to hash-tables and
2222 lh_doall() or lh_doall_arg() are typically used with a destructor callback
2223 to clean up those corresponding objects before destroying the hash table
2224 (and losing the object pointers). So some over-zealous constifications in
2225 LHASH have been relaxed so that lh_insert() does not take (nor store) the
2226 objects as "const" and the lh_doall[_arg] callback wrappers are not
2227 prototyped to have "const" restrictions on the object pointers they are
2228 given (and so aren't required to cast them away any more).
2231 *) The tmdiff.h API was so ugly and minimal that our own timing utility
2232 (speed) prefers to use its own implementation. The two implementations
2233 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
2234 its object type properly exposed (MS_TM) instead of casting to/from "char
2235 *". This may still change yet if someone realises MS_TM and "ms_time_***"
2236 aren't necessarily the greatest nomenclatures - but this is what was used
2237 internally to the implementation so I've used that for now.
2240 *) Ensure that deprecated functions do not get compiled when
2241 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
2242 the self-tests were still using deprecated key-generation functions so
2243 these have been updated also.
2246 *) Reorganise PKCS#7 code to separate the digest location functionality
2247 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
2248 New function PKCS7_set_digest() to set the digest type for PKCS#7
2249 digestedData type. Add additional code to correctly generate the
2250 digestedData type and add support for this type in PKCS7 initialization
2254 *) New function PKCS7_set0_type_other() this initializes a PKCS7
2255 structure of type "other".
2258 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
2259 sure the loop does correctly stop and breaking ("division by zero")
2260 modulus operations are not performed. The (pre-generated) prime
2261 table crypto/bn/bn_prime.h was already correct, but it could not be
2262 re-generated on some platforms because of the "division by zero"
2263 situation in the script.
2264 [Ralf S. Engelschall]
2266 *) Update support for ECC-based TLS ciphersuites according to
2267 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
2268 SHA-1 now is only used for "small" curves (where the
2269 representation of a field element takes up to 24 bytes); for
2270 larger curves, the field element resulting from ECDH is directly
2271 used as premaster secret.
2272 [Douglas Stebila (Sun Microsystems Laboratories)]
2274 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
2275 curve secp160r1 to the tests.
2276 [Douglas Stebila (Sun Microsystems Laboratories)]
2278 *) Add the possibility to load symbols globally with DSO.
2279 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
2281 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
2282 control of the error stack.
2285 *) Add support for STORE in ENGINE.
2288 *) Add the STORE type. The intention is to provide a common interface
2289 to certificate and key stores, be they simple file-based stores, or
2290 HSM-type store, or LDAP stores, or...
2291 NOTE: The code is currently UNTESTED and isn't really used anywhere.
2294 *) Add a generic structure called OPENSSL_ITEM. This can be used to
2295 pass a list of arguments to any function as well as provide a way
2296 for a function to pass data back to the caller.
2299 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
2300 works like BUF_strdup() but can be used to duplicate a portion of
2301 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
2305 *) Add the function sk_find_ex() which works like sk_find(), but will
2306 return an index to an element even if an exact match couldn't be
2307 found. The index is guaranteed to point at the element where the
2308 searched-for key would be inserted to preserve sorting order.
2311 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
2312 takes an extra flags argument for optional functionality. Currently,
2313 the following flags are defined:
2315 OBJ_BSEARCH_VALUE_ON_NOMATCH
2316 This one gets OBJ_bsearch_ex() to return a pointer to the first
2317 element where the comparing function returns a negative or zero
2320 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
2321 This one gets OBJ_bsearch_ex() to return a pointer to the first
2322 element where the comparing function returns zero. This is useful
2323 if there are more than one element where the comparing function
2327 *) Make it possible to create self-signed certificates with 'openssl ca'
2328 in such a way that the self-signed certificate becomes part of the
2329 CA database and uses the same mechanisms for serial number generation
2330 as all other certificate signing. The new flag '-selfsign' enables
2331 this functionality. Adapt CA.sh and CA.pl.in.
2334 *) Add functionality to check the public key of a certificate request
2335 against a given private. This is useful to check that a certificate
2336 request can be signed by that key (self-signing).
2339 *) Make it possible to have multiple active certificates with the same
2340 subject in the CA index file. This is done only if the keyword
2341 'unique_subject' is set to 'no' in the main CA section (default
2342 if 'CA_default') of the configuration file. The value is saved
2343 with the database itself in a separate index attribute file,
2344 named like the index file with '.attr' appended to the name.
2347 *) Generate muti valued AVAs using '+' notation in config files for
2351 *) Support for nameConstraints certificate extension.
2354 *) Support for policyConstraints certificate extension.
2357 *) Support for policyMappings certificate extension.
2360 *) Make sure the default DSA_METHOD implementation only uses its
2361 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
2362 and change its own handlers to be NULL so as to remove unnecessary
2363 indirection. This lets alternative implementations fallback to the
2364 default implementation more easily.
2367 *) Support for directoryName in GeneralName related extensions
2371 *) Make it possible to link applications using Makefile.shared.
2372 Make that possible even when linking against static libraries!
2375 *) Support for single pass processing for S/MIME signing. This now
2376 means that S/MIME signing can be done from a pipe, in addition
2377 cleartext signing (multipart/signed type) is effectively streaming
2378 and the signed data does not need to be all held in memory.
2380 This is done with a new flag PKCS7_STREAM. When this flag is set
2381 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
2382 is done after the data is output (and digests calculated) in
2383 SMIME_write_PKCS7().
2386 *) Add full support for -rpath/-R, both in shared libraries and
2387 applications, at least on the platforms where it's known how
2391 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
2392 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
2393 will now compute a table of multiples of the generator that
2394 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
2395 faster (notably in the case of a single point multiplication,
2396 scalar * generator).
2397 [Nils Larsch, Bodo Moeller]
2399 *) IPv6 support for certificate extensions. The various extensions
2400 which use the IP:a.b.c.d can now take IPv6 addresses using the
2401 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
2405 *) Added an ENGINE that implements RSA by performing private key
2406 exponentiations with the GMP library. The conversions to and from
2407 GMP's mpz_t format aren't optimised nor are any montgomery forms
2408 cached, and on x86 it appears OpenSSL's own performance has caught up.
2409 However there are likely to be other architectures where GMP could
2410 provide a boost. This ENGINE is not built in by default, but it can be
2411 specified at Configure time and should be accompanied by the necessary
2412 linker additions, eg;
2413 ./config -DOPENSSL_USE_GMP -lgmp
2416 *) "openssl engine" will not display ENGINE/DSO load failure errors when
2417 testing availability of engines with "-t" - the old behaviour is
2418 produced by increasing the feature's verbosity with "-tt".
2421 *) ECDSA routines: under certain error conditions uninitialized BN objects
2422 could be freed. Solution: make sure initialization is performed early
2423 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
2427 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
2428 and DH_METHOD (eg. by ENGINE implementations) to override the normal
2429 software implementations. For DSA and DH, parameter generation can
2430 also be overriden by providing the appropriate method callbacks.
2433 *) Change the "progress" mechanism used in key-generation and
2434 primality testing to functions that take a new BN_GENCB pointer in
2435 place of callback/argument pairs. The new API functions have "_ex"
2436 postfixes and the older functions are reimplemented as wrappers for
2437 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
2438 declarations of the old functions to help (graceful) attempts to
2439 migrate to the new functions. Also, the new key-generation API
2440 functions operate on a caller-supplied key-structure and return
2441 success/failure rather than returning a key or NULL - this is to
2442 help make "keygen" another member function of RSA_METHOD etc.
2444 Example for using the new callback interface:
2446 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
2450 BN_GENCB_set(&my_cb, my_callback, my_arg);
2452 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
2453 /* For the meaning of a, b in calls to my_callback(), see the
2454 * documentation of the function that calls the callback.
2455 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
2456 * my_callback should return 1 if it wants BN_is_prime_ex()
2457 * to continue, or 0 to stop.
2462 *) Change the ZLIB compression method to be stateful, and make it
2463 available to TLS with the number defined in
2464 draft-ietf-tls-compression-04.txt.
2467 *) Add the ASN.1 structures and functions for CertificatePair, which
2468 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
2470 CertificatePair ::= SEQUENCE {
2471 forward [0] Certificate OPTIONAL,
2472 reverse [1] Certificate OPTIONAL,
2473 -- at least one of the pair shall be present -- }
2475 Also implement the PEM functions to read and write certificate
2476 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
2478 This needed to be defined, mostly for the sake of the LDAP
2479 attribute crossCertificatePair, but may prove useful elsewhere as
2483 *) Make it possible to inhibit symlinking of shared libraries in
2484 Makefile.shared, for Cygwin's sake.
2487 *) Extend the BIGNUM API by creating a function
2488 void BN_set_negative(BIGNUM *a, int neg);
2489 and a macro that behave like
2490 int BN_is_negative(const BIGNUM *a);
2492 to avoid the need to access 'a->neg' directly in applications.
2495 *) Implement fast modular reduction for pseudo-Mersenne primes
2496 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
2497 EC_GROUP_new_curve_GFp() will now automatically use this
2499 [Nils Larsch <nla@trustcenter.de>]
2501 *) Add new lock type (CRYPTO_LOCK_BN).
2504 *) Change the ENGINE framework to automatically load engines
2505 dynamically from specific directories unless they could be
2506 found to already be built in or loaded. Move all the
2507 current engines except for the cryptodev one to a new
2509 The engines in engines/ are built as shared libraries if
2510 the "shared" options was given to ./Configure or ./config.
2511 Otherwise, they are inserted in libcrypto.a.
2512 /usr/local/ssl/engines is the default directory for dynamic
2513 engines, but that can be overriden at configure time through
2514 the usual use of --prefix and/or --openssldir, and at run
2515 time with the environment variable OPENSSL_ENGINES.
2516 [Geoff Thorpe and Richard Levitte]
2518 *) Add Makefile.shared, a helper makefile to build shared
2519 libraries. Addapt Makefile.org.
2522 *) Add version info to Win32 DLLs.
2523 [Peter 'Luna' Runestig" <peter@runestig.com>]
2525 *) Add new 'medium level' PKCS#12 API. Certificates and keys
2526 can be added using this API to created arbitrary PKCS#12
2527 files while avoiding the low level API.
2529 New options to PKCS12_create(), key or cert can be NULL and
2530 will then be omitted from the output file. The encryption
2531 algorithm NIDs can be set to -1 for no encryption, the mac
2532 iteration count can be set to 0 to omit the mac.
2534 Enhance pkcs12 utility by making the -nokeys and -nocerts
2535 options work when creating a PKCS#12 file. New option -nomac
2536 to omit the mac, NONE can be set for an encryption algorithm.
2537 New code is modified to use the enhanced PKCS12_create()
2538 instead of the low level API.
2541 *) Extend ASN1 encoder to support indefinite length constructed
2542 encoding. This can output sequences tags and octet strings in
2543 this form. Modify pk7_asn1.c to support indefinite length
2544 encoding. This is experimental and needs additional code to
2545 be useful, such as an ASN1 bio and some enhanced streaming
2548 Extend template encode functionality so that tagging is passed
2549 down to the template encoder.
2552 *) Let 'openssl req' fail if an argument to '-newkey' is not
2553 recognized instead of using RSA as a default.
2556 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
2557 As these are not official, they are not included in "ALL";
2558 the "ECCdraft" ciphersuite group alias can be used to select them.
2559 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
2561 *) Add ECDH engine support.
2562 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
2564 *) Add ECDH in new directory crypto/ecdh/.
2565 [Douglas Stebila (Sun Microsystems Laboratories)]
2567 *) Let BN_rand_range() abort with an error after 100 iterations
2568 without success (which indicates a broken PRNG).
2571 *) Change BN_mod_sqrt() so that it verifies that the input value
2572 is really the square of the return value. (Previously,
2573 BN_mod_sqrt would show GIGO behaviour.)
2576 *) Add named elliptic curves over binary fields from X9.62, SECG,
2577 and WAP/WTLS; add OIDs that were still missing.
2579 [Sheueling Chang Shantz and Douglas Stebila
2580 (Sun Microsystems Laboratories)]
2582 *) Extend the EC library for elliptic curves over binary fields
2583 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
2586 EC_GF2m_simple_method
2590 EC_GROUP_new_curve_GF2m
2591 EC_GROUP_set_curve_GF2m
2592 EC_GROUP_get_curve_GF2m
2593 EC_POINT_set_affine_coordinates_GF2m
2594 EC_POINT_get_affine_coordinates_GF2m
2595 EC_POINT_set_compressed_coordinates_GF2m
2597 Point compression for binary fields is disabled by default for
2598 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
2601 As binary polynomials are represented as BIGNUMs, various members
2602 of the EC_GROUP and EC_POINT data structures can be shared
2603 between the implementations for prime fields and binary fields;
2604 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
2605 are essentially identical to their ..._GFp counterparts.
2606 (For simplicity, the '..._GFp' prefix has been dropped from
2607 various internal method names.)
2609 An internal 'field_div' method (similar to 'field_mul' and
2610 'field_sqr') has been added; this is used only for binary fields.
2612 [Sheueling Chang Shantz and Douglas Stebila
2613 (Sun Microsystems Laboratories)]
2615 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
2616 through methods ('mul', 'precompute_mult').
2618 The generic implementations (now internally called 'ec_wNAF_mul'
2619 and 'ec_wNAF_precomputed_mult') remain the default if these
2620 methods are undefined.
2622 [Sheueling Chang Shantz and Douglas Stebila
2623 (Sun Microsystems Laboratories)]
2625 *) New function EC_GROUP_get_degree, which is defined through
2626 EC_METHOD. For curves over prime fields, this returns the bit
2627 length of the modulus.
2629 [Sheueling Chang Shantz and Douglas Stebila
2630 (Sun Microsystems Laboratories)]
2632 *) New functions EC_GROUP_dup, EC_POINT_dup.
2633 (These simply call ..._new and ..._copy).
2635 [Sheueling Chang Shantz and Douglas Stebila
2636 (Sun Microsystems Laboratories)]
2638 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
2639 Polynomials are represented as BIGNUMs (where the sign bit is not
2640 used) in the following functions [macros]:
2643 BN_GF2m_sub [= BN_GF2m_add]
2644 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
2645 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
2646 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
2648 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
2649 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
2650 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
2651 BN_GF2m_cmp [= BN_ucmp]
2653 (Note that only the 'mod' functions are actually for fields GF(2^m).
2654 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
2656 For some functions, an the irreducible polynomial defining a
2657 field can be given as an 'unsigned int[]' with strictly
2658 decreasing elements giving the indices of those bits that are set;
2659 i.e., p[] represents the polynomial
2660 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
2662 p[0] > p[1] > ... > p[k] = 0.
2663 This applies to the following functions:
2668 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
2669 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
2671 BN_GF2m_mod_sqrt_arr
2672 BN_GF2m_mod_solve_quad_arr
2676 Conversion can be performed by the following functions:
2681 bntest.c has additional tests for binary polynomial arithmetic.
2683 Two implementations for BN_GF2m_mod_div() are available.
2684 The default algorithm simply uses BN_GF2m_mod_inv() and
2685 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
2686 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
2687 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
2689 [Sheueling Chang Shantz and Douglas Stebila
2690 (Sun Microsystems Laboratories)]
2692 *) Add new error code 'ERR_R_DISABLED' that can be used when some
2693 functionality is disabled at compile-time.
2694 [Douglas Stebila <douglas.stebila@sun.com>]
2696 *) Change default behaviour of 'openssl asn1parse' so that more
2697 information is visible when viewing, e.g., a certificate:
2699 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
2700 mode the content of non-printable OCTET STRINGs is output in a
2701 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
2702 avoid the appearance of a printable string.
2703 [Nils Larsch <nla@trustcenter.de>]
2705 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
2707 EC_GROUP_set_asn1_flag()
2708 EC_GROUP_get_asn1_flag()
2709 EC_GROUP_set_point_conversion_form()
2710 EC_GROUP_get_point_conversion_form()
2711 These control ASN1 encoding details:
2712 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
2713 has been set to OPENSSL_EC_NAMED_CURVE.
2714 - Points are encoded in uncompressed form by default; options for
2715 asn1_for are as for point2oct, namely
2716 POINT_CONVERSION_COMPRESSED
2717 POINT_CONVERSION_UNCOMPRESSED
2718 POINT_CONVERSION_HYBRID
2720 Also add 'seed' and 'seed_len' members to EC_GROUP with access
2723 EC_GROUP_get0_seed()
2724 EC_GROUP_get_seed_len()
2725 This is used only for ASN1 purposes (so far).
2726 [Nils Larsch <nla@trustcenter.de>]
2728 *) Add 'field_type' member to EC_METHOD, which holds the NID
2729 of the appropriate field type OID. The new function
2730 EC_METHOD_get_field_type() returns this value.
2731 [Nils Larsch <nla@trustcenter.de>]
2736 EC_POINT_point2hex()
2737 EC_POINT_hex2point()
2738 providing useful interfaces to EC_POINT_point2oct() and
2739 EC_POINT_oct2point().
2740 [Nils Larsch <nla@trustcenter.de>]
2742 *) Change internals of the EC library so that the functions
2743 EC_GROUP_set_generator()
2744 EC_GROUP_get_generator()
2745 EC_GROUP_get_order()
2746 EC_GROUP_get_cofactor()
2747 are implemented directly in crypto/ec/ec_lib.c and not dispatched
2748 to methods, which would lead to unnecessary code duplication when
2749 adding different types of curves.
2750 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
2752 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
2753 arithmetic, and such that modified wNAFs are generated
2754 (which avoid length expansion in many cases).
2757 *) Add a function EC_GROUP_check_discriminant() (defined via
2758 EC_METHOD) that verifies that the curve discriminant is non-zero.
2760 Add a function EC_GROUP_check() that makes some sanity tests
2761 on a EC_GROUP, its generator and order. This includes
2762 EC_GROUP_check_discriminant().
2763 [Nils Larsch <nla@trustcenter.de>]
2765 *) Add ECDSA in new directory crypto/ecdsa/.
2767 Add applications 'openssl ecparam' and 'openssl ecdsa'
2768 (these are based on 'openssl dsaparam' and 'openssl dsa').
2770 ECDSA support is also included in various other files across the
2771 library. Most notably,
2772 - 'openssl req' now has a '-newkey ecdsa:file' option;
2773 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
2774 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
2775 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
2776 them suitable for ECDSA where domain parameters must be
2777 extracted before the specific public key;
2778 - ECDSA engine support has been added.
2779 [Nils Larsch <nla@trustcenter.de>]
2781 *) Include some named elliptic curves, and add OIDs from X9.62,
2782 SECG, and WAP/WTLS. Each curve can be obtained from the new
2784 EC_GROUP_new_by_curve_name(),
2785 and the list of available named curves can be obtained with
2786 EC_get_builtin_curves().
2787 Also add a 'curve_name' member to EC_GROUP objects, which can be
2789 EC_GROUP_set_curve_name()
2790 EC_GROUP_get_curve_name()
2791 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
2793 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2794 was actually never needed) and in BN_mul(). The removal in BN_mul()
2795 required a small change in bn_mul_part_recursive() and the addition
2796 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2797 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2798 bn_sub_words() and bn_add_words() except they take arrays with
2802 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
2804 *) Cleanse PEM buffers before freeing them since they may contain
2806 [Benjamin Bennett <ben@psc.edu>]
2808 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2809 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2810 authentication-only ciphersuites.
2813 *) Since AES128 and AES256 share a single mask bit in the logic of
2814 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2815 kludge to work properly if AES128 is available and AES256 isn't.
2818 *) Expand security boundary to match 1.1.1 module.
2821 *) Remove redundant features: hash file source, editing of test vectors
2822 modify fipsld to use external fips_premain.c signature.
2825 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
2826 run algorithm test programs.
2829 *) Make algorithm test programs more tolerant of whitespace.
2832 *) Have SSL/TLS server implementation tolerate "mismatched" record
2833 protocol version while receiving ClientHello even if the
2834 ClientHello is fragmented. (The server can't insist on the
2835 particular protocol version it has chosen before the ServerHello
2836 message has informed the client about his choice.)
2839 *) Load error codes if they are not already present instead of using a
2840 static variable. This allows them to be cleanly unloaded and reloaded.
2843 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
2845 *) Introduce limits to prevent malicious keys being able to
2846 cause a denial of service. (CVE-2006-2940)
2847 [Steve Henson, Bodo Moeller]
2849 *) Fix ASN.1 parsing of certain invalid structures that can result
2850 in a denial of service. (CVE-2006-2937) [Steve Henson]
2852 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2853 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2855 *) Fix SSL client code which could crash if connecting to a
2856 malicious SSLv2 server. (CVE-2006-4343)
2857 [Tavis Ormandy and Will Drewry, Google Security Team]
2859 *) Change ciphersuite string processing so that an explicit
2860 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
2861 will no longer include "AES128-SHA"), and any other similar
2862 ciphersuite (same bitmap) from *other* protocol versions (so that
2863 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
2864 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
2865 changes from 0.9.8b and 0.9.8d.
2868 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
2870 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2871 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2873 *) Change the Unix randomness entropy gathering to use poll() when
2874 possible instead of select(), since the latter has some
2875 undesirable limitations.
2876 [Darryl Miles via Richard Levitte and Bodo Moeller]
2878 *) Disable rogue ciphersuites:
2880 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2881 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2882 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2884 The latter two were purportedly from
2885 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2888 Also deactive the remaining ciphersuites from
2889 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2890 unofficial, and the ID has long expired.
2893 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2894 dual-core machines) and other potential thread-safety issues.
2897 Changes between 0.9.7i and 0.9.7j [04 May 2006]
2899 *) Adapt fipsld and the build system to link against the validated FIPS
2900 module in FIPS mode.
2903 *) Fixes for VC++ 2005 build under Windows.
2906 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
2907 from a Windows bash shell such as MSYS. It is autodetected from the
2908 "config" script when run from a VC++ environment. Modify standard VC++
2909 build to use fipscanister.o from the GNU make build.
2912 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
2914 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
2915 The value now differs depending on if you build for FIPS or not.
2916 BEWARE! A program linked with a shared FIPSed libcrypto can't be
2917 safely run with a non-FIPSed libcrypto, as it may crash because of
2918 the difference induced by this change.
2921 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
2923 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2924 (part of SSL_OP_ALL). This option used to disable the
2925 countermeasure against man-in-the-middle protocol-version
2926 rollback in the SSL 2.0 server implementation, which is a bad
2927 idea. (CVE-2005-2969)
2929 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2930 for Information Security, National Institute of Advanced Industrial
2931 Science and Technology [AIST], Japan)]
2933 *) Minimal support for X9.31 signatures and PSS padding modes. This is
2934 mainly for FIPS compliance and not fully integrated at this stage.
2937 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
2938 the exponentiation using a fixed-length exponent. (Otherwise,
2939 the information leaked through timing could expose the secret key
2940 after many signatures; cf. Bleichenbacher's attack on DSA with
2944 *) Make a new fixed-window mod_exp implementation the default for
2945 RSA, DSA, and DH private-key operations so that the sequence of
2946 squares and multiplies and the memory access pattern are
2947 independent of the particular secret key. This will mitigate
2948 cache-timing and potential related attacks.
2950 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
2951 and this is automatically used by BN_mod_exp_mont() if the new flag
2952 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
2953 will use this BN flag for private exponents unless the flag
2954 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
2955 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
2957 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
2959 *) Change the client implementation for SSLv23_method() and
2960 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
2961 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
2962 (Previously, the SSL 2.0 backwards compatible Client Hello
2963 message format would be used even with SSL_OP_NO_SSLv2.)
2966 *) Add support for smime-type MIME parameter in S/MIME messages which some
2970 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
2971 a threadsafe manner. Modify rsa code to use new function and add calls
2972 to dsa and dh code (which had race conditions before).
2975 *) Include the fixed error library code in the C error file definitions
2976 instead of fixing them up at runtime. This keeps the error code
2977 structures constant.
2980 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
2982 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
2985 *) Fixes for newer kerberos headers. NB: the casts are needed because
2986 the 'length' field is signed on one version and unsigned on another
2987 with no (?) obvious way to tell the difference, without these VC++
2988 complains. Also the "definition" of FAR (blank) is no longer included
2989 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
2990 some needed definitions.
2993 *) Undo Cygwin change.
2996 *) Added support for proxy certificates according to RFC 3820.
2997 Because they may be a security thread to unaware applications,
2998 they must be explicitely allowed in run-time. See
2999 docs/HOWTO/proxy_certificates.txt for further information.
3002 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3004 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3005 server and client random values. Previously
3006 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3007 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3009 This change has negligible security impact because:
3011 1. Server and client random values still have 24 bytes of pseudo random
3014 2. Server and client random values are sent in the clear in the initial
3017 3. The master secret is derived using the premaster secret (48 bytes in
3018 size for static RSA ciphersuites) as well as client server and random
3021 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3024 [Stephen Henson, reported by UK NISCC]
3026 *) Use Windows randomness collection on Cygwin.
3029 *) Fix hang in EGD/PRNGD query when communication socket is closed
3030 prematurely by EGD/PRNGD.
3031 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3033 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3036 *) Back-port of selected performance improvements from development
3037 branch, as well as improved support for PowerPC platforms.
3040 *) Add lots of checks for memory allocation failure, error codes to indicate
3041 failure and freeing up memory if a failure occurs.
3042 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3044 *) Add new -passin argument to dgst.
3047 *) Perform some character comparisons of different types in X509_NAME_cmp:
3048 this is needed for some certificates that reencode DNs into UTF8Strings
3049 (in violation of RFC3280) and can't or wont issue name rollover
3053 *) Make an explicit check during certificate validation to see that
3054 the CA setting in each certificate on the chain is correct. As a
3055 side effect always do the following basic checks on extensions,
3056 not just when there's an associated purpose to the check:
3058 - if there is an unhandled critical extension (unless the user
3059 has chosen to ignore this fault)
3060 - if the path length has been exceeded (if one is set at all)
3061 - that certain extensions fit the associated purpose (if one has
3065 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3067 *) Avoid a race condition when CRLs are checked in a multi threaded
3068 environment. This would happen due to the reordering of the revoked
3069 entries during signature checking and serial number lookup. Now the
3070 encoding is cached and the serial number sort performed under a lock.
3071 Add new STACK function sk_is_sorted().
3074 *) Add Delta CRL to the extension code.
3077 *) Various fixes to s3_pkt.c so alerts are sent properly.
3078 [David Holmes <d.holmes@f5.com>]
3080 *) Reduce the chances of duplicate issuer name and serial numbers (in
3081 violation of RFC3280) using the OpenSSL certificate creation utilities.
3082 This is done by creating a random 64 bit value for the initial serial
3083 number when a serial number file is created or when a self signed
3084 certificate is created using 'openssl req -x509'. The initial serial
3085 number file is created using 'openssl x509 -next_serial' in CA.pl
3086 rather than being initialized to 1.
3089 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3091 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3092 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3093 [Joe Orton, Steve Henson]
3095 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3097 [Joe Orton, Steve Henson]
3099 *) Make it possible to have multiple active certificates with the same
3100 subject in the CA index file. This is done only if the keyword
3101 'unique_subject' is set to 'no' in the main CA section (default
3102 if 'CA_default') of the configuration file. The value is saved
3103 with the database itself in a separate index attribute file,
3104 named like the index file with '.attr' appended to the name.
3107 *) X509 verify fixes. Disable broken certificate workarounds when
3108 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3109 keyUsage extension present. Don't accept CRLs with unhandled critical
3110 extensions: since verify currently doesn't process CRL extensions this
3111 rejects a CRL with *any* critical extensions. Add new verify error codes
3115 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
3116 A clarification of RFC2560 will require the use of OCTET STRINGs and
3117 some implementations cannot handle the current raw format. Since OpenSSL
3118 copies and compares OCSP nonces as opaque blobs without any attempt at
3119 parsing them this should not create any compatibility issues.
3122 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
3123 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
3124 this HMAC (and other) operations are several times slower than OpenSSL
3128 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
3129 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3131 *) Use the correct content when signing type "other".
3134 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
3136 *) Fix various bugs revealed by running the NISCC test suite:
3138 Stop out of bounds reads in the ASN1 code when presented with
3139 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3141 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
3143 If verify callback ignores invalid public key errors don't try to check
3144 certificate signature with the NULL public key.
3148 *) New -ignore_err option in ocsp application to stop the server
3149 exiting on the first error in a request.
3152 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3153 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3157 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3158 extra data after the compression methods not only for TLS 1.0
3159 but also for SSL 3.0 (as required by the specification).
3160 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3162 *) Change X509_certificate_type() to mark the key as exported/exportable
3163 when it's 512 *bits* long, not 512 bytes.
3166 *) Change AES_cbc_encrypt() so it outputs exact multiple of
3167 blocks during encryption.
3170 *) Various fixes to base64 BIO and non blocking I/O. On write
3171 flushes were not handled properly if the BIO retried. On read
3172 data was not being buffered properly and had various logic bugs.
3173 This also affects blocking I/O when the data being decoded is a
3177 *) Various S/MIME bugfixes and compatibility changes:
3178 output correct application/pkcs7 MIME type if
3179 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
3180 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
3181 of files as .eml work). Correctly handle very long lines in MIME
3185 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
3187 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3188 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3189 a protocol version number mismatch like a decryption error
3190 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3193 *) Turn on RSA blinding by default in the default implementation
3194 to avoid a timing attack. Applications that don't want it can call
3195 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3196 They would be ill-advised to do so in most cases.
3197 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3199 *) Change RSA blinding code so that it works when the PRNG is not
3200 seeded (in this case, the secret RSA exponent is abused as
3201 an unpredictable seed -- if it is not unpredictable, there
3202 is no point in blinding anyway). Make RSA blinding thread-safe
3203 by remembering the creator's thread ID in rsa->blinding and
3204 having all other threads use local one-time blinding factors
3205 (this requires more computation than sharing rsa->blinding, but
3206 avoids excessive locking; and if an RSA object is not shared
3207 between threads, blinding will still be very fast).
3210 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
3211 ENGINE as defaults for all supported algorithms irrespective of
3212 the 'flags' parameter. 'flags' is now honoured, so applications
3213 should make sure they are passing it correctly.
3216 *) Target "mingw" now allows native Windows code to be generated in
3217 the Cygwin environment as well as with the MinGW compiler.
3220 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
3222 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3223 via timing by performing a MAC computation even if incorrrect
3224 block cipher padding has been found. This is a countermeasure
3225 against active attacks where the attacker has to distinguish
3226 between bad padding and a MAC verification error. (CVE-2003-0078)
3228 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3229 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3230 Martin Vuagnoux (EPFL, Ilion)]
3232 *) Make the no-err option work as intended. The intention with no-err
3233 is not to have the whole error stack handling routines removed from
3234 libcrypto, it's only intended to remove all the function name and
3235 reason texts, thereby removing some of the footprint that may not
3236 be interesting if those errors aren't displayed anyway.
3238 NOTE: it's still possible for any application or module to have it's
3239 own set of error texts inserted. The routines are there, just not
3240 used by default when no-err is given.
3243 *) Add support for FreeBSD on IA64.
3244 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
3246 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
3247 Kerberos function mit_des_cbc_cksum(). Before this change,
3248 the value returned by DES_cbc_cksum() was like the one from
3249 mit_des_cbc_cksum(), except the bytes were swapped.
3250 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
3252 *) Allow an application to disable the automatic SSL chain building.
3253 Before this a rather primitive chain build was always performed in
3254 ssl3_output_cert_chain(): an application had no way to send the
3255 correct chain if the automatic operation produced an incorrect result.
3257 Now the chain builder is disabled if either:
3259 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
3261 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
3263 The reasoning behind this is that an application would not want the
3264 auto chain building to take place if extra chain certificates are
3265 present and it might also want a means of sending no additional
3266 certificates (for example the chain has two certificates and the
3270 *) Add the possibility to build without the ENGINE framework.
3271 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3273 *) Under Win32 gmtime() can return NULL: check return value in
3274 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
3277 *) DSA routines: under certain error conditions uninitialized BN objects
3278 could be freed. Solution: make sure initialization is performed early
3279 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
3280 Nils Larsch <nla@trustcenter.de> via PR#459)
3283 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
3284 checked on reconnect on the client side, therefore session resumption
3285 could still fail with a "ssl session id is different" error. This
3286 behaviour is masked when SSL_OP_ALL is used due to
3287 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
3288 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3289 followup to PR #377.
3292 *) IA-32 assembler support enhancements: unified ELF targets, support
3293 for SCO/Caldera platforms, fix for Cygwin shared build.
3296 *) Add support for FreeBSD on sparc64. As a consequence, support for
3297 FreeBSD on non-x86 processors is separate from x86 processors on
3298 the config script, much like the NetBSD support.
3299 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
3301 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
3303 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
3306 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
3307 code (06) was taken as the first octet of the session ID and the last
3308 octet was ignored consequently. As a result SSLv2 client side session
3309 caching could not have worked due to the session ID mismatch between
3311 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3315 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
3316 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
3320 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
3321 seems that in spite of existing for more than a year, many application
3322 author have done nothing to provide the necessary callbacks, which
3323 means that this particular engine will not work properly anywhere.
3324 This is a very unfortunate situation which forces us, in the name
3325 of usability, to give the hw_ncipher.c a static lock, which is part
3327 NOTE: This is for the 0.9.7 series ONLY. This hack will never
3328 appear in 0.9.8 or later. We EXPECT application authors to have
3329 dealt properly with this when 0.9.8 is released (unless we actually
3330 make such changes in the libcrypto locking code that changes will
3331 have to be made anyway).
3334 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
3335 octets have been read, EOF or an error occurs. Without this change
3336 some truncated ASN1 structures will not produce an error.
3339 *) Disable Heimdal support, since it hasn't been fully implemented.
3340 Still give the possibility to force the use of Heimdal, but with
3341 warnings and a request that patches get sent to openssl-dev.
3344 *) Add the VC-CE target, introduce the WINCE sysname, and add
3345 INSTALL.WCE and appropriate conditionals to make it build.
3346 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3348 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
3349 cygssl-x.y.z.dll, where x, y and z are the major, minor and
3350 edit numbers of the version.
3351 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3353 *) Introduce safe string copy and catenation functions
3354 (BUF_strlcpy() and BUF_strlcat()).
3355 [Ben Laurie (CHATS) and Richard Levitte]
3357 *) Avoid using fixed-size buffers for one-line DNs.
3358 [Ben Laurie (CHATS)]
3360 *) Add BUF_MEM_grow_clean() to avoid information leakage when
3361 resizing buffers containing secrets, and use where appropriate.
3362 [Ben Laurie (CHATS)]
3364 *) Avoid using fixed size buffers for configuration file location.
3365 [Ben Laurie (CHATS)]
3367 *) Avoid filename truncation for various CA files.
3368 [Ben Laurie (CHATS)]
3370 *) Use sizeof in preference to magic numbers.
3371 [Ben Laurie (CHATS)]
3373 *) Avoid filename truncation in cert requests.
3374 [Ben Laurie (CHATS)]
3376 *) Add assertions to check for (supposedly impossible) buffer
3378 [Ben Laurie (CHATS)]
3380 *) Don't cache truncated DNS entries in the local cache (this could
3381 potentially lead to a spoofing attack).
3382 [Ben Laurie (CHATS)]
3384 *) Fix various buffers to be large enough for hex/decimal
3385 representations in a platform independent manner.
3386 [Ben Laurie (CHATS)]
3388 *) Add CRYPTO_realloc_clean() to avoid information leakage when
3389 resizing buffers containing secrets, and use where appropriate.
3390 [Ben Laurie (CHATS)]
3392 *) Add BIO_indent() to avoid much slightly worrying code to do
3394 [Ben Laurie (CHATS)]
3396 *) Convert sprintf()/BIO_puts() to BIO_printf().
3397 [Ben Laurie (CHATS)]
3399 *) buffer_gets() could terminate with the buffer only half
3401 [Ben Laurie (CHATS)]
3403 *) Add assertions to prevent user-supplied crypto functions from
3404 overflowing internal buffers by having large block sizes, etc.
3405 [Ben Laurie (CHATS)]
3407 *) New OPENSSL_assert() macro (similar to assert(), but enabled
3409 [Ben Laurie (CHATS)]
3411 *) Eliminate unused copy of key in RC4.
3412 [Ben Laurie (CHATS)]
3414 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
3415 [Ben Laurie (CHATS)]
3417 *) Fix off-by-one error in EGD path.
3418 [Ben Laurie (CHATS)]
3420 *) If RANDFILE path is too long, ignore instead of truncating.
3421 [Ben Laurie (CHATS)]
3423 *) Eliminate unused and incorrectly sized X.509 structure
3425 [Ben Laurie (CHATS)]
3427 *) Eliminate unused and dangerous function knumber().
3428 [Ben Laurie (CHATS)]
3430 *) Eliminate unused and dangerous structure, KSSL_ERR.
3431 [Ben Laurie (CHATS)]
3433 *) Protect against overlong session ID context length in an encoded
3434 session object. Since these are local, this does not appear to be
3436 [Ben Laurie (CHATS)]
3438 *) Change from security patch (see 0.9.6e below) that did not affect
3439 the 0.9.6 release series:
3441 Remote buffer overflow in SSL3 protocol - an attacker could
3442 supply an oversized master key in Kerberos-enabled versions.
3444 [Ben Laurie (CHATS)]
3446 *) Change the SSL kerb5 codes to match RFC 2712.
3449 *) Make -nameopt work fully for req and add -reqopt switch.
3450 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
3452 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
3453 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
3455 *) Make sure tests can be performed even if the corresponding algorithms
3456 have been removed entirely. This was also the last step to make
3457 OpenSSL compilable with DJGPP under all reasonable conditions.
3458 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
3460 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
3461 to allow version independent disabling of normally unselected ciphers,
3462 which may be activated as a side-effect of selecting a single cipher.
3464 (E.g., cipher list string "RSA" enables ciphersuites that are left
3465 out of "ALL" because they do not provide symmetric encryption.
3466 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
3467 [Lutz Jaenicke, Bodo Moeller]
3469 *) Add appropriate support for separate platform-dependent build
3470 directories. The recommended way to make a platform-dependent
3471 build directory is the following (tested on Linux), maybe with
3474 # Place yourself outside of the OpenSSL source tree. In
3475 # this example, the environment variable OPENSSL_SOURCE
3476 # is assumed to contain the absolute OpenSSL source directory.
3477 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
3478 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
3479 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
3480 mkdir -p `dirname $F`
3481 ln -s $OPENSSL_SOURCE/$F $F
3484 To be absolutely sure not to disturb the source tree, a "make clean"
3485 is a good thing. If it isn't successfull, don't worry about it,
3486 it probably means the source directory is very clean.
3489 *) Make sure any ENGINE control commands make local copies of string
3490 pointers passed to them whenever necessary. Otherwise it is possible
3491 the caller may have overwritten (or deallocated) the original string
3492 data when a later ENGINE operation tries to use the stored values.
3493 [Götz Babin-Ebell <babinebell@trustcenter.de>]
3495 *) Improve diagnostics in file reading and command-line digests.
3496 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
3498 *) Add AES modes CFB and OFB to the object database. Correct an
3499 error in AES-CFB decryption.
3502 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
3503 allows existing EVP_CIPHER_CTX structures to be reused after
3504 calling EVP_*Final(). This behaviour is used by encryption
3505 BIOs and some applications. This has the side effect that
3506 applications must explicitly clean up cipher contexts with
3507 EVP_CIPHER_CTX_cleanup() or they will leak memory.
3510 *) Check the values of dna and dnb in bn_mul_recursive before calling
3511 bn_mul_comba (a non zero value means the a or b arrays do not contain
3512 n2 elements) and fallback to bn_mul_normal if either is not zero.
3515 *) Fix escaping of non-ASCII characters when using the -subj option
3516 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
3519 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
3520 form for "surname", serialNumber has no short form.
3521 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
3522 therefore remove "mail" short name for "internet 7".
3523 The OID for unique identifiers in X509 certificates is
3524 x500UniqueIdentifier, not uniqueIdentifier.
3525 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
3528 *) Add an "init" command to the ENGINE config module and auto initialize
3529 ENGINEs. Without any "init" command the ENGINE will be initialized
3530 after all ctrl commands have been executed on it. If init=1 the
3531 ENGINE is initailized at that point (ctrls before that point are run
3532 on the uninitialized ENGINE and after on the initialized one). If
3533 init=0 then the ENGINE will not be iniatialized at all.
3536 *) Fix the 'app_verify_callback' interface so that the user-defined
3537 argument is actually passed to the callback: In the
3538 SSL_CTX_set_cert_verify_callback() prototype, the callback
3539 declaration has been changed from
3542 int (*cb)(X509_STORE_CTX *,void *);
3543 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
3544 i=s->ctx->app_verify_callback(&ctx)
3545 has been changed into
3546 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
3548 To update applications using SSL_CTX_set_cert_verify_callback(),
3549 a dummy argument can be added to their callback functions.
3550 [D. K. Smetters <smetters@parc.xerox.com>]
3552 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
3553 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
3555 *) Add and OPENSSL_LOAD_CONF define which will cause
3556 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
3557 This allows older applications to transparently support certain
3558 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
3559 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
3560 load the config file and OPENSSL_add_all_algorithms_conf() which will
3561 always load it have also been added.
3564 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
3565 Adjust NIDs and EVP layer.
3566 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3568 *) Config modules support in openssl utility.
3570 Most commands now load modules from the config file,
3571 though in a few (such as version) this isn't done
3572 because it couldn't be used for anything.
3574 In the case of ca and req the config file used is
3575 the same as the utility itself: that is the -config
3576 command line option can be used to specify an
3580 *) Move default behaviour from OPENSSL_config(). If appname is NULL
3581 use "openssl_conf" if filename is NULL use default openssl config file.
3584 *) Add an argument to OPENSSL_config() to allow the use of an alternative
3585 config section name. Add a new flag to tolerate a missing config file
3586 and move code to CONF_modules_load_file().
3589 *) Support for crypto accelerator cards from Accelerated Encryption
3590 Processing, www.aep.ie. (Use engine 'aep')
3591 The support was copied from 0.9.6c [engine] and adapted/corrected
3592 to work with the new engine framework.
3593 [AEP Inc. and Richard Levitte]
3595 *) Support for SureWare crypto accelerator cards from Baltimore
3596 Technologies. (Use engine 'sureware')
3597 The support was copied from 0.9.6c [engine] and adapted
3598 to work with the new engine framework.
3601 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
3602 make the newer ENGINE framework commands for the CHIL engine work.
3603 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
3605 *) Make it possible to produce shared libraries on ReliantUNIX.
3606 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
3608 *) Add the configuration target debug-linux-ppro.
3609 Make 'openssl rsa' use the general key loading routines
3610 implemented in apps.c, and make those routines able to
3611 handle the key format FORMAT_NETSCAPE and the variant
3613 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3615 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3616 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3618 *) Add -keyform to rsautl, and document -engine.
3619 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
3621 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
3622 BIO_R_NO_SUCH_FILE error code rather than the generic
3623 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
3626 *) Add new functions
3628 ERR_peek_last_error_line
3629 ERR_peek_last_error_line_data.
3630 These are similar to
3633 ERR_peek_error_line_data,
3634 but report on the latest error recorded rather than the first one
3635 still in the error queue.
3636 [Ben Laurie, Bodo Moeller]
3638 *) default_algorithms option in ENGINE config module. This allows things
3640 default_algorithms = ALL
3641 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
3644 *) Prelminary ENGINE config module.
3647 *) New experimental application configuration code.
3650 *) Change the AES code to follow the same name structure as all other
3651 symmetric ciphers, and behave the same way. Move everything to
3652 the directory crypto/aes, thereby obsoleting crypto/rijndael.
3653 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3655 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
3656 [Ben Laurie and Theo de Raadt]
3658 *) Add option to output public keys in req command.
3659 [Massimiliano Pala madwolf@openca.org]
3661 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
3662 (up to about 10% better than before for P-192 and P-224).
3665 *) New functions/macros
3667 SSL_CTX_set_msg_callback(ctx, cb)
3668 SSL_CTX_set_msg_callback_arg(ctx, arg)
3669 SSL_set_msg_callback(ssl, cb)
3670 SSL_set_msg_callback_arg(ssl, arg)
3672 to request calling a callback function
3674 void cb(int write_p, int version, int content_type,
3675 const void *buf, size_t len, SSL *ssl, void *arg)
3677 whenever a protocol message has been completely received
3678 (write_p == 0) or sent (write_p == 1). Here 'version' is the
3679 protocol version according to which the SSL library interprets
3680 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
3681 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
3682 the content type as defined in the SSL 3.0/TLS 1.0 protocol
3683 specification (change_cipher_spec(20), alert(21), handshake(22)).
3684 'buf' and 'len' point to the actual message, 'ssl' to the
3685 SSL object, and 'arg' is the application-defined value set by
3686 SSL[_CTX]_set_msg_callback_arg().
3688 'openssl s_client' and 'openssl s_server' have new '-msg' options
3689 to enable a callback that displays all protocol messages.
3692 *) Change the shared library support so shared libraries are built as
3693 soon as the corresponding static library is finished, and thereby get
3694 openssl and the test programs linked against the shared library.
3695 This still only happens when the keyword "shard" has been given to
3696 the configuration scripts.
3698 NOTE: shared library support is still an experimental thing, and
3699 backward binary compatibility is still not guaranteed.
3700 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
3702 *) Add support for Subject Information Access extension.
3703 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3705 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
3706 additional bytes when new memory had to be allocated, not just
3707 when reusing an existing buffer.
3710 *) New command line and configuration option 'utf8' for the req command.
3711 This allows field values to be specified as UTF8 strings.
3714 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
3715 runs for the former and machine-readable output for the latter.
3718 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
3719 of the e-mail address in the DN (i.e., it will go into a certificate
3720 extension only). The new configuration file option 'email_in_dn = no'
3721 has the same effect.
3722 [Massimiliano Pala madwolf@openca.org]
3724 *) Change all functions with names starting with des_ to be starting
3725 with DES_ instead. Add wrappers that are compatible with libdes,
3726 but are named _ossl_old_des_*. Finally, add macros that map the
3727 des_* symbols to the corresponding _ossl_old_des_* if libdes
3728 compatibility is desired. If OpenSSL 0.9.6c compatibility is
3729 desired, the des_* symbols will be mapped to DES_*, with one
3732 Since we provide two compatibility mappings, the user needs to
3733 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
3734 compatibility is desired. The default (i.e., when that macro
3735 isn't defined) is OpenSSL 0.9.6c compatibility.
3737 There are also macros that enable and disable the support of old
3738 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
3739 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
3740 are defined, the default will apply: to support the old des routines.
3742 In either case, one must include openssl/des.h to get the correct
3743 definitions. Do not try to just include openssl/des_old.h, that
3746 NOTE: This is a major break of an old API into a new one. Software
3747 authors are encouraged to switch to the DES_ style functions. Some
3748 time in the future, des_old.h and the libdes compatibility functions
3749 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
3750 default), and then completely removed.
3753 *) Test for certificates which contain unsupported critical extensions.
3754 If such a certificate is found during a verify operation it is
3755 rejected by default: this behaviour can be overridden by either
3756 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
3757 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
3758 X509_supported_extension() has also been added which returns 1 if a
3759 particular extension is supported.
3762 *) Modify the behaviour of EVP cipher functions in similar way to digests
3763 to retain compatibility with existing code.
3766 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
3767 compatibility with existing code. In particular the 'ctx' parameter does
3768 not have to be to be initialized before the call to EVP_DigestInit() and
3769 it is tidied up after a call to EVP_DigestFinal(). New function
3770 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
3771 EVP_MD_CTX_copy() changed to not require the destination to be
3772 initialized valid and new function EVP_MD_CTX_copy_ex() added which
3773 requires the destination to be valid.
3775 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
3776 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
3779 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
3780 so that complete 'Handshake' protocol structures are kept in memory
3781 instead of overwriting 'msg_type' and 'length' with 'body' data.
3784 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
3785 [Massimo Santin via Richard Levitte]
3787 *) Major restructuring to the underlying ENGINE code. This includes
3788 reduction of linker bloat, separation of pure "ENGINE" manipulation
3789 (initialisation, etc) from functionality dealing with implementations
3790 of specific crypto iterfaces. This change also introduces integrated
3791 support for symmetric ciphers and digest implementations - so ENGINEs
3792 can now accelerate these by providing EVP_CIPHER and EVP_MD
3793 implementations of their own. This is detailed in crypto/engine/README
3794 as it couldn't be adequately described here. However, there are a few
3795 API changes worth noting - some RSA, DSA, DH, and RAND functions that
3796 were changed in the original introduction of ENGINE code have now
3797 reverted back - the hooking from this code to ENGINE is now a good
3798 deal more passive and at run-time, operations deal directly with
3799 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
3800 dereferencing through an ENGINE pointer any more. Also, the ENGINE
3801 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
3802 they were not being used by the framework as there is no concept of a
3803 BIGNUM_METHOD and they could not be generalised to the new
3804 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
3805 ENGINE_cpy() has been removed as it cannot be consistently defined in
3809 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
3812 *) Change mkdef.pl to sort symbols that get the same entry number,
3813 and make sure the automatically generated functions ERR_load_*
3814 become part of libeay.num as well.
3817 *) New function SSL_renegotiate_pending(). This returns true once
3818 renegotiation has been requested (either SSL_renegotiate() call
3819 or HelloRequest/ClientHello receveived from the peer) and becomes
3820 false once a handshake has been completed.
3821 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
3822 sends a HelloRequest, but does not ensure that a handshake takes
3823 place. SSL_renegotiate_pending() is useful for checking if the
3824 client has followed the request.)
3827 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
3828 By default, clients may request session resumption even during
3829 renegotiation (if session ID contexts permit); with this option,
3830 session resumption is possible only in the first handshake.
3832 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
3833 more bits available for options that should not be part of
3834 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
3837 *) Add some demos for certificate and certificate request creation.
3840 *) Make maximum certificate chain size accepted from the peer application
3841 settable (SSL*_get/set_max_cert_list()), as proposed by
3842 "Douglas E. Engert" <deengert@anl.gov>.
3845 *) Add support for shared libraries for Unixware-7
3846 (Boyd Lynn Gerber <gerberb@zenez.com>).
3849 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
3850 be done prior to destruction. Use this to unload error strings from
3851 ENGINEs that load their own error strings. NB: This adds two new API
3852 functions to "get" and "set" this destroy handler in an ENGINE.
3855 *) Alter all existing ENGINE implementations (except "openssl" and
3856 "openbsd") to dynamically instantiate their own error strings. This
3857 makes them more flexible to be built both as statically-linked ENGINEs
3858 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
3859 Also, add stub code to each that makes building them as self-contained
3860 shared-libraries easier (see README.ENGINE).
3863 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
3864 implementations into applications that are completely implemented in
3865 self-contained shared-libraries. The "dynamic" ENGINE exposes control
3866 commands that can be used to configure what shared-library to load and
3867 to control aspects of the way it is handled. Also, made an update to
3868 the README.ENGINE file that brings its information up-to-date and
3869 provides some information and instructions on the "dynamic" ENGINE
3870 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
3873 *) Make it possible to unload ranges of ERR strings with a new
3874 "ERR_unload_strings" function.
3877 *) Add a copy() function to EVP_MD.
3880 *) Make EVP_MD routines take a context pointer instead of just the
3881 md_data void pointer.
3884 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
3885 that the digest can only process a single chunk of data
3886 (typically because it is provided by a piece of
3887 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
3888 is only going to provide a single chunk of data, and hence the
3889 framework needn't accumulate the data for oneshot drivers.
3892 *) As with "ERR", make it possible to replace the underlying "ex_data"
3893 functions. This change also alters the storage and management of global
3894 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
3895 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
3896 index counters. The API functions that use this state have been changed
3897 to take a "class_index" rather than pointers to the class's local STACK
3898 and counter, and there is now an API function to dynamically create new
3899 classes. This centralisation allows us to (a) plug a lot of the
3900 thread-safety problems that existed, and (b) makes it possible to clean
3901 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
3902 such data would previously have always leaked in application code and
3903 workarounds were in place to make the memory debugging turn a blind eye
3904 to it. Application code that doesn't use this new function will still
3905 leak as before, but their memory debugging output will announce it now
3906 rather than letting it slide.
3908 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
3909 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
3910 has a return value to indicate success or failure.
3913 *) Make it possible to replace the underlying "ERR" functions such that the
3914 global state (2 LHASH tables and 2 locks) is only used by the "default"
3915 implementation. This change also adds two functions to "get" and "set"
3916 the implementation prior to it being automatically set the first time
3917 any other ERR function takes place. Ie. an application can call "get",
3918 pass the return value to a module it has just loaded, and that module
3919 can call its own "set" function using that value. This means the
3920 module's "ERR" operations will use (and modify) the error state in the
3921 application and not in its own statically linked copy of OpenSSL code.
3924 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
3925 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
3926 the operation, and provides a more encapsulated way for external code
3927 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
3928 to use these functions rather than manually incrementing the counts.
3930 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
3933 *) Add EVP test program.
3936 *) Add symmetric cipher support to ENGINE. Expect the API to change!
3939 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
3940 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
3941 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
3942 These allow a CRL to be built without having to access X509_CRL fields
3943 directly. Modify 'ca' application to use new functions.
3946 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
3947 bug workarounds. Rollback attack detection is a security feature.
3948 The problem will only arise on OpenSSL servers when TLSv1 is not
3949 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
3950 Software authors not wanting to support TLSv1 will have special reasons
3951 for their choice and can explicitly enable this option.
3952 [Bodo Moeller, Lutz Jaenicke]
3954 *) Rationalise EVP so it can be extended: don't include a union of
3955 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
3956 (similar to those existing for EVP_CIPHER_CTX).
3961 EVP_MD_CTX_init(&md); /* new function call */
3962 EVP_DigestInit(&md, EVP_sha1());
3963 EVP_DigestUpdate(&md, in, len);
3964 EVP_DigestFinal(&md, out, NULL);
3965 EVP_MD_CTX_cleanup(&md); /* new function call */
3969 *) Make DES key schedule conform to the usual scheme, as well as
3970 correcting its structure. This means that calls to DES functions
3971 now have to pass a pointer to a des_key_schedule instead of a
3972 plain des_key_schedule (which was actually always a pointer
3975 des_key_schedule ks;
3977 des_set_key_checked(..., &ks);
3978 des_ncbc_encrypt(..., &ks, ...);
3980 (Note that a later change renames 'des_...' into 'DES_...'.)
3983 *) Initial reduction of linker bloat: the use of some functions, such as
3984 PEM causes large amounts of unused functions to be linked in due to
3985 poor organisation. For example pem_all.c contains every PEM function
3986 which has a knock on effect of linking in large amounts of (unused)
3987 ASN1 code. Grouping together similar functions and splitting unrelated
3988 functions prevents this.
3991 *) Cleanup of EVP macros.
3994 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
3995 correct _ecb suffix.
3998 *) Add initial OCSP responder support to ocsp application. The
3999 revocation information is handled using the text based index
4000 use by the ca application. The responder can either handle
4001 requests generated internally, supplied in files (for example
4002 via a CGI script) or using an internal minimal server.
4005 *) Add configuration choices to get zlib compression for TLS.
4008 *) Changes to Kerberos SSL for RFC 2712 compliance:
4009 1. Implemented real KerberosWrapper, instead of just using
4010 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4011 2. Implemented optional authenticator field of KerberosWrapper.
4013 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4014 and authenticator structs; see crypto/krb5/.
4016 Generalized Kerberos calls to support multiple Kerberos libraries.
4017 [Vern Staats <staatsvr@asc.hpc.mil>,
4018 Jeffrey Altman <jaltman@columbia.edu>
4019 via Richard Levitte]
4021 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4022 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4023 values for each of the key sizes rather than having just
4024 parameters (and 'speed' generating keys each time).
4027 *) Speed up EVP routines.
4030 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4031 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4032 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4033 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4035 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4036 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4037 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4040 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4042 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4045 *) Added the OS2-EMX target.
4046 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4048 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4049 to support NCONF routines in extension code. New function CONF_set_nconf()
4050 to allow functions which take an NCONF to also handle the old LHASH
4051 structure: this means that the old CONF compatible routines can be
4052 retained (in particular wrt extensions) without having to duplicate the
4053 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4056 *) Enhance the general user interface with mechanisms for inner control
4057 and with possibilities to have yes/no kind of prompts.
4060 *) Change all calls to low level digest routines in the library and
4061 applications to use EVP. Add missing calls to HMAC_cleanup() and
4062 don't assume HMAC_CTX can be copied using memcpy().
4063 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4065 *) Add the possibility to control engines through control names but with
4066 arbitrary arguments instead of just a string.
4067 Change the key loaders to take a UI_METHOD instead of a callback
4068 function pointer. NOTE: this breaks binary compatibility with earlier
4069 versions of OpenSSL [engine].
4070 Adapt the nCipher code for these new conditions and add a card insertion
4074 *) Enhance the general user interface with mechanisms to better support
4075 dialog box interfaces, application-defined prompts, the possibility
4076 to use defaults (for example default passwords from somewhere else)
4077 and interrupts/cancellations.
4080 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4081 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4084 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4085 tidy up some unnecessarily weird code in 'sk_new()').
4086 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4088 *) Change the key loading routines for ENGINEs to use the same kind
4089 callback (pem_password_cb) as all other routines that need this
4093 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4094 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4095 than this minimum value is recommended.
4098 *) New random seeder for OpenVMS, using the system process statistics
4099 that are easily reachable.
4102 *) Windows apparently can't transparently handle global
4103 variables defined in DLLs. Initialisations such as:
4105 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4107 wont compile. This is used by the any applications that need to
4108 declare their own ASN1 modules. This was fixed by adding the option
4109 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4110 needed for static libraries under Win32.
4113 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
4114 setting of purpose and trust fields. New X509_STORE trust and
4115 purpose functions and tidy up setting in other SSL functions.
4118 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
4119 structure. These are inherited by X509_STORE_CTX when it is
4120 initialised. This allows various defaults to be set in the
4121 X509_STORE structure (such as flags for CRL checking and custom
4122 purpose or trust settings) for functions which only use X509_STORE_CTX
4123 internally such as S/MIME.
4125 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
4126 trust settings if they are not set in X509_STORE. This allows X509_STORE
4127 purposes and trust (in S/MIME for example) to override any set by default.
4129 Add command line options for CRL checking to smime, s_client and s_server
4133 *) Initial CRL based revocation checking. If the CRL checking flag(s)
4134 are set then the CRL is looked up in the X509_STORE structure and
4135 its validity and signature checked, then if the certificate is found
4136 in the CRL the verify fails with a revoked error.
4138 Various new CRL related callbacks added to X509_STORE_CTX structure.
4140 Command line options added to 'verify' application to support this.
4142 This needs some additional work, such as being able to handle multiple
4143 CRLs with different times, extension based lookup (rather than just
4144 by subject name) and ultimately more complete V2 CRL extension
4148 *) Add a general user interface API (crypto/ui/). This is designed
4149 to replace things like des_read_password and friends (backward
4150 compatibility functions using this new API are provided).
4151 The purpose is to remove prompting functions from the DES code
4152 section as well as provide for prompting through dialog boxes in
4153 a window system and the like.
4156 *) Add "ex_data" support to ENGINE so implementations can add state at a
4157 per-structure level rather than having to store it globally.
4160 *) Make it possible for ENGINE structures to be copied when retrieved by
4161 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
4162 This causes the "original" ENGINE structure to act like a template,
4163 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
4164 operational state can be localised to each ENGINE structure, despite the
4165 fact they all share the same "methods". New ENGINE structures returned in
4166 this case have no functional references and the return value is the single
4167 structural reference. This matches the single structural reference returned
4168 by ENGINE_by_id() normally, when it is incremented on the pre-existing
4172 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
4173 needs to match any other type at all we need to manually clear the
4177 *) Changes to the "openssl engine" utility to include;
4178 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
4179 about an ENGINE's available control commands.
4180 - executing control commands from command line arguments using the
4181 '-pre' and '-post' switches. '-post' is only used if '-t' is
4182 specified and the ENGINE is successfully initialised. The syntax for
4183 the individual commands are colon-separated, for example;
4184 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
4187 *) New dynamic control command support for ENGINEs. ENGINEs can now
4188 declare their own commands (numbers), names (strings), descriptions,
4189 and input types for run-time discovery by calling applications. A
4190 subset of these commands are implicitly classed as "executable"
4191 depending on their input type, and only these can be invoked through
4192 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
4193 can be based on user input, config files, etc). The distinction is
4194 that "executable" commands cannot return anything other than a boolean
4195 result and can only support numeric or string input, whereas some
4196 discoverable commands may only be for direct use through
4197 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
4198 pointers, or other custom uses. The "executable" commands are to
4199 support parameterisations of ENGINE behaviour that can be
4200 unambiguously defined by ENGINEs and used consistently across any
4201 OpenSSL-based application. Commands have been added to all the
4202 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
4203 control over shared-library paths without source code alterations.
4206 *) Changed all ENGINE implementations to dynamically allocate their
4207 ENGINEs rather than declaring them statically. Apart from this being
4208 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
4209 this also allows the implementations to compile without using the
4210 internal engine_int.h header.
4213 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
4214 'const' value. Any code that should be able to modify a RAND_METHOD
4215 should already have non-const pointers to it (ie. they should only
4216 modify their own ones).
4219 *) Made a variety of little tweaks to the ENGINE code.
4220 - "atalla" and "ubsec" string definitions were moved from header files
4221 to C code. "nuron" string definitions were placed in variables
4222 rather than hard-coded - allowing parameterisation of these values
4223 later on via ctrl() commands.
4224 - Removed unused "#if 0"'d code.
4225 - Fixed engine list iteration code so it uses ENGINE_free() to release
4226 structural references.
4227 - Constified the RAND_METHOD element of ENGINE structures.
4228 - Constified various get/set functions as appropriate and added
4229 missing functions (including a catch-all ENGINE_cpy that duplicates
4230 all ENGINE values onto a new ENGINE except reference counts/state).
4231 - Removed NULL parameter checks in get/set functions. Setting a method
4232 or function to NULL is a way of cancelling out a previously set
4233 value. Passing a NULL ENGINE parameter is just plain stupid anyway
4234 and doesn't justify the extra error symbols and code.
4235 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
4236 flags from engine_int.h to engine.h.
4237 - Changed prototypes for ENGINE handler functions (init(), finish(),
4238 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
4241 *) Implement binary inversion algorithm for BN_mod_inverse in addition
4242 to the algorithm using long division. The binary algorithm can be
4243 used only if the modulus is odd. On 32-bit systems, it is faster
4244 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
4245 roughly 5-15% for 256-bit moduli), so we use it only for moduli
4246 up to 450 bits. In 64-bit environments, the binary algorithm
4247 appears to be advantageous for much longer moduli; here we use it
4248 for moduli up to 2048 bits.
4251 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
4252 could not support the combine flag in choice fields.
4255 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
4256 extensions from a certificate request to the certificate.
4259 *) Allow multiple 'certopt' and 'nameopt' options to be separated
4260 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
4261 file: this allows the display of the certificate about to be
4262 signed to be customised, to allow certain fields to be included
4263 or excluded and extension details. The old system didn't display
4264 multicharacter strings properly, omitted fields not in the policy
4265 and couldn't display additional details such as extensions.
4268 *) Function EC_POINTs_mul for multiple scalar multiplication
4269 of an arbitrary number of elliptic curve points
4270 \sum scalars[i]*points[i],
4271 optionally including the generator defined for the EC_GROUP:
4272 scalar*generator + \sum scalars[i]*points[i].
4274 EC_POINT_mul is a simple wrapper function for the typical case
4275 that the point list has just one item (besides the optional
4279 *) First EC_METHODs for curves over GF(p):
4281 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
4282 operations and provides various method functions that can also
4283 operate with faster implementations of modular arithmetic.
4285 EC_GFp_mont_method() reuses most functions that are part of
4286 EC_GFp_simple_method, but uses Montgomery arithmetic.
4288 [Bodo Moeller; point addition and point doubling
4289 implementation directly derived from source code provided by
4290 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
4292 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
4293 crypto/ec/ec_lib.c):
4295 Curves are EC_GROUP objects (with an optional group generator)
4296 based on EC_METHODs that are built into the library.
4298 Points are EC_POINT objects based on EC_GROUP objects.
4300 Most of the framework would be able to handle curves over arbitrary
4301 finite fields, but as there are no obvious types for fields other
4302 than GF(p), some functions are limited to that for now.
4305 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
4306 that the file contains a complete HTTP response.
4309 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
4310 change the def and num file printf format specifier from "%-40sXXX"
4311 to "%-39s XXX". The latter will always guarantee a space after the
4312 field while the former will cause them to run together if the field
4313 is 40 of more characters long.
4316 *) Constify the cipher and digest 'method' functions and structures
4317 and modify related functions to take constant EVP_MD and EVP_CIPHER
4321 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
4322 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
4325 *) Modify EVP_Digest*() routines so they now return values. Although the
4326 internal software routines can never fail additional hardware versions
4330 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
4332 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
4333 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
4336 ERR_R_NESTED_ASN1_ERROR
4338 ERR_R_MISSING_ASN1_EOS
4339 were 4 .. 9, conflicting with
4340 ERR_LIB_RSA (= ERR_R_RSA_LIB)
4342 ERR_LIB_PEM (= ERR_R_PEM_LIB).
4343 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
4345 Add new error code 'ERR_R_INTERNAL_ERROR'.
4348 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
4352 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
4353 sets the subject name for a new request or supersedes the
4354 subject name in a given request. Formats that can be parsed are
4355 'CN=Some Name, OU=myOU, C=IT'
4357 'CN=Some Name/OU=myOU/C=IT'.
4359 Add options '-batch' and '-verbose' to 'openssl req'.
4360 [Massimiliano Pala <madwolf@hackmasters.net>]
4362 *) Introduce the possibility to access global variables through
4363 functions on platform were that's the best way to handle exporting
4364 global variables in shared libraries. To enable this functionality,
4365 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
4366 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
4367 is normally done by Configure or something similar).
4369 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
4370 in the source file (foo.c) like this:
4372 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
4373 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
4375 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
4376 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
4378 OPENSSL_DECLARE_GLOBAL(int,foo);
4379 #define foo OPENSSL_GLOBAL_REF(foo)
4380 OPENSSL_DECLARE_GLOBAL(double,bar);
4381 #define bar OPENSSL_GLOBAL_REF(bar)
4383 The #defines are very important, and therefore so is including the
4384 header file everywhere where the defined globals are used.
4386 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
4387 of ASN.1 items, but that structure is a bit different.
4389 The largest change is in util/mkdef.pl which has been enhanced with
4390 better and easier to understand logic to choose which symbols should
4391 go into the Windows .def files as well as a number of fixes and code
4392 cleanup (among others, algorithm keywords are now sorted
4393 lexicographically to avoid constant rewrites).
4396 *) In BN_div() keep a copy of the sign of 'num' before writing the
4397 result to 'rm' because if rm==num the value will be overwritten
4398 and produce the wrong result if 'num' is negative: this caused
4399 problems with BN_mod() and BN_nnmod().
4402 *) Function OCSP_request_verify(). This checks the signature on an
4403 OCSP request and verifies the signer certificate. The signer
4404 certificate is just checked for a generic purpose and OCSP request
4408 *) Add OCSP_check_validity() function to check the validity of OCSP
4409 responses. OCSP responses are prepared in real time and may only
4410 be a few seconds old. Simply checking that the current time lies
4411 between thisUpdate and nextUpdate max reject otherwise valid responses
4412 caused by either OCSP responder or client clock inaccuracy. Instead
4413 we allow thisUpdate and nextUpdate to fall within a certain period of
4414 the current time. The age of the response can also optionally be
4415 checked. Two new options -validity_period and -status_age added to
4419 *) If signature or public key algorithm is unrecognized print out its
4420 OID rather that just UNKNOWN.
4423 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
4424 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
4425 ID to be generated from the issuer certificate alone which can then be
4426 passed to OCSP_id_issuer_cmp().
4429 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
4430 ASN1 modules to export functions returning ASN1_ITEM pointers
4431 instead of the ASN1_ITEM structures themselves. This adds several
4432 new macros which allow the underlying ASN1 function/structure to
4433 be accessed transparently. As a result code should not use ASN1_ITEM
4434 references directly (such as &X509_it) but instead use the relevant
4435 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
4436 use of the new ASN1 code on platforms where exporting structures
4437 is problematical (for example in shared libraries) but exporting
4438 functions returning pointers to structures is not.
4441 *) Add support for overriding the generation of SSL/TLS session IDs.
4442 These callbacks can be registered either in an SSL_CTX or per SSL.
4443 The purpose of this is to allow applications to control, if they wish,
4444 the arbitrary values chosen for use as session IDs, particularly as it
4445 can be useful for session caching in multiple-server environments. A
4446 command-line switch for testing this (and any client code that wishes
4447 to use such a feature) has been added to "s_server".
4448 [Geoff Thorpe, Lutz Jaenicke]
4450 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
4451 of the form '#if defined(...) || defined(...) || ...' and
4452 '#if !defined(...) && !defined(...) && ...'. This also avoids
4453 the growing number of special cases it was previously handling.
4456 *) Make all configuration macros available for application by making
4457 sure they are available in opensslconf.h, by giving them names starting
4458 with "OPENSSL_" to avoid conflicts with other packages and by making
4459 sure e_os2.h will cover all platform-specific cases together with
4461 Additionally, it is now possible to define configuration/platform-
4462 specific names (called "system identities"). In the C code, these
4463 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
4464 macro with the name beginning with "OPENSSL_SYS_", which is determined
4465 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
4469 *) New option -set_serial to 'req' and 'x509' this allows the serial
4470 number to use to be specified on the command line. Previously self
4471 signed certificates were hard coded with serial number 0 and the
4472 CA options of 'x509' had to use a serial number in a file which was
4476 *) New options to 'ca' utility to support V2 CRL entry extensions.
4477 Currently CRL reason, invalidity date and hold instruction are
4478 supported. Add new CRL extensions to V3 code and some new objects.
4481 *) New function EVP_CIPHER_CTX_set_padding() this is used to
4482 disable standard block padding (aka PKCS#5 padding) in the EVP
4483 API, which was previously mandatory. This means that the data is
4484 not padded in any way and so the total length much be a multiple
4485 of the block size, otherwise an error occurs.
4488 *) Initial (incomplete) OCSP SSL support.
4491 *) New function OCSP_parse_url(). This splits up a URL into its host,
4492 port and path components: primarily to parse OCSP URLs. New -url
4493 option to ocsp utility.
4496 *) New nonce behavior. The return value of OCSP_check_nonce() now
4497 reflects the various checks performed. Applications can decide
4498 whether to tolerate certain situations such as an absent nonce
4499 in a response when one was present in a request: the ocsp application
4500 just prints out a warning. New function OCSP_add1_basic_nonce()
4501 this is to allow responders to include a nonce in a response even if
4502 the request is nonce-less.
4505 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
4506 skipped when using openssl x509 multiple times on a single input file,
4507 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
4510 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
4511 set string type: to handle setting ASN1_TIME structures. Fix ca
4512 utility to correctly initialize revocation date of CRLs.
4515 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
4516 the clients preferred ciphersuites and rather use its own preferences.
4517 Should help to work around M$ SGC (Server Gated Cryptography) bug in
4518 Internet Explorer by ensuring unchanged hash method during stepup.
4519 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
4522 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
4523 to aes and add a new 'exist' option to print out symbols that don't
4527 *) Additional options to ocsp utility to allow flags to be set and
4528 additional certificates supplied.
4531 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
4532 OCSP client a number of certificate to only verify the response
4536 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
4537 handle the new API. Currently only ECB, CBC modes supported. Add new
4540 Add TLS AES ciphersuites as described in RFC3268, "Advanced
4541 Encryption Standard (AES) Ciphersuites for Transport Layer
4542 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
4543 not enabled by default and were not part of the "ALL" ciphersuite
4544 alias because they were not yet official; they could be
4545 explicitly requested by specifying the "AESdraft" ciphersuite
4546 group alias. In the final release of OpenSSL 0.9.7, the group
4547 alias is called "AES" and is part of "ALL".)
4548 [Ben Laurie, Steve Henson, Bodo Moeller]
4550 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
4551 request to response.
4554 *) Functions for OCSP responders. OCSP_request_onereq_count(),
4555 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
4556 extract information from a certificate request. OCSP_response_create()
4557 creates a response and optionally adds a basic response structure.
4558 OCSP_basic_add1_status() adds a complete single response to a basic
4559 response and returns the OCSP_SINGLERESP structure just added (to allow
4560 extensions to be included for example). OCSP_basic_add1_cert() adds a
4561 certificate to a basic response and OCSP_basic_sign() signs a basic
4562 response with various flags. New helper functions ASN1_TIME_check()
4563 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
4564 (converts ASN1_TIME to GeneralizedTime).
4567 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
4568 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
4569 structure from a certificate. X509_pubkey_digest() digests the public_key
4570 contents: this is used in various key identifiers.
4573 *) Make sk_sort() tolerate a NULL argument.
4574 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
4576 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
4577 passed by the function are trusted implicitly. If any of them signed the
4578 response then it is assumed to be valid and is not verified.
4581 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
4582 to data. This was previously part of the PKCS7 ASN1 code. This
4583 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
4584 [Steve Henson, reported by Kenneth R. Robinette
4585 <support@securenetterm.com>]
4587 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
4588 routines: without these tracing memory leaks is very painful.
4589 Fix leaks in PKCS12 and PKCS7 routines.
4592 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
4593 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
4594 effectively meant GeneralizedTime would never be used. Now it
4595 is initialised to -1 but X509_time_adj() now has to check the value
4596 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
4597 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
4598 [Steve Henson, reported by Kenneth R. Robinette
4599 <support@securenetterm.com>]
4601 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
4602 result in a zero length in the ASN1_INTEGER structure which was
4603 not consistent with the structure when d2i_ASN1_INTEGER() was used
4604 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
4605 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
4606 where it did not print out a minus for negative ASN1_INTEGER.
4609 *) Add summary printout to ocsp utility. The various functions which
4610 convert status values to strings have been renamed to:
4611 OCSP_response_status_str(), OCSP_cert_status_str() and
4612 OCSP_crl_reason_str() and are no longer static. New options
4613 to verify nonce values and to disable verification. OCSP response
4614 printout format cleaned up.
4617 *) Add additional OCSP certificate checks. These are those specified
4618 in RFC2560. This consists of two separate checks: the CA of the
4619 certificate being checked must either be the OCSP signer certificate
4620 or the issuer of the OCSP signer certificate. In the latter case the
4621 OCSP signer certificate must contain the OCSP signing extended key
4622 usage. This check is performed by attempting to match the OCSP
4623 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
4624 in the OCSP_CERTID structures of the response.
4627 *) Initial OCSP certificate verification added to OCSP_basic_verify()
4628 and related routines. This uses the standard OpenSSL certificate
4629 verify routines to perform initial checks (just CA validity) and
4630 to obtain the certificate chain. Then additional checks will be
4631 performed on the chain. Currently the root CA is checked to see
4632 if it is explicitly trusted for OCSP signing. This is used to set
4633 a root CA as a global signing root: that is any certificate that
4634 chains to that CA is an acceptable OCSP signing certificate.
4637 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
4638 extensions from a separate configuration file.
4639 As when reading extensions from the main configuration file,
4640 the '-extensions ...' option may be used for specifying the
4642 [Massimiliano Pala <madwolf@comune.modena.it>]
4644 *) New OCSP utility. Allows OCSP requests to be generated or
4645 read. The request can be sent to a responder and the output
4646 parsed, outputed or printed in text form. Not complete yet:
4647 still needs to check the OCSP response validity.
4650 *) New subcommands for 'openssl ca':
4651 'openssl ca -status <serial>' prints the status of the cert with
4652 the given serial number (according to the index file).
4653 'openssl ca -updatedb' updates the expiry status of certificates
4655 [Massimiliano Pala <madwolf@comune.modena.it>]
4657 *) New '-newreq-nodes' command option to CA.pl. This is like
4658 '-newreq', but calls 'openssl req' with the '-nodes' option
4659 so that the resulting key is not encrypted.
4660 [Damien Miller <djm@mindrot.org>]
4662 *) New configuration for the GNU Hurd.
4663 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
4665 *) Initial code to implement OCSP basic response verify. This
4666 is currently incomplete. Currently just finds the signer's
4667 certificate and verifies the signature on the response.
4670 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
4671 value of OPENSSLDIR. This is available via the new '-d' option
4672 to 'openssl version', and is also included in 'openssl version -a'.
4675 *) Allowing defining memory allocation callbacks that will be given
4676 file name and line number information in additional arguments
4677 (a const char* and an int). The basic functionality remains, as
4678 well as the original possibility to just replace malloc(),
4679 realloc() and free() by functions that do not know about these
4680 additional arguments. To register and find out the current
4681 settings for extended allocation functions, the following
4682 functions are provided:
4684 CRYPTO_set_mem_ex_functions
4685 CRYPTO_set_locked_mem_ex_functions
4686 CRYPTO_get_mem_ex_functions
4687 CRYPTO_get_locked_mem_ex_functions
4689 These work the same way as CRYPTO_set_mem_functions and friends.
4690 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
4691 extended allocation function is enabled.
4692 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
4693 a conventional allocation function is enabled.
4694 [Richard Levitte, Bodo Moeller]
4696 *) Finish off removing the remaining LHASH function pointer casts.
4697 There should no longer be any prototype-casting required when using
4698 the LHASH abstraction, and any casts that remain are "bugs". See
4699 the callback types and macros at the head of lhash.h for details
4700 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
4703 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
4704 If /dev/[u]random devices are not available or do not return enough
4705 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
4707 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
4708 /etc/entropy will be queried once each in this sequence, quering stops
4709 when enough entropy was collected without querying more sockets.
4712 *) Change the Unix RAND_poll() variant to be able to poll several
4713 random devices, as specified by DEVRANDOM, until a sufficient amount
4714 of data has been collected. We spend at most 10 ms on each file
4715 (select timeout) and read in non-blocking mode. DEVRANDOM now
4716 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
4717 (previously it was just the string "/dev/urandom"), so on typical
4718 platforms the 10 ms delay will never occur.
4719 Also separate out the Unix variant to its own file, rand_unix.c.
4720 For VMS, there's a currently-empty rand_vms.c.
4723 *) Move OCSP client related routines to ocsp_cl.c. These
4724 provide utility functions which an application needing
4725 to issue a request to an OCSP responder and analyse the
4726 response will typically need: as opposed to those which an
4727 OCSP responder itself would need which will be added later.
4729 OCSP_request_sign() signs an OCSP request with an API similar
4730 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
4731 response. OCSP_response_get1_basic() extracts basic response
4732 from response. OCSP_resp_find_status(): finds and extracts status
4733 information from an OCSP_CERTID structure (which will be created
4734 when the request structure is built). These are built from lower
4735 level functions which work on OCSP_SINGLERESP structures but
4736 wont normally be used unless the application wishes to examine
4737 extensions in the OCSP response for example.
4739 Replace nonce routines with a pair of functions.
4740 OCSP_request_add1_nonce() adds a nonce value and optionally
4741 generates a random value. OCSP_check_nonce() checks the
4742 validity of the nonce in an OCSP response.
4745 *) Change function OCSP_request_add() to OCSP_request_add0_id().
4746 This doesn't copy the supplied OCSP_CERTID and avoids the
4747 need to free up the newly created id. Change return type
4748 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
4749 This can then be used to add extensions to the request.
4750 Deleted OCSP_request_new(), since most of its functionality
4751 is now in OCSP_REQUEST_new() (and the case insensitive name
4752 clash) apart from the ability to set the request name which
4753 will be added elsewhere.
4756 *) Update OCSP API. Remove obsolete extensions argument from
4757 various functions. Extensions are now handled using the new
4758 OCSP extension code. New simple OCSP HTTP function which
4759 can be used to send requests and parse the response.
4762 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
4763 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
4764 uses the special reorder version of SET OF to sort the attributes
4765 and reorder them to match the encoded order. This resolves a long
4766 standing problem: a verify on a PKCS7 structure just after signing
4767 it used to fail because the attribute order did not match the
4768 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
4769 it uses the received order. This is necessary to tolerate some broken
4770 software that does not order SET OF. This is handled by encoding
4771 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
4772 to produce the required SET OF.
4775 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
4776 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
4777 files to get correct declarations of the ASN.1 item variables.
4780 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
4781 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
4782 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
4783 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
4784 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
4785 ASN1_ITEM and no wrapper functions.
4788 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
4789 replace the old function pointer based I/O routines. Change most of
4790 the *_d2i_bio() and *_d2i_fp() functions to use these.
4793 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
4794 lines, recognice more "algorithms" that can be deselected, and make
4795 it complain about algorithm deselection that isn't recognised.
4798 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
4799 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
4800 to use new functions. Add NO_ASN1_OLD which can be set to remove
4801 some old style ASN1 functions: this can be used to determine if old
4802 code will still work when these eventually go away.
4805 *) New extension functions for OCSP structures, these follow the
4806 same conventions as certificates and CRLs.
4809 *) New function X509V3_add1_i2d(). This automatically encodes and
4810 adds an extension. Its behaviour can be customised with various
4811 flags to append, replace or delete. Various wrappers added for
4812 certifcates and CRLs.
4815 *) Fix to avoid calling the underlying ASN1 print routine when
4816 an extension cannot be parsed. Correct a typo in the
4817 OCSP_SERVICELOC extension. Tidy up print OCSP format.
4820 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
4821 entries for variables.
4824 *) Add functionality to apps/openssl.c for detecting locking
4825 problems: As the program is single-threaded, all we have
4826 to do is register a locking callback using an array for
4827 storing which locks are currently held by the program.
4830 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
4831 SSL_get_ex_data_X509_STORE_idx(), which is used in
4832 ssl_verify_cert_chain() and thus can be called at any time
4833 during TLS/SSL handshakes so that thread-safety is essential.
4834 Unfortunately, the ex_data design is not at all suited
4835 for multi-threaded use, so it probably should be abolished.
4838 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
4839 [Broadcom, tweaked and integrated by Geoff Thorpe]
4841 *) Move common extension printing code to new function
4842 X509V3_print_extensions(). Reorganise OCSP print routines and
4843 implement some needed OCSP ASN1 functions. Add OCSP extensions.
4846 *) New function X509_signature_print() to remove duplication in some
4850 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
4851 set (this was treated exactly the same as SET OF previously). This
4852 is used to reorder the STACK representing the structure to match the
4853 encoding. This will be used to get round a problem where a PKCS7
4854 structure which was signed could not be verified because the STACK
4855 order did not reflect the encoded order.
4858 *) Reimplement the OCSP ASN1 module using the new code.
4861 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
4862 for its ASN1 operations. The old style function pointers still exist
4863 for now but they will eventually go away.
4866 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
4867 completely replaces the old ASN1 functionality with a table driven
4868 encoder and decoder which interprets an ASN1_ITEM structure describing
4869 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
4870 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
4871 has also been converted to the new form.
4874 *) Change BN_mod_exp_recp so that negative moduli are tolerated
4875 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
4876 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
4877 for negative moduli.
4880 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
4881 of not touching the result's sign bit.
4884 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
4888 *) Changed the LHASH code to use prototypes for callbacks, and created
4889 macros to declare and implement thin (optionally static) functions
4890 that provide type-safety and avoid function pointer casting for the
4891 type-specific callbacks.
4894 *) Added Kerberos Cipher Suites to be used with TLS, as written in
4896 [Veers Staats <staatsvr@asc.hpc.mil>,
4897 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
4899 *) Reformat the FAQ so the different questions and answers can be divided
4900 in sections depending on the subject.
4903 *) Have the zlib compression code load ZLIB.DLL dynamically under
4907 *) New function BN_mod_sqrt for computing square roots modulo a prime
4908 (using the probabilistic Tonelli-Shanks algorithm unless
4909 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
4910 be handled deterministically).
4911 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
4913 *) Make BN_mod_inverse faster by explicitly handling small quotients
4914 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
4915 512 bits], about 30% for larger ones [1024 or 2048 bits].)
4918 *) New function BN_kronecker.
4921 *) Fix BN_gcd so that it works on negative inputs; the result is
4922 positive unless both parameters are zero.
4923 Previously something reasonably close to an infinite loop was
4924 possible because numbers could be growing instead of shrinking
4925 in the implementation of Euclid's algorithm.
4928 *) Fix BN_is_word() and BN_is_one() macros to take into account the
4929 sign of the number in question.
4931 Fix BN_is_word(a,w) to work correctly for w == 0.
4933 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
4934 because its test if the absolute value of 'a' equals 'w'.
4935 Note that BN_abs_is_word does *not* handle w == 0 reliably;
4936 it exists mostly for use in the implementations of BN_is_zero(),
4937 BN_is_one(), and BN_is_word().
4940 *) New function BN_swap.
4943 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
4944 the exponentiation functions are more likely to produce reasonable
4945 results on negative inputs.
4948 *) Change BN_mod_mul so that the result is always non-negative.
4949 Previously, it could be negative if one of the factors was negative;
4950 I don't think anyone really wanted that behaviour.
4953 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
4954 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
4955 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
4956 and add new functions:
4965 BN_mod_lshift1_quick
4969 These functions always generate non-negative results.
4971 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
4972 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
4974 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
4975 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
4976 be reduced modulo m.
4977 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
4980 The following entry accidentily appeared in the CHANGES file
4981 distributed with OpenSSL 0.9.7. The modifications described in
4982 it do *not* apply to OpenSSL 0.9.7.
4984 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4985 was actually never needed) and in BN_mul(). The removal in BN_mul()
4986 required a small change in bn_mul_part_recursive() and the addition
4987 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4988 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4989 bn_sub_words() and bn_add_words() except they take arrays with
4994 *) In 'openssl passwd', verify passwords read from the terminal
4995 unless the '-salt' option is used (which usually means that
4996 verification would just waste user's time since the resulting
4997 hash is going to be compared with some given password hash)
4998 or the new '-noverify' option is used.
5000 This is an incompatible change, but it does not affect
5001 non-interactive use of 'openssl passwd' (passwords on the command
5002 line, '-stdin' option, '-in ...' option) and thus should not
5006 *) Remove all references to RSAref, since there's no more need for it.
5009 *) Make DSO load along a path given through an environment variable
5010 (SHLIB_PATH) with shl_load().
5013 *) Constify the ENGINE code as a result of BIGNUM constification.
5014 Also constify the RSA code and most things related to it. In a
5015 few places, most notable in the depth of the ASN.1 code, ugly
5016 casts back to non-const were required (to be solved at a later
5020 *) Make it so the openssl application has all engines loaded by default.
5023 *) Constify the BIGNUM routines a little more.
5026 *) Add the following functions:
5028 ENGINE_load_cswift()
5030 ENGINE_load_atalla()
5032 ENGINE_load_builtin_engines()
5034 That way, an application can itself choose if external engines that
5035 are built-in in OpenSSL shall ever be used or not. The benefit is
5036 that applications won't have to be linked with libdl or other dso
5037 libraries unless it's really needed.
5039 Changed 'openssl engine' to load all engines on demand.
5040 Changed the engine header files to avoid the duplication of some
5041 declarations (they differed!).
5044 *) 'openssl engine' can now list capabilities.
5047 *) Better error reporting in 'openssl engine'.
5050 *) Never call load_dh_param(NULL) in s_server.
5053 *) Add engine application. It can currently list engines by name and
5054 identity, and test if they are actually available.
5057 *) Improve RPM specification file by forcing symbolic linking and making
5058 sure the installed documentation is also owned by root.root.
5059 [Damien Miller <djm@mindrot.org>]
5061 *) Give the OpenSSL applications more possibilities to make use of
5062 keys (public as well as private) handled by engines.
5065 *) Add OCSP code that comes from CertCo.
5068 *) Add VMS support for the Rijndael code.
5071 *) Added untested support for Nuron crypto accelerator.
5074 *) Add support for external cryptographic devices. This code was
5075 previously distributed separately as the "engine" branch.
5076 [Geoff Thorpe, Richard Levitte]
5078 *) Rework the filename-translation in the DSO code. It is now possible to
5079 have far greater control over how a "name" is turned into a filename
5080 depending on the operating environment and any oddities about the
5081 different shared library filenames on each system.
5084 *) Support threads on FreeBSD-elf in Configure.
5087 *) Fix for SHA1 assembly problem with MASM: it produces
5088 warnings about corrupt line number information when assembling
5089 with debugging information. This is caused by the overlapping
5091 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5094 NCONF_get_number() has no error checking at all. As a replacement,
5095 NCONF_get_number_e() is defined (_e for "error checking") and is
5096 promoted strongly. The old NCONF_get_number is kept around for
5097 binary backward compatibility.
5098 Make it possible for methods to load from something other than a BIO,
5099 by providing a function pointer that is given a name instead of a BIO.
5100 For example, this could be used to load configuration data from an
5104 *) Fix for non blocking accept BIOs. Added new I/O special reason
5105 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5106 with non blocking I/O was not possible because no retry code was
5107 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5111 *) Added the beginnings of Rijndael support.
5114 *) Fix for bug in DirectoryString mask setting. Add support for
5115 X509_NAME_print_ex() in 'req' and X509_print_ex() function
5116 to allow certificate printing to more controllable, additional
5117 'certopt' option to 'x509' to allow new printing options to be
5121 *) Clean old EAY MD5 hack from e_os.h.
5124 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5126 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5127 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5128 [Joe Orton, Steve Henson]
5130 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5132 *) Fix additional bug revealed by the NISCC test suite:
5134 Stop bug triggering large recursion when presented with
5135 certain ASN.1 tags (CVE-2003-0851)
5138 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5140 *) Fix various bugs revealed by running the NISCC test suite:
5142 Stop out of bounds reads in the ASN1 code when presented with
5143 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5145 If verify callback ignores invalid public key errors don't try to check
5146 certificate signature with the NULL public key.
5150 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5151 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5155 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5156 extra data after the compression methods not only for TLS 1.0
5157 but also for SSL 3.0 (as required by the specification).
5158 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5160 *) Change X509_certificate_type() to mark the key as exported/exportable
5161 when it's 512 *bits* long, not 512 bytes.
5164 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5166 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5167 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5168 a protocol version number mismatch like a decryption error
5169 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5172 *) Turn on RSA blinding by default in the default implementation
5173 to avoid a timing attack. Applications that don't want it can call
5174 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5175 They would be ill-advised to do so in most cases.
5176 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5178 *) Change RSA blinding code so that it works when the PRNG is not
5179 seeded (in this case, the secret RSA exponent is abused as
5180 an unpredictable seed -- if it is not unpredictable, there
5181 is no point in blinding anyway). Make RSA blinding thread-safe
5182 by remembering the creator's thread ID in rsa->blinding and
5183 having all other threads use local one-time blinding factors
5184 (this requires more computation than sharing rsa->blinding, but
5185 avoids excessive locking; and if an RSA object is not shared
5186 between threads, blinding will still be very fast).
5189 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5191 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5192 via timing by performing a MAC computation even if incorrrect
5193 block cipher padding has been found. This is a countermeasure
5194 against active attacks where the attacker has to distinguish
5195 between bad padding and a MAC verification error. (CVE-2003-0078)
5197 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5198 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5199 Martin Vuagnoux (EPFL, Ilion)]
5201 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
5203 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
5204 memory from it's contents. This is done with a counter that will
5205 place alternating values in each byte. This can be used to solve
5206 two issues: 1) the removal of calls to memset() by highly optimizing
5207 compilers, and 2) cleansing with other values than 0, since those can
5208 be read through on certain media, for example a swap space on disk.
5211 *) Bugfix: client side session caching did not work with external caching,
5212 because the session->cipher setting was not restored when reloading
5213 from the external cache. This problem was masked, when
5214 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
5215 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
5218 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
5219 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
5220 [Zeev Lieber <zeev-l@yahoo.com>]
5222 *) Undo an undocumented change introduced in 0.9.6e which caused
5223 repeated calls to OpenSSL_add_all_ciphers() and
5224 OpenSSL_add_all_digests() to be ignored, even after calling
5228 *) Change the default configuration reader to deal with last line not
5229 being properly terminated.
5232 *) Change X509_NAME_cmp() so it applies the special rules on handling
5233 DN values that are of type PrintableString, as well as RDNs of type
5234 emailAddress where the value has the type ia5String.
5235 [stefank@valicert.com via Richard Levitte]
5237 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
5238 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
5239 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
5240 the bitwise-OR of the two for use by the majority of applications
5241 wanting this behaviour, and update the docs. The documented
5242 behaviour and actual behaviour were inconsistent and had been
5243 changing anyway, so this is more a bug-fix than a behavioural
5245 [Geoff Thorpe, diagnosed by Nadav Har'El]
5247 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
5248 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
5251 *) Fix initialization code race conditions in
5252 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
5253 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
5254 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
5255 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
5256 ssl2_get_cipher_by_char(),
5257 ssl3_get_cipher_by_char().
5258 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
5260 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
5261 the cached sessions are flushed, as the remove_cb() might use ex_data
5262 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
5263 (see [openssl.org #212]).
5264 [Geoff Thorpe, Lutz Jaenicke]
5266 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
5267 length, instead of the encoding length to d2i_ASN1_OBJECT.
5270 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
5272 *) [In 0.9.6g-engine release:]
5273 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
5274 [Lynn Gazis <lgazis@rainbow.com>]
5276 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
5278 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
5279 and get fix the header length calculation.
5280 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
5281 Alon Kantor <alonk@checkpoint.com> (and others),
5284 *) Use proper error handling instead of 'assertions' in buffer
5285 overflow checks added in 0.9.6e. This prevents DoS (the
5286 assertions could call abort()).
5287 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
5289 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
5291 *) Add various sanity checks to asn1_get_length() to reject
5292 the ASN1 length bytes if they exceed sizeof(long), will appear
5293 negative or the content length exceeds the length of the
5295 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5297 *) Fix cipher selection routines: ciphers without encryption had no flags
5298 for the cipher strength set and where therefore not handled correctly
5299 by the selection routines (PR #130).
5302 *) Fix EVP_dsa_sha macro.
5306 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
5307 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
5308 that was added in OpenSSL 0.9.6d.
5310 As the countermeasure turned out to be incompatible with some
5311 broken SSL implementations, the new option is part of SSL_OP_ALL.
5312 SSL_OP_ALL is usually employed when compatibility with weird SSL
5313 implementations is desired (e.g. '-bugs' option to 's_client' and
5314 's_server'), so the new option is automatically set in many
5318 *) Changes in security patch:
5320 Changes marked "(CHATS)" were sponsored by the Defense Advanced
5321 Research Projects Agency (DARPA) and Air Force Research Laboratory,
5322 Air Force Materiel Command, USAF, under agreement number
5325 *) Add various sanity checks to asn1_get_length() to reject
5326 the ASN1 length bytes if they exceed sizeof(long), will appear
5327 negative or the content length exceeds the length of the
5328 supplied buffer. (CVE-2002-0659)
5329 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5331 *) Assertions for various potential buffer overflows, not known to
5333 [Ben Laurie (CHATS)]
5335 *) Various temporary buffers to hold ASCII versions of integers were
5336 too small for 64 bit platforms. (CVE-2002-0655)
5337 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
5339 *) Remote buffer overflow in SSL3 protocol - an attacker could
5340 supply an oversized session ID to a client. (CVE-2002-0656)
5341 [Ben Laurie (CHATS)]
5343 *) Remote buffer overflow in SSL2 protocol - an attacker could
5344 supply an oversized client master key. (CVE-2002-0656)
5345 [Ben Laurie (CHATS)]
5347 Changes between 0.9.6c and 0.9.6d [9 May 2002]
5349 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
5350 encoded as NULL) with id-dsa-with-sha1.
5351 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
5353 *) Check various X509_...() return values in apps/req.c.
5354 [Nils Larsch <nla@trustcenter.de>]
5356 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
5357 an end-of-file condition would erronously be flagged, when the CRLF
5358 was just at the end of a processed block. The bug was discovered when
5359 processing data through a buffering memory BIO handing the data to a
5360 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
5361 <ptsekov@syntrex.com> and Nedelcho Stanev.
5364 *) Implement a countermeasure against a vulnerability recently found
5365 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
5366 before application data chunks to avoid the use of known IVs
5367 with data potentially chosen by the attacker.
5370 *) Fix length checks in ssl3_get_client_hello().
5373 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
5374 to prevent ssl3_read_internal() from incorrectly assuming that
5375 ssl3_read_bytes() found application data while handshake
5376 processing was enabled when in fact s->s3->in_read_app_data was
5377 merely automatically cleared during the initial handshake.
5378 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
5380 *) Fix object definitions for Private and Enterprise: they were not
5381 recognized in their shortname (=lowercase) representation. Extend
5382 obj_dat.pl to issue an error when using undefined keywords instead
5383 of silently ignoring the problem (Svenning Sorensen
5384 <sss@sss.dnsalias.net>).
5387 *) Fix DH_generate_parameters() so that it works for 'non-standard'
5388 generators, i.e. generators other than 2 and 5. (Previously, the
5389 code did not properly initialise the 'add' and 'rem' values to
5390 BN_generate_prime().)
5392 In the new general case, we do not insist that 'generator' is
5393 actually a primitive root: This requirement is rather pointless;
5394 a generator of the order-q subgroup is just as good, if not
5398 *) Map new X509 verification errors to alerts. Discovered and submitted by
5399 Tom Wu <tom@arcot.com>.
5402 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
5403 returning non-zero before the data has been completely received
5404 when using non-blocking I/O.
5405 [Bodo Moeller; problem pointed out by John Hughes]
5407 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
5408 [Ben Laurie, Lutz Jaenicke]
5410 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
5411 Yoram Zahavi <YoramZ@gilian.com>).
5414 *) Add information about CygWin 1.3 and on, and preserve proper
5415 configuration for the versions before that.
5416 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5418 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
5419 check whether we deal with a copy of a session and do not delete from
5420 the cache in this case. Problem reported by "Izhar Shoshani Levi"
5421 <izhar@checkpoint.com>.
5424 *) Do not store session data into the internal session cache, if it
5425 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
5426 flag is set). Proposed by Aslam <aslam@funk.com>.
5429 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
5433 *) [In 0.9.6d-engine release:]
5434 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5435 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5437 *) Add the configuration target linux-s390x.
5438 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
5440 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
5441 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
5442 variable as an indication that a ClientHello message has been
5443 received. As the flag value will be lost between multiple
5444 invocations of ssl3_accept when using non-blocking I/O, the
5445 function may not be aware that a handshake has actually taken
5446 place, thus preventing a new session from being added to the
5449 To avoid this problem, we now set s->new_session to 2 instead of
5450 using a local variable.
5451 [Lutz Jaenicke, Bodo Moeller]
5453 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
5454 if the SSL_R_LENGTH_MISMATCH error is detected.
5455 [Geoff Thorpe, Bodo Moeller]
5457 *) New 'shared_ldflag' column in Configure platform table.
5460 *) Fix EVP_CIPHER_mode macro.
5461 ["Dan S. Camper" <dan@bti.net>]
5463 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
5464 type, we must throw them away by setting rr->length to 0.
5465 [D P Chang <dpc@qualys.com>]
5467 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
5469 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
5470 <Dominikus.Scherkl@biodata.com>. (The previous implementation
5471 worked incorrectly for those cases where range = 10..._2 and
5472 3*range is two bits longer than range.)
5475 *) Only add signing time to PKCS7 structures if it is not already
5479 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
5480 OBJ_ld_ce should be OBJ_id_ce.
5481 Also some ip-pda OIDs in crypto/objects/objects.txt were
5482 incorrect (cf. RFC 3039).
5483 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
5485 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
5486 returns early because it has nothing to do.
5487 [Andy Schneider <andy.schneider@bjss.co.uk>]
5489 *) [In 0.9.6c-engine release:]
5490 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
5491 [Andy Schneider <andy.schneider@bjss.co.uk>]
5493 *) [In 0.9.6c-engine release:]
5494 Add support for Cryptographic Appliance's keyserver technology.
5495 (Use engine 'keyclient')
5496 [Cryptographic Appliances and Geoff Thorpe]
5498 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
5499 is called via tools/c89.sh because arguments have to be
5500 rearranged (all '-L' options must appear before the first object
5502 [Richard Shapiro <rshapiro@abinitio.com>]
5504 *) [In 0.9.6c-engine release:]
5505 Add support for Broadcom crypto accelerator cards, backported
5507 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
5509 *) [In 0.9.6c-engine release:]
5510 Add support for SureWare crypto accelerator cards from
5511 Baltimore Technologies. (Use engine 'sureware')
5512 [Baltimore Technologies and Mark Cox]
5514 *) [In 0.9.6c-engine release:]
5515 Add support for crypto accelerator cards from Accelerated
5516 Encryption Processing, www.aep.ie. (Use engine 'aep')
5517 [AEP Inc. and Mark Cox]
5519 *) Add a configuration entry for gcc on UnixWare.
5520 [Gary Benson <gbenson@redhat.com>]
5522 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
5523 messages are stored in a single piece (fixed-length part and
5524 variable-length part combined) and fix various bugs found on the way.
5527 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
5528 instead. BIO_gethostbyname() does not know what timeouts are
5529 appropriate, so entries would stay in cache even when they have
5531 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
5533 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
5534 faced with a pathologically small ClientHello fragment that does
5535 not contain client_version: Instead of aborting with an error,
5536 simply choose the highest available protocol version (i.e.,
5537 TLS 1.0 unless it is disabled). In practice, ClientHello
5538 messages are never sent like this, but this change gives us
5539 strictly correct behaviour at least for TLS.
5542 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
5543 never resets s->method to s->ctx->method when called from within
5544 one of the SSL handshake functions.
5545 [Bodo Moeller; problem pointed out by Niko Baric]
5547 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
5548 (sent using the client's version number) if client_version is
5549 smaller than the protocol version in use. Also change
5550 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
5551 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
5552 the client will at least see that alert.
5555 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
5559 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
5560 client receives HelloRequest while in a handshake.
5561 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
5563 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
5564 should end in 'break', not 'goto end' which circuments various
5565 cleanups done in state SSL_ST_OK. But session related stuff
5566 must be disabled for SSL_ST_OK in the case that we just sent a
5569 Also avoid some overhead by not calling ssl_init_wbio_buffer()
5570 before just sending a HelloRequest.
5571 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
5573 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
5574 reveal whether illegal block cipher padding was found or a MAC
5575 verification error occured. (Neither SSLerr() codes nor alerts
5576 are directly visible to potential attackers, but the information
5577 may leak via logfiles.)
5579 Similar changes are not required for the SSL 2.0 implementation
5580 because the number of padding bytes is sent in clear for SSL 2.0,
5581 and the extra bytes are just ignored. However ssl/s2_pkt.c
5582 failed to verify that the purported number of padding bytes is in
5586 *) Add OpenUNIX-8 support including shared libraries
5587 (Boyd Lynn Gerber <gerberb@zenez.com>).
5590 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
5591 'wristwatch attack' using huge encoding parameters (cf.
5592 James H. Manger's CRYPTO 2001 paper). Note that the
5593 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
5594 encoding parameters and hence was not vulnerable.
5597 *) BN_sqr() bug fix.
5598 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
5600 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
5601 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
5602 followed by modular reduction.
5603 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
5605 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
5606 equivalent based on BN_pseudo_rand() instead of BN_rand().
5609 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
5610 This function was broken, as the check for a new client hello message
5611 to handle SGC did not allow these large messages.
5612 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
5615 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
5618 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
5619 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
5622 *) Rework the configuration and shared library support for Tru64 Unix.
5623 The configuration part makes use of modern compiler features and
5624 still retains old compiler behavior for those that run older versions
5625 of the OS. The shared library support part includes a variant that
5626 uses the RPATH feature, and is available through the special
5627 configuration target "alpha-cc-rpath", which will never be selected
5629 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
5631 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
5632 with the same message size as in ssl3_get_certificate_request().
5633 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
5634 messages might inadvertently be reject as too long.
5635 [Petr Lampa <lampa@fee.vutbr.cz>]
5637 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
5640 *) Modified SSL library such that the verify_callback that has been set
5641 specificly for an SSL object with SSL_set_verify() is actually being
5642 used. Before the change, a verify_callback set with this function was
5643 ignored and the verify_callback() set in the SSL_CTX at the time of
5644 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
5645 to allow the necessary settings.
5648 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
5649 explicitly to NULL, as at least on Solaris 8 this seems not always to be
5650 done automatically (in contradiction to the requirements of the C
5651 standard). This made problems when used from OpenSSH.
5654 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
5655 dh->length and always used
5657 BN_rand_range(priv_key, dh->p).
5659 BN_rand_range() is not necessary for Diffie-Hellman, and this
5660 specific range makes Diffie-Hellman unnecessarily inefficient if
5661 dh->length (recommended exponent length) is much smaller than the
5662 length of dh->p. We could use BN_rand_range() if the order of
5663 the subgroup was stored in the DH structure, but we only have
5668 BN_rand(priv_key, l, ...)
5670 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
5676 RSA_eay_public_encrypt
5677 RSA_eay_private_decrypt
5678 RSA_eay_private_encrypt (signing)
5679 RSA_eay_public_decrypt (signature verification)
5681 (default implementations for RSA_public_encrypt,
5682 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
5683 always reject numbers >= n.
5686 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
5687 to synchronize access to 'locking_thread'. This is necessary on
5688 systems where access to 'locking_thread' (an 'unsigned long'
5689 variable) is not atomic.
5692 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
5693 *before* setting the 'crypto_lock_rand' flag. The previous code had
5694 a race condition if 0 is a valid thread ID.
5695 [Travis Vitek <vitek@roguewave.com>]
5697 *) Add support for shared libraries under Irix.
5698 [Albert Chin-A-Young <china@thewrittenword.com>]
5700 *) Add configuration option to build on Linux on both big-endian and
5702 [Ralf Baechle <ralf@uni-koblenz.de>]
5704 *) Add the possibility to create shared libraries on HP-UX.
5707 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
5709 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
5710 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
5711 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
5712 PRNG state recovery was possible based on the output of
5713 one PRNG request appropriately sized to gain knowledge on
5714 'md' followed by enough consecutive 1-byte PRNG requests
5715 to traverse all of 'state'.
5717 1. When updating 'md_local' (the current thread's copy of 'md')
5718 during PRNG output generation, hash all of the previous
5719 'md_local' value, not just the half used for PRNG output.
5721 2. Make the number of bytes from 'state' included into the hash
5722 independent from the number of PRNG bytes requested.
5724 The first measure alone would be sufficient to avoid
5725 Markku-Juhani's attack. (Actually it had never occurred
5726 to me that the half of 'md_local' used for chaining was the
5727 half from which PRNG output bytes were taken -- I had always
5728 assumed that the secret half would be used.) The second
5729 measure makes sure that additional data from 'state' is never
5730 mixed into 'md_local' in small portions; this heuristically
5731 further strengthens the PRNG.
5734 *) Fix crypto/bn/asm/mips3.s.
5737 *) When only the key is given to "enc", the IV is undefined. Print out
5738 an error message in this case.
5741 *) Handle special case when X509_NAME is empty in X509 printing routines.
5744 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
5745 positive and less than q.
5748 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
5749 used: it isn't thread safe and the add_lock_callback should handle
5751 [Paul Rose <Paul.Rose@bridge.com>]
5753 *) Verify that incoming data obeys the block size in
5754 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
5758 [Ulf Möller, Bodo Möller]
5760 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
5761 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
5762 when fixing the server behaviour for backwards-compatible 'client
5763 hello' messages. (Note that the attack is impractical against
5764 SSL 3.0 and TLS 1.0 anyway because length and version checking
5765 means that the probability of guessing a valid ciphertext is
5766 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
5769 Before 0.9.5, the countermeasure (hide the error by generating a
5770 random 'decryption result') did not work properly because
5771 ERR_clear_error() was missing, meaning that SSL_get_error() would
5772 detect the supposedly ignored error.
5774 Both problems are now fixed.
5777 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
5778 (previously it was 1024).
5781 *) Fix for compatibility mode trust settings: ignore trust settings
5782 unless some valid trust or reject settings are present.
5785 *) Fix for blowfish EVP: its a variable length cipher.
5788 *) Fix various bugs related to DSA S/MIME verification. Handle missing
5789 parameters in DSA public key structures and return an error in the
5790 DSA routines if parameters are absent.
5793 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
5794 in the current directory if neither $RANDFILE nor $HOME was set.
5795 RAND_file_name() in 0.9.6a returned NULL in this case. This has
5796 caused some confusion to Windows users who haven't defined $HOME.
5797 Thus RAND_file_name() is changed again: e_os.h can define a
5798 DEFAULT_HOME, which will be used if $HOME is not set.
5799 For Windows, we use "C:"; on other platforms, we still require
5800 environment variables.
5802 *) Move 'if (!initialized) RAND_poll()' into regions protected by
5803 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
5804 having multiple threads call RAND_poll() concurrently.
5807 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
5808 combination of a flag and a thread ID variable.
5809 Otherwise while one thread is in ssleay_rand_bytes (which sets the
5810 flag), *other* threads can enter ssleay_add_bytes without obeying
5811 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
5812 that they do not hold after the first thread unsets add_do_not_lock).
5815 *) Change bctest again: '-x' expressions are not available in all
5819 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
5821 *) Fix a couple of memory leaks in PKCS7_dataDecode()
5822 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
5824 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
5825 the default extension for executables, if any. Also, make the perl
5826 scripts that use symlink() to test if it really exists and use "cp"
5827 if it doesn't. All this made OpenSSL compilable and installable in
5831 *) Fix for asn1_GetSequence() for indefinite length constructed data.
5832 If SEQUENCE is length is indefinite just set c->slen to the total
5833 amount of data available.
5834 [Steve Henson, reported by shige@FreeBSD.org]
5835 [This change does not apply to 0.9.7.]
5837 *) Change bctest to avoid here-documents inside command substitution
5838 (workaround for FreeBSD /bin/sh bug).
5839 For compatibility with Ultrix, avoid shell functions (introduced
5840 in the bctest version that searches along $PATH).
5843 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
5844 with des_encrypt() defined on some operating systems, like Solaris
5848 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
5849 On the Importance of Eliminating Errors in Cryptographic
5850 Computations, J. Cryptology 14 (2001) 2, 101-119,
5851 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
5854 *) MIPS assembler BIGNUM division bug fix.
5857 *) Disabled incorrect Alpha assembler code.
5860 *) Fix PKCS#7 decode routines so they correctly update the length
5861 after reading an EOC for the EXPLICIT tag.
5863 [This change does not apply to 0.9.7.]
5865 *) Fix bug in PKCS#12 key generation routines. This was triggered
5866 if a 3DES key was generated with a 0 initial byte. Include
5867 PKCS12_BROKEN_KEYGEN compilation option to retain the old
5868 (but broken) behaviour.
5871 *) Enhance bctest to search for a working bc along $PATH and print
5873 [Tim Rice <tim@multitalents.net> via Richard Levitte]
5875 *) Fix memory leaks in err.c: free err_data string if necessary;
5876 don't write to the wrong index in ERR_set_error_data.
5879 *) Implement ssl23_peek (analogous to ssl23_read), which previously
5883 *) Replace rdtsc with _emit statements for VC++ version 5.
5884 [Jeremy Cooper <jeremy@baymoo.org>]
5886 *) Make it possible to reuse SSLv2 sessions.
5889 *) In copy_email() check for >= 0 as a return value for
5890 X509_NAME_get_index_by_NID() since 0 is a valid index.
5891 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
5893 *) Avoid coredump with unsupported or invalid public keys by checking if
5894 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
5895 PKCS7_verify() fails with non detached data.
5898 *) Don't use getenv in library functions when run as setuid/setgid.
5899 New function OPENSSL_issetugid().
5902 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
5903 due to incorrect handling of multi-threading:
5905 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
5907 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
5909 3. Count how many times MemCheck_off() has been called so that
5910 nested use can be treated correctly. This also avoids
5911 inband-signalling in the previous code (which relied on the
5912 assumption that thread ID 0 is impossible).
5915 *) Add "-rand" option also to s_client and s_server.
5918 *) Fix CPU detection on Irix 6.x.
5919 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
5920 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
5922 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
5925 [This change does not apply to 0.9.7.]
5927 *) Use the cached encoding of an X509_NAME structure rather than
5928 copying it. This is apparently the reason for the libsafe "errors"
5929 but the code is actually correct.
5932 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
5933 Bleichenbacher's DSA attack.
5934 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
5935 to be set and top=0 forces the highest bit to be set; top=-1 is new
5936 and leaves the highest bit random.
5937 [Ulf Moeller, Bodo Moeller]
5939 *) In the NCONF_...-based implementations for CONF_... queries
5940 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
5941 a temporary CONF structure with the data component set to NULL
5942 (which gives segmentation faults in lh_retrieve).
5943 Instead, use NULL for the CONF pointer in CONF_get_string and
5944 CONF_get_number (which may use environment variables) and directly
5945 return NULL from CONF_get_section.
5948 *) Fix potential buffer overrun for EBCDIC.
5951 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
5952 keyUsage if basicConstraints absent for a CA.
5955 *) Make SMIME_write_PKCS7() write mail header values with a format that
5956 is more generally accepted (no spaces before the semicolon), since
5957 some programs can't parse those values properly otherwise. Also make
5958 sure BIO's that break lines after each write do not create invalid
5962 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
5963 macros previously used would not encode an empty SEQUENCE OF
5964 and break the signature.
5966 [This change does not apply to 0.9.7.]
5968 *) Zero the premaster secret after deriving the master secret in
5972 *) Add some EVP_add_digest_alias registrations (as found in
5973 OpenSSL_add_all_digests()) to SSL_library_init()
5974 aka OpenSSL_add_ssl_algorithms(). This provides improved
5975 compatibility with peers using X.509 certificates
5976 with unconventional AlgorithmIdentifier OIDs.
5979 *) Fix for Irix with NO_ASM.
5980 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
5982 *) ./config script fixes.
5983 [Ulf Moeller, Richard Levitte]
5985 *) Fix 'openssl passwd -1'.
5988 *) Change PKCS12_key_gen_asc() so it can cope with non null
5989 terminated strings whose length is passed in the passlen
5990 parameter, for example from PEM callbacks. This was done
5991 by adding an extra length parameter to asc2uni().
5992 [Steve Henson, reported by <oddissey@samsung.co.kr>]
5994 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
5995 call failed, free the DSA structure.
5998 *) Fix to uni2asc() to cope with zero length Unicode strings.
5999 These are present in some PKCS#12 files.
6002 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6003 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6004 when writing a 32767 byte record.
6005 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6007 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6008 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6010 (RSA objects have a reference count access to which is protected
6011 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6012 so they are meant to be shared between threads.)
6013 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6014 "Reddie, Steven" <Steven.Reddie@ca.com>]
6016 *) Fix a deadlock in CRYPTO_mem_leaks().
6019 *) Use better test patterns in bntest.
6022 *) rand_win.c fix for Borland C.
6025 *) BN_rshift bugfix for n == 0.
6028 *) Add a 'bctest' script that checks for some known 'bc' bugs
6029 so that 'make test' does not abort just because 'bc' is broken.
6032 *) Store verify_result within SSL_SESSION also for client side to
6033 avoid potential security hole. (Re-used sessions on the client side
6034 always resulted in verify_result==X509_V_OK, not using the original
6035 result of the server certificate verification.)
6038 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6039 SSL3_RT_APPLICATION_DATA, return 0.
6040 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6044 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6045 releases, have been re-implemented by renaming the previous
6046 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6047 and ssl3_read_internal, respectively, and adding 'peek' parameters
6048 to them. The new ssl[23]_{read,peek} functions are calls to
6049 ssl[23]_read_internal with the 'peek' flag set appropriately.
6050 A 'peek' parameter has also been added to ssl3_read_bytes, which
6051 does the actual work for ssl3_read_internal.
6054 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6055 the method-specific "init()" handler. Also clean up ex_data after
6056 calling the method-specific "finish()" handler. Previously, this was
6057 happening the other way round.
6060 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6061 The previous value, 12, was not always sufficient for BN_mod_exp().
6064 *) Make sure that shared libraries get the internal name engine with
6065 the full version number and not just 0. This should mark the
6066 shared libraries as not backward compatible. Of course, this should
6067 be changed again when we can guarantee backward binary compatibility.
6070 *) Fix typo in get_cert_by_subject() in by_dir.c
6071 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6073 *) Rework the system to generate shared libraries:
6075 - Make note of the expected extension for the shared libraries and
6076 if there is a need for symbolic links from for example libcrypto.so.0
6077 to libcrypto.so.0.9.7. There is extended info in Configure for
6080 - Make as few rebuilds of the shared libraries as possible.
6082 - Still avoid linking the OpenSSL programs with the shared libraries.
6084 - When installing, install the shared libraries separately from the
6088 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6090 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6091 and not in SSL_clear because the latter is also used by the
6092 accept/connect functions; previously, the settings made by
6093 SSL_set_read_ahead would be lost during the handshake.
6094 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6096 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6097 Previously, it would create entries for disableed algorithms no
6101 *) Added several new manual pages for SSL_* function.
6104 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6106 *) In ssl23_get_client_hello, generate an error message when faced
6107 with an initial SSL 3.0/TLS record that is too small to contain the
6108 first two bytes of the ClientHello message, i.e. client_version.
6109 (Note that this is a pathologic case that probably has never happened
6110 in real life.) The previous approach was to use the version number
6111 from the record header as a substitute; but our protocol choice
6112 should not depend on that one because it is not authenticated
6113 by the Finished messages.
6116 *) More robust randomness gathering functions for Windows.
6117 [Jeffrey Altman <jaltman@columbia.edu>]
6119 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
6120 not set then we don't setup the error code for issuer check errors
6121 to avoid possibly overwriting other errors which the callback does
6122 handle. If an application does set the flag then we assume it knows
6123 what it is doing and can handle the new informational codes
6127 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
6128 a general "ANY" type, as such it should be able to decode anything
6129 including tagged types. However it didn't check the class so it would
6130 wrongly interpret tagged types in the same way as their universal
6131 counterpart and unknown types were just rejected. Changed so that the
6132 tagged and unknown types are handled in the same way as a SEQUENCE:
6133 that is the encoding is stored intact. There is also a new type
6134 "V_ASN1_OTHER" which is used when the class is not universal, in this
6135 case we have no idea what the actual type is so we just lump them all
6139 *) On VMS, stdout may very well lead to a file that is written to
6140 in a record-oriented fashion. That means that every write() will
6141 write a separate record, which will be read separately by the
6142 programs trying to read from it. This can be very confusing.
6144 The solution is to put a BIO filter in the way that will buffer
6145 text until a linefeed is reached, and then write everything a
6146 line at a time, so every record written will be an actual line,
6147 not chunks of lines and not (usually doesn't happen, but I've
6148 seen it once) several lines in one record. BIO_f_linebuffer() is
6151 Currently, it's a VMS-only method, because that's where it has
6152 been tested well enough.
6155 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
6156 it can return incorrect results.
6157 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
6158 but it was in 0.9.6-beta[12].)
6161 *) Disable the check for content being present when verifying detached
6162 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
6163 include zero length content when signing messages.
6166 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
6167 BIO_ctrl (for BIO pairs).
6170 *) Add DSO method for VMS.
6173 *) Bug fix: Montgomery multiplication could produce results with the
6177 *) Add RPM specification openssl.spec and modify it to build three
6178 packages. The default package contains applications, application
6179 documentation and run-time libraries. The devel package contains
6180 include files, static libraries and function documentation. The
6181 doc package contains the contents of the doc directory. The original
6182 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
6185 *) Add a large number of documentation files for many SSL routines.
6186 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6188 *) Add a configuration entry for Sony News 4.
6189 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
6191 *) Don't set the two most significant bits to one when generating a
6192 random number < q in the DSA library.
6195 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
6196 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
6197 the underlying transport is blocking) if a handshake took place.
6198 (The default behaviour is needed by applications such as s_client
6199 and s_server that use select() to determine when to use SSL_read;
6200 but for applications that know in advance when to expect data, it
6201 just makes things more complicated.)
6204 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
6208 *) Add a few more EBCDIC conditionals that make `req' and `x509'
6209 work better on such systems.
6210 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6212 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
6213 Update PKCS12_parse() so it copies the friendlyName and the
6214 keyid to the certificates aux info.
6217 *) Fix bug in PKCS7_verify() which caused an infinite loop
6218 if there was more than one signature.
6219 [Sven Uszpelkat <su@celocom.de>]
6221 *) Major change in util/mkdef.pl to include extra information
6222 about each symbol, as well as presentig variables as well
6223 as functions. This change means that there's n more need
6224 to rebuild the .num files when some algorithms are excluded.
6227 *) Allow the verify time to be set by an application,
6228 rather than always using the current time.
6231 *) Phase 2 verify code reorganisation. The certificate
6232 verify code now looks up an issuer certificate by a
6233 number of criteria: subject name, authority key id
6234 and key usage. It also verifies self signed certificates
6235 by the same criteria. The main comparison function is
6236 X509_check_issued() which performs these checks.
6238 Lot of changes were necessary in order to support this
6239 without completely rewriting the lookup code.
6241 Authority and subject key identifier are now cached.
6243 The LHASH 'certs' is X509_STORE has now been replaced
6244 by a STACK_OF(X509_OBJECT). This is mainly because an
6245 LHASH can't store or retrieve multiple objects with
6246 the same hash value.
6248 As a result various functions (which were all internal
6249 use only) have changed to handle the new X509_STORE
6250 structure. This will break anything that messed round
6251 with X509_STORE internally.
6253 The functions X509_STORE_add_cert() now checks for an
6254 exact match, rather than just subject name.
6256 The X509_STORE API doesn't directly support the retrieval
6257 of multiple certificates matching a given criteria, however
6258 this can be worked round by performing a lookup first
6259 (which will fill the cache with candidate certificates)
6260 and then examining the cache for matches. This is probably
6261 the best we can do without throwing out X509_LOOKUP
6262 entirely (maybe later...).
6264 The X509_VERIFY_CTX structure has been enhanced considerably.
6266 All certificate lookup operations now go via a get_issuer()
6267 callback. Although this currently uses an X509_STORE it
6268 can be replaced by custom lookups. This is a simple way
6269 to bypass the X509_STORE hackery necessary to make this
6270 work and makes it possible to use more efficient techniques
6271 in future. A very simple version which uses a simple
6272 STACK for its trusted certificate store is also provided
6273 using X509_STORE_CTX_trusted_stack().
6275 The verify_cb() and verify() callbacks now have equivalents
6276 in the X509_STORE_CTX structure.
6278 X509_STORE_CTX also has a 'flags' field which can be used
6279 to customise the verify behaviour.
6282 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
6283 excludes S/MIME capabilities.
6286 *) When a certificate request is read in keep a copy of the
6287 original encoding of the signed data and use it when outputing
6288 again. Signatures then use the original encoding rather than
6289 a decoded, encoded version which may cause problems if the
6290 request is improperly encoded.
6293 *) For consistency with other BIO_puts implementations, call
6294 buffer_write(b, ...) directly in buffer_puts instead of calling
6297 In BIO_puts, increment b->num_write as in BIO_write.
6298 [Peter.Sylvester@EdelWeb.fr]
6300 *) Fix BN_mul_word for the case where the word is 0. (We have to use
6301 BN_zero, we may not return a BIGNUM with an array consisting of
6305 *) Avoid calling abort() from within the library when problems are
6306 detected, except if preprocessor symbols have been defined
6307 (such as REF_CHECK, BN_DEBUG etc.).
6310 *) New openssl application 'rsautl'. This utility can be
6311 used for low level RSA operations. DER public key
6312 BIO/fp routines also added.
6315 *) New Configure entry and patches for compiling on QNX 4.
6316 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
6318 *) A demo state-machine implementation was sponsored by
6319 Nuron (http://www.nuron.com/) and is now available in
6320 demos/state_machine.
6323 *) New options added to the 'dgst' utility for signature
6324 generation and verification.
6327 *) Unrecognized PKCS#7 content types are now handled via a
6328 catch all ASN1_TYPE structure. This allows unsupported
6329 types to be stored as a "blob" and an application can
6330 encode and decode it manually.
6333 *) Fix various signed/unsigned issues to make a_strex.c
6335 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
6337 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
6338 length if passed a buffer. ASN1_INTEGER_to_BN failed
6339 if passed a NULL BN and its argument was negative.
6340 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
6342 *) Modification to PKCS#7 encoding routines to output definite
6343 length encoding. Since currently the whole structures are in
6344 memory there's not real point in using indefinite length
6345 constructed encoding. However if OpenSSL is compiled with
6346 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
6349 *) Added BIO_vprintf() and BIO_vsnprintf().
6352 *) Added more prefixes to parse for in the the strings written
6353 through a logging bio, to cover all the levels that are available
6354 through syslog. The prefixes are now:
6356 PANIC, EMERG, EMR => LOG_EMERG
6357 ALERT, ALR => LOG_ALERT
6358 CRIT, CRI => LOG_CRIT
6359 ERROR, ERR => LOG_ERR
6360 WARNING, WARN, WAR => LOG_WARNING
6361 NOTICE, NOTE, NOT => LOG_NOTICE
6362 INFO, INF => LOG_INFO
6363 DEBUG, DBG => LOG_DEBUG
6365 and as before, if none of those prefixes are present at the
6366 beginning of the string, LOG_ERR is chosen.
6368 On Win32, the LOG_* levels are mapped according to this:
6370 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
6371 LOG_WARNING => EVENTLOG_WARNING_TYPE
6372 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
6376 *) Made it possible to reconfigure with just the configuration
6377 argument "reconf" or "reconfigure". The command line arguments
6378 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
6379 and are retrieved from there when reconfiguring.
6383 [Assar Westerlund <assar@sics.se>, Richard Levitte]
6385 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
6388 *) The obj_dat.pl script was messing up the sorting of object
6389 names. The reason was that it compared the quoted version
6390 of strings as a result "OCSP" > "OCSP Signing" because
6391 " > SPACE. Changed script to store unquoted versions of
6392 names and add quotes on output. It was also omitting some
6393 names from the lookup table if they were given a default
6394 value (that is if SN is missing it is given the same
6395 value as LN and vice versa), these are now added on the
6396 grounds that if an object has a name we should be able to
6397 look it up. Finally added warning output when duplicate
6398 short or long names are found.
6401 *) Changes needed for Tandem NSK.
6402 [Scott Uroff <scott@xypro.com>]
6404 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
6405 RSA_padding_check_SSLv23(), special padding was never detected
6406 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
6407 version rollback attacks was not effective.
6409 In s23_clnt.c, don't use special rollback-attack detection padding
6410 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
6411 client; similarly, in s23_srvr.c, don't do the rollback check if
6412 SSL 2.0 is the only protocol enabled in the server.
6415 *) Make it possible to get hexdumps of unprintable data with 'openssl
6416 asn1parse'. By implication, the functions ASN1_parse_dump() and
6417 BIO_dump_indent() are added.
6420 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
6421 these print out strings and name structures based on various
6422 flags including RFC2253 support and proper handling of
6423 multibyte characters. Added options to the 'x509' utility
6424 to allow the various flags to be set.
6427 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
6428 Also change the functions X509_cmp_current_time() and
6429 X509_gmtime_adj() work with an ASN1_TIME structure,
6430 this will enable certificates using GeneralizedTime in validity
6431 dates to be checked.
6434 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
6435 negative public key encodings) on by default,
6436 NO_NEG_PUBKEY_BUG can be set to disable it.
6439 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
6440 content octets. An i2c_ASN1_OBJECT is unnecessary because
6441 the encoding can be trivially obtained from the structure.
6444 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
6445 not read locks (CRYPTO_r_[un]lock).
6448 *) A first attempt at creating official support for shared
6449 libraries through configuration. I've kept it so the
6450 default is static libraries only, and the OpenSSL programs
6451 are always statically linked for now, but there are
6452 preparations for dynamic linking in place.
6453 This has been tested on Linux and Tru64.
6456 *) Randomness polling function for Win9x, as described in:
6457 Peter Gutmann, Software Generation of Practically Strong
6461 *) Fix so PRNG is seeded in req if using an already existing
6465 *) New options to smime application. -inform and -outform
6466 allow alternative formats for the S/MIME message including
6467 PEM and DER. The -content option allows the content to be
6468 specified separately. This should allow things like Netscape
6469 form signing output easier to verify.
6472 *) Fix the ASN1 encoding of tags using the 'long form'.
6475 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
6476 STRING types. These convert content octets to and from the
6477 underlying type. The actual tag and length octets are
6478 already assumed to have been read in and checked. These
6479 are needed because all other string types have virtually
6480 identical handling apart from the tag. By having versions
6481 of the ASN1 functions that just operate on content octets
6482 IMPLICIT tagging can be handled properly. It also allows
6483 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
6484 and ASN1_INTEGER are identical apart from the tag.
6487 *) Change the handling of OID objects as follows:
6489 - New object identifiers are inserted in objects.txt, following
6490 the syntax given in objects.README.
6491 - objects.pl is used to process obj_mac.num and create a new
6493 - obj_dat.pl is used to create a new obj_dat.h, using the data in
6496 This is currently kind of a hack, and the perl code in objects.pl
6497 isn't very elegant, but it works as I intended. The simplest way
6498 to check that it worked correctly is to look in obj_dat.h and
6499 check the array nid_objs and make sure the objects haven't moved
6500 around (this is important!). Additions are OK, as well as
6501 consistent name changes.
6504 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
6507 *) Addition of the command line parameter '-rand file' to 'openssl req'.
6508 The given file adds to whatever has already been seeded into the
6509 random pool through the RANDFILE configuration file option or
6510 environment variable, or the default random state file.
6513 *) mkstack.pl now sorts each macro group into lexical order.
6514 Previously the output order depended on the order the files
6515 appeared in the directory, resulting in needless rewriting
6519 *) Patches to make OpenSSL compile under Win32 again. Mostly
6520 work arounds for the VC++ problem that it treats func() as
6521 func(void). Also stripped out the parts of mkdef.pl that
6522 added extra typesafe functions: these no longer exist.
6525 *) Reorganisation of the stack code. The macros are now all
6526 collected in safestack.h . Each macro is defined in terms of
6527 a "stack macro" of the form SKM_<name>(type, a, b). The
6528 DEBUG_SAFESTACK is now handled in terms of function casts,
6529 this has the advantage of retaining type safety without the
6530 use of additional functions. If DEBUG_SAFESTACK is not defined
6531 then the non typesafe macros are used instead. Also modified the
6532 mkstack.pl script to handle the new form. Needs testing to see
6533 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
6534 the default if no major problems. Similar behaviour for ASN1_SET_OF
6535 and PKCS12_STACK_OF.
6538 *) When some versions of IIS use the 'NET' form of private key the
6539 key derivation algorithm is different. Normally MD5(password) is
6540 used as a 128 bit RC4 key. In the modified case
6541 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
6542 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
6543 as the old Netscape_RSA functions except they have an additional
6544 'sgckey' parameter which uses the modified algorithm. Also added
6545 an -sgckey command line option to the rsa utility. Thanks to
6546 Adrian Peck <bertie@ncipher.com> for posting details of the modified
6547 algorithm to openssl-dev.
6550 *) The evp_local.h macros were using 'c.##kname' which resulted in
6551 invalid expansion on some systems (SCO 5.0.5 for example).
6552 Corrected to 'c.kname'.
6553 [Phillip Porch <root@theporch.com>]
6555 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
6556 a STACK of email addresses from a certificate or request, these look
6557 in the subject name and the subject alternative name extensions and
6558 omit any duplicate addresses.
6561 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
6562 This makes DSA verification about 2 % faster.
6565 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
6566 (meaning that now 2^5 values will be precomputed, which is only 4 KB
6567 plus overhead for 1024 bit moduli).
6568 This makes exponentiations about 0.5 % faster for 1024 bit
6569 exponents (as measured by "openssl speed rsa2048").
6572 *) Rename memory handling macros to avoid conflicts with other
6574 Malloc => OPENSSL_malloc
6575 Malloc_locked => OPENSSL_malloc_locked
6576 Realloc => OPENSSL_realloc
6577 Free => OPENSSL_free
6580 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
6581 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
6584 *) CygWin32 support.
6585 [John Jarvie <jjarvie@newsguy.com>]
6587 *) The type-safe stack code has been rejigged. It is now only compiled
6588 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
6589 by default all type-specific stack functions are "#define"d back to
6590 standard stack functions. This results in more streamlined output
6591 but retains the type-safety checking possibilities of the original
6595 *) The STACK code has been cleaned up, and certain type declarations
6596 that didn't make a lot of sense have been brought in line. This has
6597 also involved a cleanup of sorts in safestack.h to more correctly
6598 map type-safe stack functions onto their plain stack counterparts.
6599 This work has also resulted in a variety of "const"ifications of
6600 lots of the code, especially "_cmp" operations which should normally
6601 be prototyped with "const" parameters anyway.
6604 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
6605 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
6606 (The PRNG state consists of two parts, the large pool 'state' and 'md',
6607 where all of 'md' is used each time the PRNG is used, but 'state'
6608 is used only indexed by a cyclic counter. As entropy may not be
6609 well distributed from the beginning, 'md' is important as a
6610 chaining variable. However, the output function chains only half
6611 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
6612 all of 'md', and seeding with STATE_SIZE dummy bytes will result
6613 in all of 'state' being rewritten, with the new values depending
6614 on virtually all of 'md'. This overcomes the 80 bit limitation.)
6617 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
6618 the handshake is continued after ssl_verify_cert_chain();
6619 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
6620 can lead to 'unexplainable' connection aborts later.
6621 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
6623 *) Major EVP API cipher revision.
6624 Add hooks for extra EVP features. This allows various cipher
6625 parameters to be set in the EVP interface. Support added for variable
6626 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
6627 setting of RC2 and RC5 parameters.
6629 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
6632 Remove lots of duplicated code from the EVP library. For example *every*
6633 cipher init() function handles the 'iv' in the same way according to the
6634 cipher mode. They also all do nothing if the 'key' parameter is NULL and
6635 for CFB and OFB modes they zero ctx->num.
6637 New functionality allows removal of S/MIME code RC2 hack.
6639 Most of the routines have the same form and so can be declared in terms
6642 By shifting this to the top level EVP_CipherInit() it can be removed from
6643 all individual ciphers. If the cipher wants to handle IVs or keys
6644 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
6647 Change lots of functions like EVP_EncryptUpdate() to now return a
6648 value: although software versions of the algorithms cannot fail
6649 any installed hardware versions can.
6652 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
6653 this option is set, tolerate broken clients that send the negotiated
6654 protocol version number instead of the requested protocol version
6658 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
6659 i.e. non-zero for export ciphersuites, zero otherwise.
6660 Previous versions had this flag inverted, inconsistent with
6661 rsa_tmp_cb (..._TMP_RSA_CB).
6662 [Bodo Moeller; problem reported by Amit Chopra]
6664 *) Add missing DSA library text string. Work around for some IIS
6665 key files with invalid SEQUENCE encoding.
6668 *) Add a document (doc/standards.txt) that list all kinds of standards
6669 and so on that are implemented in OpenSSL.
6672 *) Enhance c_rehash script. Old version would mishandle certificates
6673 with the same subject name hash and wouldn't handle CRLs at all.
6674 Added -fingerprint option to crl utility, to support new c_rehash
6678 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
6681 *) Fix for SSL server purpose checking. Server checking was
6682 rejecting certificates which had extended key usage present
6683 but no ssl client purpose.
6684 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
6686 *) Make PKCS#12 code work with no password. The PKCS#12 spec
6687 is a little unclear about how a blank password is handled.
6688 Since the password in encoded as a BMPString with terminating
6689 double NULL a zero length password would end up as just the
6690 double NULL. However no password at all is different and is
6691 handled differently in the PKCS#12 key generation code. NS
6692 treats a blank password as zero length. MSIE treats it as no
6693 password on export: but it will try both on import. We now do
6694 the same: PKCS12_parse() tries zero length and no password if
6695 the password is set to "" or NULL (NULL is now a valid password:
6696 it wasn't before) as does the pkcs12 application.
6699 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
6700 perror when PEM_read_bio_X509_REQ fails, the error message must
6701 be obtained from the error queue.
6704 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
6705 it in ERR_remove_state if appropriate, and change ERR_get_state
6706 accordingly to avoid race conditions (this is necessary because
6707 thread_hash is no longer constant once set).
6710 *) Bugfix for linux-elf makefile.one.
6713 *) RSA_get_default_method() will now cause a default
6714 RSA_METHOD to be chosen if one doesn't exist already.
6715 Previously this was only set during a call to RSA_new()
6716 or RSA_new_method(NULL) meaning it was possible for
6717 RSA_get_default_method() to return NULL.
6720 *) Added native name translation to the existing DSO code
6721 that will convert (if the flag to do so is set) filenames
6722 that are sufficiently small and have no path information
6723 into a canonical native form. Eg. "blah" converted to
6724 "libblah.so" or "blah.dll" etc.
6727 *) New function ERR_error_string_n(e, buf, len) which is like
6728 ERR_error_string(e, buf), but writes at most 'len' bytes
6729 including the 0 terminator. For ERR_error_string_n, 'buf'
6731 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
6733 *) CONF library reworked to become more general. A new CONF
6734 configuration file reader "class" is implemented as well as a
6735 new functions (NCONF_*, for "New CONF") to handle it. The now
6736 old CONF_* functions are still there, but are reimplemented to
6737 work in terms of the new functions. Also, a set of functions
6738 to handle the internal storage of the configuration data is
6739 provided to make it easier to write new configuration file
6740 reader "classes" (I can definitely see something reading a
6741 configuration file in XML format, for example), called _CONF_*,
6742 or "the configuration storage API"...
6744 The new configuration file reading functions are:
6746 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
6747 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
6749 NCONF_default, NCONF_WIN32
6751 NCONF_dump_fp, NCONF_dump_bio
6753 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
6754 NCONF_new creates a new CONF object. This works in the same way
6755 as other interfaces in OpenSSL, like the BIO interface.
6756 NCONF_dump_* dump the internal storage of the configuration file,
6757 which is useful for debugging. All other functions take the same
6758 arguments as the old CONF_* functions wth the exception of the
6759 first that must be a `CONF *' instead of a `LHASH *'.
6761 To make it easer to use the new classes with the old CONF_* functions,
6762 the function CONF_set_default_method is provided.
6765 *) Add '-tls1' option to 'openssl ciphers', which was already
6766 mentioned in the documentation but had not been implemented.
6767 (This option is not yet really useful because even the additional
6768 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
6771 *) Initial DSO code added into libcrypto for letting OpenSSL (and
6772 OpenSSL-based applications) load shared libraries and bind to
6773 them in a portable way.
6774 [Geoff Thorpe, with contributions from Richard Levitte]
6776 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
6778 *) Make sure _lrotl and _lrotr are only used with MSVC.
6780 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
6781 (the default implementation of RAND_status).
6783 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
6784 to '-clrext' (= clear extensions), as intended and documented.
6785 [Bodo Moeller; inconsistency pointed out by Michael Attili
6788 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
6789 was larger than the MD block size.
6790 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
6792 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
6793 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
6794 using the passed key: if the passed key was a private key the result
6795 of X509_print(), for example, would be to print out all the private key
6799 *) des_quad_cksum() byte order bug fix.
6800 [Ulf Möller, using the problem description in krb4-0.9.7, where
6801 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
6803 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
6805 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
6807 *) For easily testing in shell scripts whether some command
6808 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
6809 returns with exit code 0 iff no command of the given name is available.
6810 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
6811 the output goes to stdout and nothing is printed to stderr.
6812 Additional arguments are always ignored.
6814 Since for each cipher there is a command of the same name,
6815 the 'no-cipher' compilation switches can be tested this way.
6817 ('openssl no-XXX' is not able to detect pseudo-commands such
6818 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
6821 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
6824 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
6825 is set; it will be thrown away anyway because each handshake creates
6827 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
6828 to parameters -- in previous versions (since OpenSSL 0.9.3) the
6829 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
6830 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
6833 *) New s_client option -ign_eof: EOF at stdin is ignored, and
6834 'Q' and 'R' lose their special meanings (quit/renegotiate).
6835 This is part of what -quiet does; unlike -quiet, -ign_eof
6836 does not suppress any output.
6839 *) Add compatibility options to the purpose and trust code. The
6840 purpose X509_PURPOSE_ANY is "any purpose" which automatically
6841 accepts a certificate or CA, this was the previous behaviour,
6842 with all the associated security issues.
6844 X509_TRUST_COMPAT is the old trust behaviour: only and
6845 automatically trust self signed roots in certificate store. A
6846 new trust setting X509_TRUST_DEFAULT is used to specify that
6847 a purpose has no associated trust setting and it should instead
6848 use the value in the default purpose.
6851 *) Fix the PKCS#8 DSA private key code so it decodes keys again
6852 and fix a memory leak.
6855 *) In util/mkerr.pl (which implements 'make errors'), preserve
6856 reason strings from the previous version of the .c file, as
6857 the default to have only downcase letters (and digits) in
6858 automatically generated reasons codes is not always appropriate.
6861 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
6862 using strerror. Previously, ERR_reason_error_string() returned
6863 library names as reason strings for SYSerr; but SYSerr is a special
6864 case where small numbers are errno values, not library numbers.
6867 *) Add '-dsaparam' option to 'openssl dhparam' application. This
6868 converts DSA parameters into DH parameters. (When creating parameters,
6869 DSA_generate_parameters is used.)
6872 *) Include 'length' (recommended exponent length) in C code generated
6873 by 'openssl dhparam -C'.
6876 *) The second argument to set_label in perlasm was already being used
6877 so couldn't be used as a "file scope" flag. Moved to third argument
6881 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
6882 instead of RAND_bytes for encryption IVs and salts.
6885 *) Include RAND_status() into RAND_METHOD instead of implementing
6886 it only for md_rand.c Otherwise replacing the PRNG by calling
6887 RAND_set_rand_method would be impossible.
6890 *) Don't let DSA_generate_key() enter an infinite loop if the random
6891 number generation fails.
6894 *) New 'rand' application for creating pseudo-random output.
6897 *) Added configuration support for Linux/IA64
6898 [Rolf Haberrecker <rolf@suse.de>]
6900 *) Assembler module support for Mingw32.
6903 *) Shared library support for HPUX (in shlib/).
6904 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
6906 *) Shared library support for Solaris gcc.
6907 [Lutz Behnke <behnke@trustcenter.de>]
6909 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
6911 *) PKCS7_encrypt() was adding text MIME headers twice because they
6912 were added manually and by SMIME_crlf_copy().
6915 *) In bntest.c don't call BN_rand with zero bits argument.
6916 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
6918 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
6919 case was implemented. This caused BN_div_recp() to fail occasionally.
6922 *) Add an optional second argument to the set_label() in the perl
6923 assembly language builder. If this argument exists and is set
6924 to 1 it signals that the assembler should use a symbol whose
6925 scope is the entire file, not just the current function. This
6926 is needed with MASM which uses the format label:: for this scope.
6927 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
6929 *) Change the ASN1 types so they are typedefs by default. Before
6930 almost all types were #define'd to ASN1_STRING which was causing
6931 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
6935 *) Change names of new functions to the new get1/get0 naming
6936 convention: After 'get1', the caller owns a reference count
6937 and has to call ..._free; 'get0' returns a pointer to some
6938 data structure without incrementing reference counters.
6939 (Some of the existing 'get' functions increment a reference
6940 counter, some don't.)
6941 Similarly, 'set1' and 'add1' functions increase reference
6942 counters or duplicate objects.
6945 *) Allow for the possibility of temp RSA key generation failure:
6946 the code used to assume it always worked and crashed on failure.
6949 *) Fix potential buffer overrun problem in BIO_printf().
6950 [Ulf Möller, using public domain code by Patrick Powell; problem
6951 pointed out by David Sacerdote <das33@cornell.edu>]
6953 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
6954 RAND_egd() and RAND_status(). In the command line application,
6955 the EGD socket can be specified like a seed file using RANDFILE
6959 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
6960 Some CAs (e.g. Verisign) distribute certificates in this form.
6963 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
6964 list to exclude them. This means that no special compilation option
6965 is needed to use anonymous DH: it just needs to be included in the
6969 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
6970 EVP_MD_type. The old functionality is available in a new macro called
6971 EVP_MD_md(). Change code that uses it and update docs.
6974 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
6975 where the 'void *' argument is replaced by a function pointer argument.
6976 Previously 'void *' was abused to point to functions, which works on
6977 many platforms, but is not correct. As these functions are usually
6978 called by macros defined in OpenSSL header files, most source code
6979 should work without changes.
6982 *) <openssl/opensslconf.h> (which is created by Configure) now contains
6983 sections with information on -D... compiler switches used for
6984 compiling the library so that applications can see them. To enable
6985 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
6986 must be defined. E.g.,
6987 #define OPENSSL_ALGORITHM_DEFINES
6988 #include <openssl/opensslconf.h>
6989 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
6990 [Richard Levitte, Ulf and Bodo Möller]
6992 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
6996 *) Change the 'other' type in certificate aux info to a STACK_OF
6997 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
6998 the required ASN1 format: arbitrary types determined by an OID.
7001 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7002 argument to 'req'. This is not because the function is newer or
7003 better than others it just uses the work 'NEW' in the certificate
7004 request header lines. Some software needs this.
7007 *) Reorganise password command line arguments: now passwords can be
7008 obtained from various sources. Delete the PEM_cb function and make
7009 it the default behaviour: i.e. if the callback is NULL and the
7010 usrdata argument is not NULL interpret it as a null terminated pass
7011 phrase. If usrdata and the callback are NULL then the pass phrase
7012 is prompted for as usual.
7015 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7016 the support is automatically enabled. The resulting binaries will
7017 autodetect the card and use it if present.
7018 [Ben Laurie and Compaq Inc.]
7020 *) Work around for Netscape hang bug. This sends certificate request
7021 and server done in one record. Since this is perfectly legal in the
7022 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7023 the bugs/SSLv3 entry for more info.
7026 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7029 *) Add -rand argument to smime and pkcs12 applications and read/write
7033 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7036 *) Add command line password options to the remaining applications.
7039 *) Bug fix for BN_div_recp() for numerators with an even number of
7043 *) More tests in bntest.c, and changed test_bn output.
7046 *) ./config recognizes MacOS X now.
7049 *) Bug fix for BN_div() when the first words of num and divsor are
7050 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7053 *) Add support for various broken PKCS#8 formats, and command line
7054 options to produce them.
7057 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7058 get temporary BIGNUMs from a BN_CTX.
7061 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7065 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7066 include a #define from the old name to the new. The original intent
7067 was that statically linked binaries could for example just call
7068 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7069 link with digests. This never worked becayse SSLeay_add_all_digests()
7070 and SSLeay_add_all_ciphers() were in the same source file so calling
7071 one would link with the other. They are now in separate source files.
7074 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7077 *) Use a less unusual form of the Miller-Rabin primality test (it used
7078 a binary algorithm for exponentiation integrated into the Miller-Rabin
7079 loop, our standard modexp algorithms are faster).
7082 *) Support for the EBCDIC character set completed.
7083 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7085 *) Source code cleanups: use const where appropriate, eliminate casts,
7086 use void * instead of char * in lhash.
7089 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7090 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7091 this the server could overwrite ephemeral keys that the client
7095 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7096 using 50 iterations of the Rabin-Miller test.
7098 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7099 iterations of the Rabin-Miller test as required by the appendix
7100 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7101 As BN_is_prime_fasttest includes trial division, DSA parameter
7102 generation becomes much faster.
7104 This implies a change for the callback functions in DSA_is_prime
7105 and DSA_generate_parameters: The callback function is called once
7106 for each positive witness in the Rabin-Miller test, not just
7107 occasionally in the inner loop; and the parameters to the
7108 callback function now provide an iteration count for the outer
7109 loop rather than for the current invocation of the inner loop.
7110 DSA_generate_parameters additionally can call the callback
7111 function with an 'iteration count' of -1, meaning that a
7112 candidate has passed the trial division test (when q is generated
7113 from an application-provided seed, trial division is skipped).
7116 *) New function BN_is_prime_fasttest that optionally does trial
7117 division before starting the Rabin-Miller test and has
7118 an additional BN_CTX * argument (whereas BN_is_prime always
7119 has to allocate at least one BN_CTX).
7120 'callback(1, -1, cb_arg)' is called when a number has passed the
7121 trial division stage.
7124 *) Fix for bug in CRL encoding. The validity dates weren't being handled
7128 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
7131 *) New function BN_pseudo_rand().
7134 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
7135 bignum version of BN_from_montgomery() with the working code from
7136 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
7140 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
7141 made it impossible to use the same SSL_SESSION data structure in
7142 SSL2 clients in multiple threads.
7145 *) The return value of RAND_load_file() no longer counts bytes obtained
7146 by stat(). RAND_load_file(..., -1) is new and uses the complete file
7147 to seed the PRNG (previously an explicit byte count was required).
7148 [Ulf Möller, Bodo Möller]
7150 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
7151 used (char *) instead of (void *) and had casts all over the place.
7154 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
7157 *) Retain source code compatibility for BN_prime_checks macro:
7158 BN_is_prime(..., BN_prime_checks, ...) now uses
7159 BN_prime_checks_for_size to determine the appropriate number of
7160 Rabin-Miller iterations.
7163 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
7164 DH_CHECK_P_NOT_SAFE_PRIME.
7165 (Check if this is true? OpenPGP calls them "strong".)
7168 *) Merge the functionality of "dh" and "gendh" programs into a new program
7169 "dhparam". The old programs are retained for now but will handle DH keys
7170 (instead of parameters) in future.
7173 *) Make the ciphers, s_server and s_client programs check the return values
7174 when a new cipher list is set.
7177 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
7178 ciphers. Before when the 56bit ciphers were enabled the sorting was
7181 The syntax for the cipher sorting has been extended to support sorting by
7182 cipher-strength (using the strength_bits hard coded in the tables).
7183 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
7185 Fix a bug in the cipher-command parser: when supplying a cipher command
7186 string with an "undefined" symbol (neither command nor alphanumeric
7187 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
7188 an error is flagged.
7190 Due to the strength-sorting extension, the code of the
7191 ssl_create_cipher_list() function was completely rearranged. I hope that
7192 the readability was also increased :-)
7193 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7195 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
7196 for the first serial number and places 2 in the serial number file. This
7197 avoids problems when the root CA is created with serial number zero and
7198 the first user certificate has the same issuer name and serial number
7202 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
7203 the new code. Add documentation for this stuff.
7206 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
7207 X509_*() to X509at_*() on the grounds that they don't handle X509
7208 structures and behave in an analagous way to the X509v3 functions:
7209 they shouldn't be called directly but wrapper functions should be used
7212 So we also now have some wrapper functions that call the X509at functions
7213 when passed certificate requests. (TO DO: similar things can be done with
7214 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
7215 things. Some of these need some d2i or i2d and print functionality
7216 because they handle more complex structures.)
7219 *) Add missing #ifndefs that caused missing symbols when building libssl
7220 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
7221 NO_RSA in ssl/s2*.c.
7222 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
7224 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
7225 has a return value which indicates the quality of the random data
7226 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
7227 error queue. New function RAND_pseudo_bytes() generates output that is
7228 guaranteed to be unique but not unpredictable. RAND_add is like
7229 RAND_seed, but takes an extra argument for an entropy estimate
7230 (RAND_seed always assumes full entropy).
7233 *) Do more iterations of Rabin-Miller probable prime test (specifically,
7234 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
7235 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
7236 in crypto/bn/bn_prime.c for the complete table). This guarantees a
7237 false-positive rate of at most 2^-80 for random input.
7240 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
7243 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
7244 in the 0.9.5 release), this returns the chain
7245 from an X509_CTX structure with a dup of the stack and all
7246 the X509 reference counts upped: so the stack will exist
7247 after X509_CTX_cleanup() has been called. Modify pkcs12.c
7250 Also make SSL_SESSION_print() print out the verify return
7254 *) Add manpage for the pkcs12 command. Also change the default
7255 behaviour so MAC iteration counts are used unless the new
7256 -nomaciter option is used. This improves file security and
7257 only older versions of MSIE (4.0 for example) need it.
7260 *) Honor the no-xxx Configure options when creating .DEF files.
7263 *) Add PKCS#10 attributes to field table: challengePassword,
7264 unstructuredName and unstructuredAddress. These are taken from
7265 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
7266 international characters are used.
7268 More changes to X509_ATTRIBUTE code: allow the setting of types
7269 based on strings. Remove the 'loc' parameter when adding
7270 attributes because these will be a SET OF encoding which is sorted
7274 *) Initial changes to the 'req' utility to allow request generation
7275 automation. This will allow an application to just generate a template
7276 file containing all the field values and have req construct the
7279 Initial support for X509_ATTRIBUTE handling. Stacks of these are
7280 used all over the place including certificate requests and PKCS#7
7281 structures. They are currently handled manually where necessary with
7282 some primitive wrappers for PKCS#7. The new functions behave in a
7283 manner analogous to the X509 extension functions: they allow
7284 attributes to be looked up by NID and added.
7286 Later something similar to the X509V3 code would be desirable to
7287 automatically handle the encoding, decoding and printing of the
7288 more complex types. The string types like challengePassword can
7289 be handled by the string table functions.
7291 Also modified the multi byte string table handling. Now there is
7292 a 'global mask' which masks out certain types. The table itself
7293 can use the flag STABLE_NO_MASK to ignore the mask setting: this
7294 is useful when for example there is only one permissible type
7295 (as in countryName) and using the mask might result in no valid
7299 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
7300 SSL_get_peer_finished to allow applications to obtain the latest
7301 Finished messages sent to the peer or expected from the peer,
7302 respectively. (SSL_get_peer_finished is usually the Finished message
7303 actually received from the peer, otherwise the protocol will be aborted.)
7305 As the Finished message are message digests of the complete handshake
7306 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
7307 be used for external authentication procedures when the authentication
7308 provided by SSL/TLS is not desired or is not enough.
7311 *) Enhanced support for Alpha Linux is added. Now ./config checks if
7312 the host supports BWX extension and if Compaq C is present on the
7313 $PATH. Just exploiting of the BWX extension results in 20-30%
7314 performance kick for some algorithms, e.g. DES and RC4 to mention
7315 a couple. Compaq C in turn generates ~20% faster code for MD5 and
7319 *) Add support for MS "fast SGC". This is arguably a violation of the
7320 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
7321 weak crypto and after checking the certificate is SGC a second one
7322 with strong crypto. MS SGC stops the first handshake after receiving
7323 the server certificate message and sends a second client hello. Since
7324 a server will typically do all the time consuming operations before
7325 expecting any further messages from the client (server key exchange
7326 is the most expensive) there is little difference between the two.
7328 To get OpenSSL to support MS SGC we have to permit a second client
7329 hello message after we have sent server done. In addition we have to
7330 reset the MAC if we do get this second client hello.
7333 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
7334 if a DER encoded private key is RSA or DSA traditional format. Changed
7335 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
7336 format DER encoded private key. Newer code should use PKCS#8 format which
7337 has the key type encoded in the ASN1 structure. Added DER private key
7338 support to pkcs8 application.
7341 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
7342 ciphersuites has been selected (as required by the SSL 3/TLS 1
7343 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
7344 is set, we interpret this as a request to violate the specification
7345 (the worst that can happen is a handshake failure, and 'correct'
7346 behaviour would result in a handshake failure anyway).
7349 *) In SSL_CTX_add_session, take into account that there might be multiple
7350 SSL_SESSION structures with the same session ID (e.g. when two threads
7351 concurrently obtain them from an external cache).
7352 The internal cache can handle only one SSL_SESSION with a given ID,
7353 so if there's a conflict, we now throw out the old one to achieve
7357 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
7358 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
7359 some routines that use cipher OIDs: some ciphers do not have OIDs
7360 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
7364 *) Simplify the trust setting structure and code. Now we just have
7365 two sequences of OIDs for trusted and rejected settings. These will
7366 typically have values the same as the extended key usage extension
7367 and any application specific purposes.
7369 The trust checking code now has a default behaviour: it will just
7370 check for an object with the same NID as the passed id. Functions can
7371 be provided to override either the default behaviour or the behaviour
7372 for a given id. SSL client, server and email already have functions
7373 in place for compatibility: they check the NID and also return "trusted"
7374 if the certificate is self signed.
7377 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
7378 traditional format into an EVP_PKEY structure.
7381 *) Add a password callback function PEM_cb() which either prompts for
7382 a password if usr_data is NULL or otherwise assumes it is a null
7383 terminated password. Allow passwords to be passed on command line
7384 environment or config files in a few more utilities.
7387 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
7388 keys. Add some short names for PKCS#8 PBE algorithms and allow them
7389 to be specified on the command line for the pkcs8 and pkcs12 utilities.
7390 Update documentation.
7393 *) Support for ASN1 "NULL" type. This could be handled before by using
7394 ASN1_TYPE but there wasn't any function that would try to read a NULL
7395 and produce an error if it couldn't. For compatibility we also have
7396 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
7397 don't allocate anything because they don't need to.
7400 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
7402 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
7404 *) Rebuild of the memory allocation routines used by OpenSSL code and
7405 possibly others as well. The purpose is to make an interface that
7406 provide hooks so anyone can build a separate set of allocation and
7407 deallocation routines to be used by OpenSSL, for example memory
7408 pool implementations, or something else, which was previously hard
7409 since Malloc(), Realloc() and Free() were defined as macros having
7410 the values malloc, realloc and free, respectively (except for Win32
7411 compilations). The same is provided for memory debugging code.
7412 OpenSSL already comes with functionality to find memory leaks, but
7413 this gives people a chance to debug other memory problems.
7415 With these changes, a new set of functions and macros have appeared:
7417 CRYPTO_set_mem_debug_functions() [F]
7418 CRYPTO_get_mem_debug_functions() [F]
7419 CRYPTO_dbg_set_options() [F]
7420 CRYPTO_dbg_get_options() [F]
7421 CRYPTO_malloc_debug_init() [M]
7423 The memory debug functions are NULL by default, unless the library
7424 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
7425 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
7426 gives the standard debugging functions that come with OpenSSL) or
7427 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
7428 provided by the library user) must be used. When the standard
7429 debugging functions are used, CRYPTO_dbg_set_options can be used to
7430 request additional information:
7431 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
7432 the CRYPTO_MDEBUG_xxx macro when compiling the library.
7434 Also, things like CRYPTO_set_mem_functions will always give the
7435 expected result (the new set of functions is used for allocation
7436 and deallocation) at all times, regardless of platform and compiler
7439 To finish it up, some functions that were never use in any other
7440 way than through macros have a new API and new semantic:
7443 CRYPTO_dbg_realloc()
7446 All macros of value have retained their old syntax.
7447 [Richard Levitte and Bodo Moeller]
7449 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
7450 ordering of SMIMECapabilities wasn't in "strength order" and there
7451 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
7455 *) Some ASN1 types with illegal zero length encoding (INTEGER,
7456 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
7457 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
7459 *) Merge in my S/MIME library for OpenSSL. This provides a simple
7460 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
7461 functionality to handle multipart/signed properly) and a utility
7462 called 'smime' to call all this stuff. This is based on code I
7463 originally wrote for Celo who have kindly allowed it to be
7464 included in OpenSSL.
7467 *) Add variants des_set_key_checked and des_set_key_unchecked of
7468 des_set_key (aka des_key_sched). Global variable des_check_key
7469 decides which of these is called by des_set_key; this way
7470 des_check_key behaves as it always did, but applications and
7471 the library itself, which was buggy for des_check_key == 1,
7472 have a cleaner way to pick the version they need.
7475 *) New function PKCS12_newpass() which changes the password of a
7479 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
7480 dynamic mix. In both cases the ids can be used as an index into the
7481 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
7482 functions so they accept a list of the field values and the
7483 application doesn't need to directly manipulate the X509_TRUST
7487 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
7491 *) Modify the way the V3 extension code looks up extensions. This now
7492 works in a similar way to the object code: we have some "standard"
7493 extensions in a static table which is searched with OBJ_bsearch()
7494 and the application can add dynamic ones if needed. The file
7495 crypto/x509v3/ext_dat.h now has the info: this file needs to be
7496 updated whenever a new extension is added to the core code and kept
7497 in ext_nid order. There is a simple program 'tabtest.c' which checks
7498 this. New extensions are not added too often so this file can readily
7499 be maintained manually.
7501 There are two big advantages in doing things this way. The extensions
7502 can be looked up immediately and no longer need to be "added" using
7503 X509V3_add_standard_extensions(): this function now does nothing.
7504 [Side note: I get *lots* of email saying the extension code doesn't
7505 work because people forget to call this function]
7506 Also no dynamic allocation is done unless new extensions are added:
7507 so if we don't add custom extensions there is no need to call
7508 X509V3_EXT_cleanup().
7511 *) Modify enc utility's salting as follows: make salting the default. Add a
7512 magic header, so unsalted files fail gracefully instead of just decrypting
7513 to garbage. This is because not salting is a big security hole, so people
7514 should be discouraged from doing it.
7517 *) Fixes and enhancements to the 'x509' utility. It allowed a message
7518 digest to be passed on the command line but it only used this
7519 parameter when signing a certificate. Modified so all relevant
7520 operations are affected by the digest parameter including the
7521 -fingerprint and -x509toreq options. Also -x509toreq choked if a
7522 DSA key was used because it didn't fix the digest.
7525 *) Initial certificate chain verify code. Currently tests the untrusted
7526 certificates for consistency with the verify purpose (which is set
7527 when the X509_STORE_CTX structure is set up) and checks the pathlength.
7529 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
7530 this is because it will reject chains with invalid extensions whereas
7531 every previous version of OpenSSL and SSLeay made no checks at all.
7533 Trust code: checks the root CA for the relevant trust settings. Trust
7534 settings have an initial value consistent with the verify purpose: e.g.
7535 if the verify purpose is for SSL client use it expects the CA to be
7536 trusted for SSL client use. However the default value can be changed to
7537 permit custom trust settings: one example of this would be to only trust
7538 certificates from a specific "secure" set of CAs.
7540 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
7541 which should be used for version portability: especially since the
7542 verify structure is likely to change more often now.
7544 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
7545 to set them. If not set then assume SSL clients will verify SSL servers
7548 Two new options to the verify program: -untrusted allows a set of
7549 untrusted certificates to be passed in and -purpose which sets the
7550 intended purpose of the certificate. If a purpose is set then the
7551 new chain verify code is used to check extension consistency.
7554 *) Support for the authority information access extension.
7557 *) Modify RSA and DSA PEM read routines to transparently handle
7558 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
7559 public keys in a format compatible with certificate
7560 SubjectPublicKeyInfo structures. Unfortunately there were already
7561 functions called *_PublicKey_* which used various odd formats so
7562 these are retained for compatibility: however the DSA variants were
7563 never in a public release so they have been deleted. Changed dsa/rsa
7564 utilities to handle the new format: note no releases ever handled public
7565 keys so we should be OK.
7567 The primary motivation for this change is to avoid the same fiasco
7568 that dogs private keys: there are several incompatible private key
7569 formats some of which are standard and some OpenSSL specific and
7570 require various evil hacks to allow partial transparent handling and
7571 even then it doesn't work with DER formats. Given the option anything
7572 other than PKCS#8 should be dumped: but the other formats have to
7573 stay in the name of compatibility.
7575 With public keys and the benefit of hindsight one standard format
7576 is used which works with EVP_PKEY, RSA or DSA structures: though
7577 it clearly returns an error if you try to read the wrong kind of key.
7579 Added a -pubkey option to the 'x509' utility to output the public key.
7580 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
7581 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
7582 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
7583 that do the same as the EVP_PKEY_assign_*() except they up the
7584 reference count of the added key (they don't "swallow" the
7588 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
7589 CRLs would fail if the file contained no certificates or no CRLs:
7590 added a new function to read in both types and return the number
7591 read: this means that if none are read it will be an error. The
7592 DER versions of the certificate and CRL reader would always fail
7593 because it isn't possible to mix certificates and CRLs in DER format
7594 without choking one or the other routine. Changed this to just read
7595 a certificate: this is the best we can do. Also modified the code
7596 in apps/verify.c to take notice of return codes: it was previously
7597 attempting to read in certificates from NULL pointers and ignoring
7598 any errors: this is one reason why the cert and CRL reader seemed
7599 to work. It doesn't check return codes from the default certificate
7600 routines: these may well fail if the certificates aren't installed.
7603 *) Code to support otherName option in GeneralName.
7606 *) First update to verify code. Change the verify utility
7607 so it warns if it is passed a self signed certificate:
7608 for consistency with the normal behaviour. X509_verify
7609 has been modified to it will now verify a self signed
7610 certificate if *exactly* the same certificate appears
7611 in the store: it was previously impossible to trust a
7612 single self signed certificate. This means that:
7613 openssl verify ss.pem
7614 now gives a warning about a self signed certificate but
7615 openssl verify -CAfile ss.pem ss.pem
7619 *) For servers, store verify_result in SSL_SESSION data structure
7620 (and add it to external session representation).
7621 This is needed when client certificate verifications fails,
7622 but an application-provided verification callback (set by
7623 SSL_CTX_set_cert_verify_callback) allows accepting the session
7624 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
7625 but returns 1): When the session is reused, we have to set
7626 ssl->verify_result to the appropriate error code to avoid
7628 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
7630 *) Fix a bug in the new PKCS#7 code: it didn't consider the
7631 case in PKCS7_dataInit() where the signed PKCS7 structure
7632 didn't contain any existing data because it was being created.
7633 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
7635 *) Add a salt to the key derivation routines in enc.c. This
7636 forms the first 8 bytes of the encrypted file. Also add a
7637 -S option to allow a salt to be input on the command line.
7640 *) New function X509_cmp(). Oddly enough there wasn't a function
7641 to compare two certificates. We do this by working out the SHA1
7642 hash and comparing that. X509_cmp() will be needed by the trust
7646 *) SSL_get1_session() is like SSL_get_session(), but increments
7647 the reference count in the SSL_SESSION returned.
7648 [Geoff Thorpe <geoff@eu.c2.net>]
7650 *) Fix for 'req': it was adding a null to request attributes.
7651 Also change the X509_LOOKUP and X509_INFO code to handle
7652 certificate auxiliary information.
7655 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
7659 *) Add the possibility to add extra information to the memory leak
7660 detecting output, to form tracebacks, showing from where each
7661 allocation was originated: CRYPTO_push_info("constant string") adds
7662 the string plus current file name and line number to a per-thread
7663 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
7664 is like calling CYRPTO_pop_info() until the stack is empty.
7665 Also updated memory leak detection code to be multi-thread-safe.
7668 *) Add options -text and -noout to pkcs7 utility and delete the
7669 encryption options which never did anything. Update docs.
7672 *) Add options to some of the utilities to allow the pass phrase
7673 to be included on either the command line (not recommended on
7674 OSes like Unix) or read from the environment. Update the
7675 manpages and fix a few bugs.
7678 *) Add a few manpages for some of the openssl commands.
7681 *) Fix the -revoke option in ca. It was freeing up memory twice,
7682 leaking and not finding already revoked certificates.
7685 *) Extensive changes to support certificate auxiliary information.
7686 This involves the use of X509_CERT_AUX structure and X509_AUX
7687 functions. An X509_AUX function such as PEM_read_X509_AUX()
7688 can still read in a certificate file in the usual way but it
7689 will also read in any additional "auxiliary information". By
7690 doing things this way a fair degree of compatibility can be
7691 retained: existing certificates can have this information added
7692 using the new 'x509' options.
7694 Current auxiliary information includes an "alias" and some trust
7695 settings. The trust settings will ultimately be used in enhanced
7696 certificate chain verification routines: currently a certificate
7697 can only be trusted if it is self signed and then it is trusted
7701 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
7702 The problem was that one of the replacement routines had not been working
7703 since SSLeay releases. For now the offending routine has been replaced
7704 with non-optimised assembler. Even so, this now gives around 95%
7705 performance improvement for 1024 bit RSA signs.
7708 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
7709 handling. Most clients have the effective key size in bits equal to
7710 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
7711 A few however don't do this and instead use the size of the decrypted key
7712 to determine the RC2 key length and the AlgorithmIdentifier to determine
7713 the effective key length. In this case the effective key length can still
7714 be 40 bits but the key length can be 168 bits for example. This is fixed
7715 by manually forcing an RC2 key into the EVP_PKEY structure because the
7716 EVP code can't currently handle unusual RC2 key sizes: it always assumes
7717 the key length and effective key length are equal.
7720 *) Add a bunch of functions that should simplify the creation of
7721 X509_NAME structures. Now you should be able to do:
7722 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
7723 and have it automatically work out the correct field type and fill in
7724 the structures. The more adventurous can try:
7725 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
7726 and it will (hopefully) work out the correct multibyte encoding.
7729 *) Change the 'req' utility to use the new field handling and multibyte
7730 copy routines. Before the DN field creation was handled in an ad hoc
7731 way in req, ca, and x509 which was rather broken and didn't support
7732 BMPStrings or UTF8Strings. Since some software doesn't implement
7733 BMPStrings or UTF8Strings yet, they can be enabled using the config file
7734 using the dirstring_type option. See the new comment in the default
7735 openssl.cnf for more info.
7738 *) Make crypto/rand/md_rand.c more robust:
7739 - Assure unique random numbers after fork().
7740 - Make sure that concurrent threads access the global counter and
7741 md serializably so that we never lose entropy in them
7742 or use exactly the same state in multiple threads.
7743 Access to the large state is not always serializable because
7744 the additional locking could be a performance killer, and
7745 md should be large enough anyway.
7748 *) New file apps/app_rand.c with commonly needed functionality
7749 for handling the random seed file.
7751 Use the random seed file in some applications that previously did not:
7753 dsaparam -genkey (which also ignored its '-rand' option),
7756 x509 (when signing).
7757 Except on systems with /dev/urandom, it is crucial to have a random
7758 seed file at least for key creation, DSA signing, and for DH exchanges;
7759 for RSA signatures we could do without one.
7761 gendh and gendsa (unlike genrsa) used to read only the first byte
7762 of each file listed in the '-rand' option. The function as previously
7763 found in genrsa is now in app_rand.c and is used by all programs
7764 that support '-rand'.
7767 *) In RAND_write_file, use mode 0600 for creating files;
7768 don't just chmod when it may be too late.
7771 *) Report an error from X509_STORE_load_locations
7772 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
7775 *) New function ASN1_mbstring_copy() this copies a string in either
7776 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
7777 into an ASN1_STRING type. A mask of permissible types is passed
7778 and it chooses the "minimal" type to use or an error if not type
7782 *) Add function equivalents to the various macros in asn1.h. The old
7783 macros are retained with an M_ prefix. Code inside the library can
7784 use the M_ macros. External code (including the openssl utility)
7785 should *NOT* in order to be "shared library friendly".
7788 *) Add various functions that can check a certificate's extensions
7789 to see if it usable for various purposes such as SSL client,
7790 server or S/MIME and CAs of these types. This is currently
7791 VERY EXPERIMENTAL but will ultimately be used for certificate chain
7792 verification. Also added a -purpose flag to x509 utility to
7793 print out all the purposes.
7796 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
7800 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
7801 for, obtain and decode and extension and obtain its critical flag.
7802 This allows all the necessary extension code to be handled in a
7803 single function call.
7806 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
7807 platforms. See crypto/rc4/rc4_enc.c for further details.
7810 *) New -noout option to asn1parse. This causes no output to be produced
7811 its main use is when combined with -strparse and -out to extract data
7812 from a file (which may not be in ASN.1 format).
7815 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
7816 when producing the local key id.
7817 [Richard Levitte <levitte@stacken.kth.se>]
7819 *) New option -dhparam in s_server. This allows a DH parameter file to be
7820 stated explicitly. If it is not stated then it tries the first server
7821 certificate file. The previous behaviour hard coded the filename
7825 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
7826 a public key to be input or output. For example:
7827 openssl rsa -in key.pem -pubout -out pubkey.pem
7828 Also added necessary DSA public key functions to handle this.
7831 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
7832 in the message. This was handled by allowing
7833 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
7834 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
7836 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
7837 to the end of the strings whereas this didn't. This would cause problems
7838 if strings read with d2i_ASN1_bytes() were later modified.
7839 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
7841 *) Fix for base64 decode bug. When a base64 bio reads only one line of
7842 data and it contains EOF it will end up returning an error. This is
7843 caused by input 46 bytes long. The cause is due to the way base64
7844 BIOs find the start of base64 encoded data. They do this by trying a
7845 trial decode on each line until they find one that works. When they
7846 do a flag is set and it starts again knowing it can pass all the
7847 data directly through the decoder. Unfortunately it doesn't reset
7848 the context it uses. This means that if EOF is reached an attempt
7849 is made to pass two EOFs through the context and this causes the
7850 resulting error. This can also cause other problems as well. As is
7851 usual with these problems it takes *ages* to find and the fix is
7852 trivial: move one line.
7853 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
7855 *) Ugly workaround to get s_client and s_server working under Windows. The
7856 old code wouldn't work because it needed to select() on sockets and the
7857 tty (for keypresses and to see if data could be written). Win32 only
7858 supports select() on sockets so we select() with a 1s timeout on the
7859 sockets and then see if any characters are waiting to be read, if none
7860 are present then we retry, we also assume we can always write data to
7861 the tty. This isn't nice because the code then blocks until we've
7862 received a complete line of data and it is effectively polling the
7863 keyboard at 1s intervals: however it's quite a bit better than not
7864 working at all :-) A dedicated Windows application might handle this
7865 with an event loop for example.
7868 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
7869 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
7870 will be called when RSA_sign() and RSA_verify() are used. This is useful
7871 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
7872 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
7873 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
7874 This necessitated the support of an extra signature type NID_md5_sha1
7875 for SSL signatures and modifications to the SSL library to use it instead
7876 of calling RSA_public_decrypt() and RSA_private_encrypt().
7879 *) Add new -verify -CAfile and -CApath options to the crl program, these
7880 will lookup a CRL issuers certificate and verify the signature in a
7881 similar way to the verify program. Tidy up the crl program so it
7882 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
7883 less strict. It will now permit CRL extensions even if it is not
7884 a V2 CRL: this will allow it to tolerate some broken CRLs.
7887 *) Initialize all non-automatic variables each time one of the openssl
7888 sub-programs is started (this is necessary as they may be started
7889 multiple times from the "OpenSSL>" prompt).
7890 [Lennart Bang, Bodo Moeller]
7892 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
7893 removing all other RSA functionality (this is what NO_RSA does). This
7894 is so (for example) those in the US can disable those operations covered
7895 by the RSA patent while allowing storage and parsing of RSA keys and RSA
7899 *) Non-copying interface to BIO pairs.
7900 (still largely untested)
7903 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
7904 ASCII string. This was handled independently in various places before.
7907 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
7908 UTF8 strings a character at a time.
7911 *) Use client_version from client hello to select the protocol
7912 (s23_srvr.c) and for RSA client key exchange verification
7913 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
7916 *) Add various utility functions to handle SPKACs, these were previously
7917 handled by poking round in the structure internals. Added new function
7918 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
7919 print, verify and generate SPKACs. Based on an original idea from
7920 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
7923 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
7926 *) Allow the config file extension section to be overwritten on the
7927 command line. Based on an original idea from Massimiliano Pala
7928 <madwolf@comune.modena.it>. The new option is called -extensions
7929 and can be applied to ca, req and x509. Also -reqexts to override
7930 the request extensions in req and -crlexts to override the crl extensions
7934 *) Add new feature to the SPKAC handling in ca. Now you can include
7935 the same field multiple times by preceding it by "XXXX." for example:
7938 this is the same syntax as used in the req config file.
7941 *) Allow certificate extensions to be added to certificate requests. These
7942 are specified in a 'req_extensions' option of the req section of the
7943 config file. They can be printed out with the -text option to req but
7944 are otherwise ignored at present.
7947 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
7948 data read consists of only the final block it would not decrypted because
7949 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
7950 A misplaced 'break' also meant the decrypted final block might not be
7951 copied until the next read.
7954 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
7955 a few extra parameters to the DH structure: these will be useful if
7956 for example we want the value of 'q' or implement X9.42 DH.
7959 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
7960 provides hooks that allow the default DSA functions or functions on a
7961 "per key" basis to be replaced. This allows hardware acceleration and
7962 hardware key storage to be handled without major modification to the
7963 library. Also added low level modexp hooks and CRYPTO_EX structure and
7964 associated functions.
7967 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
7968 as "read only": it can't be written to and the buffer it points to will
7969 not be freed. Reading from a read only BIO is much more efficient than
7970 a normal memory BIO. This was added because there are several times when
7971 an area of memory needs to be read from a BIO. The previous method was
7972 to create a memory BIO and write the data to it, this results in two
7973 copies of the data and an O(n^2) reading algorithm. There is a new
7974 function BIO_new_mem_buf() which creates a read only memory BIO from
7975 an area of memory. Also modified the PKCS#7 routines to use read only
7979 *) Bugfix: ssl23_get_client_hello did not work properly when called in
7980 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
7981 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
7982 but a retry condition occured while trying to read the rest.
7985 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
7986 NID_pkcs7_encrypted by default: this was wrong since this should almost
7987 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
7988 the encrypted data type: this is a more sensible place to put it and it
7989 allows the PKCS#12 code to be tidied up that duplicated this
7993 *) Changed obj_dat.pl script so it takes its input and output files on
7994 the command line. This should avoid shell escape redirection problems
7998 *) Initial support for certificate extension requests, these are included
7999 in things like Xenroll certificate requests. Included functions to allow
8000 extensions to be obtained and added.
8003 *) -crlf option to s_client and s_server for sending newlines as
8004 CRLF (as required by many protocols).
8007 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8009 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8010 [Ralf S. Engelschall]
8012 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8013 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8015 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8019 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8020 DH parameters/keys (q is lost during that conversion, but the resulting
8021 DH parameters contain its length).
8023 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8024 much faster than DH_generate_parameters (which creates parameters
8025 where p = 2*q + 1), and also the smaller q makes DH computations
8026 much more efficient (160-bit exponentiation instead of 1024-bit
8027 exponentiation); so this provides a convenient way to support DHE
8028 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8029 utter importance to use
8030 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8032 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8033 when such DH parameters are used, because otherwise small subgroup
8034 attacks may become possible!
8037 *) Avoid memory leak in i2d_DHparams.
8040 *) Allow the -k option to be used more than once in the enc program:
8041 this allows the same encrypted message to be read by multiple recipients.
8044 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8045 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8046 it will always use the numerical form of the OID, even if it has a short
8050 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8051 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8052 otherwise bn_mod_exp was called. In the case of hardware keys for example
8053 no private key components need be present and it might store extra data
8054 in the RSA structure, which cannot be accessed from bn_mod_exp.
8055 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8056 private key operations.
8059 *) Added support for SPARC Linux.
8062 *) pem_password_cb function type incompatibly changed from
8063 typedef int pem_password_cb(char *buf, int size, int rwflag);
8065 ....(char *buf, int size, int rwflag, void *userdata);
8066 so that applications can pass data to their callbacks:
8067 The PEM[_ASN1]_{read,write}... functions and macros now take an
8068 additional void * argument, which is just handed through whenever
8069 the password callback is called.
8070 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8072 New function SSL_CTX_set_default_passwd_cb_userdata.
8074 Compatibility note: As many C implementations push function arguments
8075 onto the stack in reverse order, the new library version is likely to
8076 interoperate with programs that have been compiled with the old
8077 pem_password_cb definition (PEM_whatever takes some data that
8078 happens to be on the stack as its last argument, and the callback
8079 just ignores this garbage); but there is no guarantee whatsoever that
8082 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8083 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8084 problems not only on Windows, but also on some Unix platforms.
8085 To avoid problematic command lines, these definitions are now in an
8086 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8087 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8090 *) MIPS III/IV assembler module is reimplemented.
8093 *) More DES library cleanups: remove references to srand/rand and
8094 delete an unused file.
8097 *) Add support for the the free Netwide assembler (NASM) under Win32,
8098 since not many people have MASM (ml) and it can be hard to obtain.
8099 This is currently experimental but it seems to work OK and pass all
8100 the tests. Check out INSTALL.W32 for info.
8103 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8104 without temporary keys kept an extra copy of the server key,
8105 and connections with temporary keys did not free everything in case
8109 *) New function RSA_check_key and new openssl rsa option -check
8110 for verifying the consistency of RSA keys.
8111 [Ulf Moeller, Bodo Moeller]
8113 *) Various changes to make Win32 compile work:
8114 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
8115 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
8116 comparison" warnings.
8117 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
8120 *) Add a debugging option to PKCS#5 v2 key generation function: when
8121 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
8122 derived keys are printed to stderr.
8125 *) Copy the flags in ASN1_STRING_dup().
8126 [Roman E. Pavlov <pre@mo.msk.ru>]
8128 *) The x509 application mishandled signing requests containing DSA
8129 keys when the signing key was also DSA and the parameters didn't match.
8131 It was supposed to omit the parameters when they matched the signing key:
8132 the verifying software was then supposed to automatically use the CA's
8133 parameters if they were absent from the end user certificate.
8135 Omitting parameters is no longer recommended. The test was also
8136 the wrong way round! This was probably due to unusual behaviour in
8137 EVP_cmp_parameters() which returns 1 if the parameters match.
8138 This meant that parameters were omitted when they *didn't* match and
8139 the certificate was useless. Certificates signed with 'ca' didn't have
8141 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
8143 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
8144 The interface is as follows:
8145 Applications can use
8146 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
8147 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
8148 "off" is now the default.
8149 The library internally uses
8150 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
8151 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
8152 to disable memory-checking temporarily.
8154 Some inconsistent states that previously were possible (and were
8155 even the default) are now avoided.
8157 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
8158 with each memory chunk allocated; this is occasionally more helpful
8159 than just having a counter.
8161 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
8163 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
8167 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
8168 which largely parallels "options", but is for changing API behaviour,
8169 whereas "options" are about protocol behaviour.
8170 Initial "mode" flags are:
8172 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
8173 a single record has been written.
8174 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
8175 retries use the same buffer location.
8176 (But all of the contents must be
8180 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
8183 *) Fix problems with no-hmac etc.
8184 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
8186 *) New functions RSA_get_default_method(), RSA_set_method() and
8187 RSA_get_method(). These allows replacement of RSA_METHODs without having
8188 to mess around with the internals of an RSA structure.
8191 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
8192 Also really enable memory leak checks in openssl.c and in some
8194 [Chad C. Mulligan, Bodo Moeller]
8196 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
8197 up the length of negative integers. This has now been simplified to just
8198 store the length when it is first determined and use it later, rather
8199 than trying to keep track of where data is copied and updating it to
8201 [Steve Henson, reported by Brien Wheeler
8202 <bwheeler@authentica-security.com>]
8204 *) Add a new function PKCS7_signatureVerify. This allows the verification
8205 of a PKCS#7 signature but with the signing certificate passed to the
8206 function itself. This contrasts with PKCS7_dataVerify which assumes the
8207 certificate is present in the PKCS#7 structure. This isn't always the
8208 case: certificates can be omitted from a PKCS#7 structure and be
8209 distributed by "out of band" means (such as a certificate database).
8212 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
8213 function prototypes in pem.h, also change util/mkdef.pl to add the
8214 necessary function names.
8217 *) mk1mf.pl (used by Windows builds) did not properly read the
8218 options set by Configure in the top level Makefile, and Configure
8219 was not even able to write more than one option correctly.
8220 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
8223 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
8224 file to be loaded from a BIO or FILE pointer. The BIO version will
8225 for example allow memory BIOs to contain config info.
8228 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
8229 Whoever hopes to achieve shared-library compatibility across versions
8230 must use this, not the compile-time macro.
8231 (Exercise 0.9.4: Which is the minimum library version required by
8233 Note: All this applies only to multi-threaded programs, others don't
8237 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
8238 through a BIO pair triggered the default case, i.e.
8239 SSLerr(...,SSL_R_UNKNOWN_STATE).
8242 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
8243 can use the SSL library even if none of the specific BIOs is
8247 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
8248 for the encoded length.
8249 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
8251 *) Add initial documentation of the X509V3 functions.
8254 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
8255 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
8256 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
8257 secure PKCS#8 private key format with a high iteration count.
8260 *) Fix determination of Perl interpreter: A perl or perl5
8261 _directory_ in $PATH was also accepted as the interpreter.
8262 [Ralf S. Engelschall]
8264 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
8265 wrong with it but it was very old and did things like calling
8266 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
8270 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
8271 to use the new extension code.
8274 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
8275 with macros. This should make it easier to change their form, add extra
8276 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
8280 *) Add to configuration table a new entry that can specify an alternative
8281 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
8282 according to Mark Crispin <MRC@Panda.COM>.
8286 *) DES CBC did not update the IV. Weird.
8289 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
8290 Changing the behaviour of the former might break existing programs --
8291 where IV updating is needed, des_ncbc_encrypt can be used.
8294 *) When bntest is run from "make test" it drives bc to check its
8295 calculations, as well as internally checking them. If an internal check
8296 fails, it needs to cause bc to give a non-zero result or make test carries
8297 on without noticing the failure. Fixed.
8300 *) DES library cleanups.
8303 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
8304 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
8305 ciphers. NOTE: although the key derivation function has been verified
8306 against some published test vectors it has not been extensively tested
8307 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
8311 *) Instead of "mkdir -p", which is not fully portable, use new
8312 Perl script "util/mkdir-p.pl".
8315 *) Rewrite the way password based encryption (PBE) is handled. It used to
8316 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
8317 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
8318 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
8319 the 'parameter' field of the AlgorithmIdentifier is passed to the
8320 underlying key generation function so it must do its own ASN1 parsing.
8321 This has also changed the EVP_PBE_CipherInit() function which now has a
8322 'parameter' argument instead of literal salt and iteration count values
8323 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
8326 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
8327 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
8328 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
8329 KEY" because this clashed with PKCS#8 unencrypted string. Since this
8330 value was just used as a "magic string" and not used directly its
8331 value doesn't matter.
8334 *) Introduce some semblance of const correctness to BN. Shame C doesn't
8338 *) "linux-sparc64" configuration (ultrapenguin).
8339 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
8340 "linux-sparc" configuration.
8341 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
8343 *) config now generates no-xxx options for missing ciphers.
8346 *) Support the EBCDIC character set (work in progress).
8347 File ebcdic.c not yet included because it has a different license.
8348 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8350 *) Support BS2000/OSD-POSIX.
8351 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8353 *) Make callbacks for key generation use void * instead of char *.
8356 *) Make S/MIME samples compile (not yet tested).
8359 *) Additional typesafe stacks.
8362 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
8366 Changes between 0.9.3 and 0.9.3a [29 May 1999]
8368 *) New configuration variant "sco5-gcc".
8370 *) Updated some demos.
8371 [Sean O Riordain, Wade Scholine]
8373 *) Add missing BIO_free at exit of pkcs12 application.
8376 *) Fix memory leak in conf.c.
8379 *) Updates for Win32 to assembler version of MD5.
8382 *) Set #! path to perl in apps/der_chop to where we found it
8383 instead of using a fixed path.
8386 *) SHA library changes for irix64-mips4-cc.
8389 *) Improvements for VMS support.
8393 Changes between 0.9.2b and 0.9.3 [24 May 1999]
8395 *) Bignum library bug fix. IRIX 6 passes "make test" now!
8396 This also avoids the problems with SC4.2 and unpatched SC5.
8397 [Andy Polyakov <appro@fy.chalmers.se>]
8399 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
8400 These are required because of the typesafe stack would otherwise break
8401 existing code. If old code used a structure member which used to be STACK
8402 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
8403 sk_num or sk_value it would produce an error because the num, data members
8404 are not present in STACK_OF. Now it just produces a warning. sk_set
8405 replaces the old method of assigning a value to sk_value
8406 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
8407 that does this will no longer work (and should use sk_set instead) but
8408 this could be regarded as a "questionable" behaviour anyway.
8411 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
8412 correctly handle encrypted S/MIME data.
8415 *) Change type of various DES function arguments from des_cblock
8416 (which means, in function argument declarations, pointer to char)
8417 to des_cblock * (meaning pointer to array with 8 char elements),
8418 which allows the compiler to do more typechecking; it was like
8419 that back in SSLeay, but with lots of ugly casts.
8421 Introduce new type const_des_cblock.
8424 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
8425 problems: find RecipientInfo structure that matches recipient certificate
8426 and initialise the ASN1 structures properly based on passed cipher.
8429 *) Belatedly make the BN tests actually check the results.
8432 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
8433 to and from BNs: it was completely broken. New compilation option
8434 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
8435 key elements as negative integers.
8438 *) Reorganize and speed up MD5.
8439 [Andy Polyakov <appro@fy.chalmers.se>]
8442 [Richard Levitte <richard@levitte.org>]
8444 *) New option -out to asn1parse to allow the parsed structure to be
8445 output to a file. This is most useful when combined with the -strparse
8446 option to examine the output of things like OCTET STRINGS.
8449 *) Make SSL library a little more fool-proof by not requiring any longer
8450 that SSL_set_{accept,connect}_state be called before
8451 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
8452 in many applications because usually everything *appeared* to work as
8453 intended anyway -- now it really works as intended).
8456 *) Move openssl.cnf out of lib/.
8459 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
8460 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
8461 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
8462 [Ralf S. Engelschall]
8464 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
8465 handle PKCS#7 enveloped data properly.
8466 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
8468 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
8469 copying pointers. The cert_st handling is changed by this in
8470 various ways (and thus what used to be known as ctx->default_cert
8471 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
8472 any longer when s->cert does not give us what we need).
8473 ssl_cert_instantiate becomes obsolete by this change.
8474 As soon as we've got the new code right (possibly it already is?),
8475 we have solved a couple of bugs of the earlier code where s->cert
8476 was used as if it could not have been shared with other SSL structures.
8478 Note that using the SSL API in certain dirty ways now will result
8479 in different behaviour than observed with earlier library versions:
8480 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
8481 does not influence s as it used to.
8483 In order to clean up things more thoroughly, inside SSL_SESSION
8484 we don't use CERT any longer, but a new structure SESS_CERT
8485 that holds per-session data (if available); currently, this is
8486 the peer's certificate chain and, for clients, the server's certificate
8487 and temporary key. CERT holds only those values that can have
8488 meaningful defaults in an SSL_CTX.
8491 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
8492 from the internal representation. Various PKCS#7 fixes: remove some
8493 evil casts and set the enc_dig_alg field properly based on the signing
8497 *) Allow PKCS#12 password to be set from the command line or the
8498 environment. Let 'ca' get its config file name from the environment
8499 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
8503 *) Allow certificate policies extension to use an IA5STRING for the
8504 organization field. This is contrary to the PKIX definition but
8505 VeriSign uses it and IE5 only recognises this form. Document 'x509'
8509 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
8510 without disallowing inline assembler and the like for non-pedantic builds.
8513 *) Support Borland C++ builder.
8514 [Janez Jere <jj@void.si>, modified by Ulf Möller]
8519 *) SHA-1 cleanups and performance enhancements.
8520 [Andy Polyakov <appro@fy.chalmers.se>]
8522 *) Sparc v8plus assembler for the bignum library.
8523 [Andy Polyakov <appro@fy.chalmers.se>]
8525 *) Accept any -xxx and +xxx compiler options in Configure.
8528 *) Update HPUX configuration.
8531 *) Add missing sk_<type>_unshift() function to safestack.h
8532 [Ralf S. Engelschall]
8534 *) New function SSL_CTX_use_certificate_chain_file that sets the
8535 "extra_cert"s in addition to the certificate. (This makes sense
8536 only for "PEM" format files, as chains as a whole are not
8540 *) Support verify_depth from the SSL API.
8541 x509_vfy.c had what can be considered an off-by-one-error:
8542 Its depth (which was not part of the external interface)
8543 was actually counting the number of certificates in a chain;
8544 now it really counts the depth.
8547 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
8548 instead of X509err, which often resulted in confusing error
8549 messages since the error codes are not globally unique
8550 (e.g. an alleged error in ssl3_accept when a certificate
8551 didn't match the private key).
8553 *) New function SSL_CTX_set_session_id_context that allows to set a default
8554 value (so that you don't need SSL_set_session_id_context for each
8555 connection using the SSL_CTX).
8558 *) OAEP decoding bug fix.
8561 *) Support INSTALL_PREFIX for package builders, as proposed by
8565 *) New Configure options "threads" and "no-threads". For systems
8566 where the proper compiler options are known (currently Solaris
8567 and Linux), "threads" is the default.
8570 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
8573 *) Install various scripts to $(OPENSSLDIR)/misc, not to
8574 $(INSTALLTOP)/bin -- they shouldn't clutter directories
8575 such as /usr/local/bin.
8578 *) "make linux-shared" to build shared libraries.
8579 [Niels Poppe <niels@netbox.org>]
8581 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
8584 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
8585 extension adding in x509 utility.
8588 *) Remove NOPROTO sections and error code comments.
8591 *) Partial rewrite of the DEF file generator to now parse the ANSI
8595 *) New Configure options --prefix=DIR and --openssldir=DIR.
8598 *) Complete rewrite of the error code script(s). It is all now handled
8599 by one script at the top level which handles error code gathering,
8600 header rewriting and C source file generation. It should be much better
8601 than the old method: it now uses a modified version of Ulf's parser to
8602 read the ANSI prototypes in all header files (thus the old K&R definitions
8603 aren't needed for error creation any more) and do a better job of
8604 translating function codes into names. The old 'ASN1 error code imbedded
8605 in a comment' is no longer necessary and it doesn't use .err files which
8606 have now been deleted. Also the error code call doesn't have to appear all
8607 on one line (which resulted in some large lines...).
8610 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
8613 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
8614 0 (which usually indicates a closed connection), but continue reading.
8617 *) Fix some race conditions.
8620 *) Add support for CRL distribution points extension. Add Certificate
8621 Policies and CRL distribution points documentation.
8624 *) Move the autogenerated header file parts to crypto/opensslconf.h.
8627 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
8628 8 of keying material. Merlin has also confirmed interop with this fix
8629 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
8630 [Merlin Hughes <merlin@baltimore.ie>]
8632 *) Fix lots of warnings.
8633 [Richard Levitte <levitte@stacken.kth.se>]
8635 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
8636 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
8637 [Richard Levitte <levitte@stacken.kth.se>]
8639 *) Fix problems with sizeof(long) == 8.
8640 [Andy Polyakov <appro@fy.chalmers.se>]
8642 *) Change functions to ANSI C.
8645 *) Fix typos in error codes.
8646 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
8648 *) Remove defunct assembler files from Configure.
8651 *) SPARC v8 assembler BIGNUM implementation.
8652 [Andy Polyakov <appro@fy.chalmers.se>]
8654 *) Support for Certificate Policies extension: both print and set.
8655 Various additions to support the r2i method this uses.
8658 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
8659 return a const string when you are expecting an allocated buffer.
8662 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
8663 types DirectoryString and DisplayText.
8666 *) Add code to allow r2i extensions to access the configuration database,
8667 add an LHASH database driver and add several ctx helper functions.
8670 *) Fix an evil bug in bn_expand2() which caused various BN functions to
8671 fail when they extended the size of a BIGNUM.
8674 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
8675 support typesafe stack.
8678 *) Fix typo in SSL_[gs]et_options().
8679 [Nils Frostberg <nils@medcom.se>]
8681 *) Delete various functions and files that belonged to the (now obsolete)
8682 old X509V3 handling code.
8685 *) New Configure option "rsaref".
8688 *) Don't auto-generate pem.h.
8691 *) Introduce type-safe ASN.1 SETs.
8694 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
8695 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
8697 *) Introduce type-safe STACKs. This will almost certainly break lots of code
8698 that links with OpenSSL (well at least cause lots of warnings), but fear
8699 not: the conversion is trivial, and it eliminates loads of evil casts. A
8700 few STACKed things have been converted already. Feel free to convert more.
8701 In the fullness of time, I'll do away with the STACK type altogether.
8704 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
8705 specified in <certfile> by updating the entry in the index.txt file.
8706 This way one no longer has to edit the index.txt file manually for
8707 revoking a certificate. The -revoke option does the gory details now.
8708 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
8710 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
8711 `-text' option at all and this way the `-noout -text' combination was
8712 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
8713 [Ralf S. Engelschall]
8715 *) Make sure a corresponding plain text error message exists for the
8716 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
8717 verify callback function determined that a certificate was revoked.
8718 [Ralf S. Engelschall]
8720 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
8721 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
8722 all available cipers including rc5, which was forgotten until now.
8723 In order to let the testing shell script know which algorithms
8724 are available, a new (up to now undocumented) command
8725 "openssl list-cipher-commands" is used.
8728 *) Bugfix: s_client occasionally would sleep in select() when
8729 it should have checked SSL_pending() first.
8732 *) New functions DSA_do_sign and DSA_do_verify to provide access to
8733 the raw DSA values prior to ASN.1 encoding.
8736 *) Tweaks to Configure
8737 [Niels Poppe <niels@netbox.org>]
8739 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
8743 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
8746 *) New config option to avoid instructions that are illegal on the 80386.
8747 The default code is faster, but requires at least a 486.
8750 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
8751 SSL2_SERVER_VERSION (not used at all) macros, which are now the
8752 same as SSL2_VERSION anyway.
8755 *) New "-showcerts" option for s_client.
8758 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
8759 application. Various cleanups and fixes.
8762 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
8763 modify error routines to work internally. Add error codes and PBE init
8764 to library startup routines.
8767 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
8768 packing functions to asn1 and evp. Changed function names and error
8769 codes along the way.
8772 *) PKCS12 integration: and so it begins... First of several patches to
8773 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
8774 objects to objects.h
8777 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
8778 and display support for Thawte strong extranet extension.
8781 *) Add LinuxPPC support.
8782 [Jeff Dubrule <igor@pobox.org>]
8784 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
8785 bn_div_words in alpha.s.
8786 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
8788 *) Make sure the RSA OAEP test is skipped under -DRSAref because
8789 OAEP isn't supported when OpenSSL is built with RSAref.
8790 [Ulf Moeller <ulf@fitug.de>]
8792 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
8793 so they no longer are missing under -DNOPROTO.
8794 [Soren S. Jorvang <soren@t.dk>]
8797 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
8799 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
8800 doesn't work when the session is reused. Coming soon!
8803 *) Fix a security hole, that allows sessions to be reused in the wrong
8804 context thus bypassing client cert protection! All software that uses
8805 client certs and session caches in multiple contexts NEEDS PATCHING to
8806 allow session reuse! A fuller solution is in the works.
8807 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
8809 *) Some more source tree cleanups (removed obsolete files
8810 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
8811 permission on "config" script to be executable) and a fix for the INSTALL
8813 [Ulf Moeller <ulf@fitug.de>]
8815 *) Remove some legacy and erroneous uses of malloc, free instead of
8817 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
8819 *) Make rsa_oaep_test return non-zero on error.
8820 [Ulf Moeller <ulf@fitug.de>]
8822 *) Add support for native Solaris shared libraries. Configure
8823 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
8824 if someone would make that last step automatic.
8825 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
8827 *) ctx_size was not built with the right compiler during "make links". Fixed.
8830 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
8831 except NULL ciphers". This means the default cipher list will no longer
8832 enable NULL ciphers. They need to be specifically enabled e.g. with
8833 the string "DEFAULT:eNULL".
8836 *) Fix to RSA private encryption routines: if p < q then it would
8837 occasionally produce an invalid result. This will only happen with
8838 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
8841 *) Be less restrictive and allow also `perl util/perlpath.pl
8842 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
8843 because this way one can also use an interpreter named `perl5' (which is
8844 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
8845 installed as `perl').
8846 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8848 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
8849 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8851 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
8852 advapi32.lib to Win32 build and change the pem test comparision
8853 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
8854 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
8855 and crypto/des/ede_cbcm_enc.c.
8858 *) DES quad checksum was broken on big-endian architectures. Fixed.
8861 *) Comment out two functions in bio.h that aren't implemented. Fix up the
8862 Win32 test batch file so it (might) work again. The Win32 test batch file
8863 is horrible: I feel ill....
8866 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
8867 in e_os.h. Audit of header files to check ANSI and non ANSI
8868 sections: 10 functions were absent from non ANSI section and not exported
8869 from Windows DLLs. Fixed up libeay.num for new functions.
8872 *) Make `openssl version' output lines consistent.
8873 [Ralf S. Engelschall]
8875 *) Fix Win32 symbol export lists for BIO functions: Added
8876 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
8877 to ms/libeay{16,32}.def.
8878 [Ralf S. Engelschall]
8880 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
8881 fine under Unix and passes some trivial tests I've now added. But the
8882 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
8883 added to make sure no one expects that this stuff really works in the
8884 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
8885 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
8887 [Ralf S. Engelschall]
8889 *) Fix the generation of two part addresses in perl.
8890 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
8892 *) Add config entry for Linux on MIPS.
8893 [John Tobey <jtobey@channel1.com>]
8895 *) Make links whenever Configure is run, unless we are on Windoze.
8898 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
8899 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
8903 *) Add a useful kludge to allow package maintainers to specify compiler and
8904 other platforms details on the command line without having to patch the
8905 Configure script everytime: One now can use ``perl Configure
8906 <id>:<details>'', i.e. platform ids are allowed to have details appended
8907 to them (seperated by colons). This is treated as there would be a static
8908 pre-configured entry in Configure's %table under key <id> with value
8909 <details> and ``perl Configure <id>'' is called. So, when you want to
8910 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
8911 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
8912 now, which overrides the FreeBSD-elf entry on-the-fly.
8913 [Ralf S. Engelschall]
8915 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
8918 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
8919 on the `perl Configure ...' command line. This way one can compile
8920 OpenSSL libraries with Position Independent Code (PIC) which is needed
8921 for linking it into DSOs.
8922 [Ralf S. Engelschall]
8924 *) Remarkably, export ciphers were totally broken and no-one had noticed!
8928 *) Cleaned up the LICENSE document: The official contact for any license
8929 questions now is the OpenSSL core team under openssl-core@openssl.org.
8930 And add a paragraph about the dual-license situation to make sure people
8931 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
8932 to the OpenSSL toolkit.
8933 [Ralf S. Engelschall]
8935 *) General source tree makefile cleanups: Made `making xxx in yyy...'
8936 display consistent in the source tree and replaced `/bin/rm' by `rm'.
8937 Additonally cleaned up the `make links' target: Remove unnecessary
8938 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
8939 to speed processing and no longer clutter the display with confusing
8940 stuff. Instead only the actually done links are displayed.
8941 [Ralf S. Engelschall]
8943 *) Permit null encryption ciphersuites, used for authentication only. It used
8944 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
8945 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
8949 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
8950 signed attributes when verifying signatures (this would break them),
8951 the detached data encoding was wrong and public keys obtained using
8952 X509_get_pubkey() weren't freed.
8955 *) Add text documentation for the BUFFER functions. Also added a work around
8956 to a Win95 console bug. This was triggered by the password read stuff: the
8957 last character typed gets carried over to the next fread(). If you were
8958 generating a new cert request using 'req' for example then the last
8959 character of the passphrase would be CR which would then enter the first
8963 *) Added the new `Includes OpenSSL Cryptography Software' button as
8964 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
8965 button and can be used by applications based on OpenSSL to show the
8966 relationship to the OpenSSL project.
8967 [Ralf S. Engelschall]
8969 *) Remove confusing variables in function signatures in files
8970 ssl/ssl_lib.c and ssl/ssl.h.
8971 [Lennart Bong <lob@kulthea.stacken.kth.se>]
8973 *) Don't install bss_file.c under PREFIX/include/
8974 [Lennart Bong <lob@kulthea.stacken.kth.se>]
8976 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
8977 functions that return function pointers and has support for NT specific
8978 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
8979 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
8980 unsigned to signed types: this was killing the Win32 compile.
8983 *) Add new certificate file to stack functions,
8984 SSL_add_dir_cert_subjects_to_stack() and
8985 SSL_add_file_cert_subjects_to_stack(). These largely supplant
8986 SSL_load_client_CA_file(), and can be used to add multiple certs easily
8987 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
8988 This means that Apache-SSL and similar packages don't have to mess around
8989 to add as many CAs as they want to the preferred list.
8992 *) Experiment with doxygen documentation. Currently only partially applied to
8994 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
8995 openssl.doxy as the configuration file.
8998 *) Get rid of remaining C++-style comments which strict C compilers hate.
8999 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9001 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9002 compiled in by default: it has problems with large keys.
9005 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9006 DH private keys and/or callback functions which directly correspond to
9007 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9008 is needed for applications which have to configure certificates on a
9009 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9011 For the RSA certificate situation is makes no difference, but
9012 for the DSA certificate situation this fixes the "no shared cipher"
9013 problem where the OpenSSL cipher selection procedure failed because the
9014 temporary keys were not overtaken from the context and the API provided
9015 no way to reconfigure them.
9016 The new functions now let applications reconfigure the stuff and they
9017 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9018 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9019 non-public-API function ssl_cert_instantiate() is used as a helper
9020 function and also to reduce code redundancy inside ssl_rsa.c.
9021 [Ralf S. Engelschall]
9023 *) Move s_server -dcert and -dkey options out of the undocumented feature
9024 area because they are useful for the DSA situation and should be
9025 recognized by the users.
9026 [Ralf S. Engelschall]
9028 *) Fix the cipher decision scheme for export ciphers: the export bits are
9029 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9030 SSL_EXP_MASK. So, the original variable has to be used instead of the
9031 already masked variable.
9032 [Richard Levitte <levitte@stacken.kth.se>]
9034 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9035 [Richard Levitte <levitte@stacken.kth.se>]
9037 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9038 from `int' to `unsigned int' because it's a length and initialized by
9039 EVP_DigestFinal() which expects an `unsigned int *'.
9040 [Richard Levitte <levitte@stacken.kth.se>]
9042 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9043 script. Instead use the usual Shell->Perl transition trick.
9044 [Ralf S. Engelschall]
9046 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9047 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9048 -noout -modulus' as it's already the case for `openssl rsa -noout
9049 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9050 currently the public key is printed (a decision which was already done by
9051 `openssl dsa -modulus' in the past) which serves a similar purpose.
9052 Additionally the NO_RSA no longer completely removes the whole -modulus
9053 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9055 [Ralf S. Engelschall]
9057 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9058 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9059 [Arne Ansper <arne@ats.cyber.ee>]
9061 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9062 to be added. Now both 'req' and 'ca' can use new objects defined in the
9066 *) Add cool BIO that does syslog (or event log on NT).
9067 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9069 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9070 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9071 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9072 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9075 *) Add preliminary config info for new extension code.
9078 *) Make RSA_NO_PADDING really use no padding.
9079 [Ulf Moeller <ulf@fitug.de>]
9081 *) Generate errors when private/public key check is done.
9084 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9085 for some CRL extensions and new objects added.
9088 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9089 key usage extension and fuller support for authority key id.
9092 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9093 padding method for RSA, which is recommended for new applications in PKCS
9094 #1 v2.0 (RFC 2437, October 1998).
9095 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9096 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9097 against Bleichbacher's attack on RSA.
9098 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9101 *) Updates to the new SSL compression code
9102 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9104 *) Fix so that the version number in the master secret, when passed
9105 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9106 (because the server will not accept higher), that the version number
9107 is 0x03,0x01, not 0x03,0x00
9108 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9110 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9111 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9112 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
9115 *) Support for RAW extensions where an arbitrary extension can be
9116 created by including its DER encoding. See apps/openssl.cnf for
9120 *) Make sure latest Perl versions don't interpret some generated C array
9121 code as Perl array code in the crypto/err/err_genc.pl script.
9122 [Lars Weber <3weber@informatik.uni-hamburg.de>]
9124 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
9125 not many people have the assembler. Various Win32 compilation fixes and
9126 update to the INSTALL.W32 file with (hopefully) more accurate Win32
9130 *) Modify configure script 'Configure' to automatically create crypto/date.h
9131 file under Win32 and also build pem.h from pem.org. New script
9132 util/mkfiles.pl to create the MINFO file on environments that can't do a
9133 'make files': perl util/mkfiles.pl >MINFO should work.
9136 *) Major rework of DES function declarations, in the pursuit of correctness
9137 and purity. As a result, many evil casts evaporated, and some weirdness,
9138 too. You may find this causes warnings in your code. Zapping your evil
9139 casts will probably fix them. Mostly.
9142 *) Fix for a typo in asn1.h. Bug fix to object creation script
9143 obj_dat.pl. It considered a zero in an object definition to mean
9144 "end of object": none of the objects in objects.h have any zeros
9145 so it wasn't spotted.
9146 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
9148 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
9149 Masking (CBCM). In the absence of test vectors, the best I have been able
9150 to do is check that the decrypt undoes the encrypt, so far. Send me test
9151 vectors if you have them.
9154 *) Correct calculation of key length for export ciphers (too much space was
9155 allocated for null ciphers). This has not been tested!
9158 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
9159 message is now correct (it understands "crypto" and "ssl" on its
9160 command line). There is also now an "update" option. This will update
9161 the util/ssleay.num and util/libeay.num files with any new functions.
9163 perl util/mkdef.pl crypto ssl update
9164 it will update them.
9167 *) Overhauled the Perl interface (perl/*):
9168 - ported BN stuff to OpenSSL's different BN library
9169 - made the perl/ source tree CVS-aware
9170 - renamed the package from SSLeay to OpenSSL (the files still contain
9171 their history because I've copied them in the repository)
9172 - removed obsolete files (the test scripts will be replaced
9173 by better Test::Harness variants in the future)
9174 [Ralf S. Engelschall]
9176 *) First cut for a very conservative source tree cleanup:
9177 1. merge various obsolete readme texts into doc/ssleay.txt
9178 where we collect the old documents and readme texts.
9179 2. remove the first part of files where I'm already sure that we no
9180 longer need them because of three reasons: either they are just temporary
9181 files which were left by Eric or they are preserved original files where
9182 I've verified that the diff is also available in the CVS via "cvs diff
9183 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
9184 the crypto/md/ stuff).
9185 [Ralf S. Engelschall]
9187 *) More extension code. Incomplete support for subject and issuer alt
9188 name, issuer and authority key id. Change the i2v function parameters
9189 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
9190 what that's for :-) Fix to ASN1 macro which messed up
9191 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
9194 *) Preliminary support for ENUMERATED type. This is largely copied from the
9198 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
9199 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9201 *) Make sure `make rehash' target really finds the `openssl' program.
9202 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9204 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
9205 like to hear about it if this slows down other processors.
9208 *) Add CygWin32 platform information to Configure script.
9209 [Alan Batie <batie@aahz.jf.intel.com>]
9211 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
9212 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
9214 *) New program nseq to manipulate netscape certificate sequences
9217 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
9221 *) Fixes to BN code. Previously the default was to define BN_RECURSION
9222 but the BN code had some problems that would cause failures when
9223 doing certificate verification and some other functions.
9224 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9226 *) Add ASN1 and PEM code to support netscape certificate sequences.
9229 *) Add ASN1 and PEM code to support netscape certificate sequences.
9232 *) Add several PKIX and private extended key usage OIDs.
9235 *) Modify the 'ca' program to handle the new extension code. Modify
9236 openssl.cnf for new extension format, add comments.
9239 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
9240 and add a sample to openssl.cnf so req -x509 now adds appropriate
9244 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
9245 error code, add initial support to X509_print() and x509 application.
9248 *) Takes a deep breath and start addding X509 V3 extension support code. Add
9249 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
9250 stuff is currently isolated and isn't even compiled yet.
9253 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
9254 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
9255 Removed the versions check from X509 routines when loading extensions:
9256 this allows certain broken certificates that don't set the version
9257 properly to be processed.
9260 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
9261 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
9262 can still be regenerated with "make depend".
9265 *) Spelling mistake in C version of CAST-128.
9266 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
9268 *) Changes to the error generation code. The perl script err-code.pl
9269 now reads in the old error codes and retains the old numbers, only
9270 adding new ones if necessary. It also only changes the .err files if new
9271 codes are added. The makefiles have been modified to only insert errors
9272 when needed (to avoid needlessly modifying header files). This is done
9273 by only inserting errors if the .err file is newer than the auto generated
9274 C file. To rebuild all the error codes from scratch (the old behaviour)
9275 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
9276 or delete all the .err files.
9279 *) CAST-128 was incorrectly implemented for short keys. The C version has
9280 been fixed, but is untested. The assembler versions are also fixed, but
9281 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
9282 to regenerate it if needed.
9283 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
9284 Hagino <itojun@kame.net>]
9286 *) File was opened incorrectly in randfile.c.
9287 [Ulf Möller <ulf@fitug.de>]
9289 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
9290 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
9291 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
9292 al: it's just almost always a UTCTime. Note this patch adds new error
9293 codes so do a "make errors" if there are problems.
9296 *) Correct Linux 1 recognition in config.
9297 [Ulf Möller <ulf@fitug.de>]
9299 *) Remove pointless MD5 hash when using DSA keys in ca.
9300 [Anonymous <nobody@replay.com>]
9302 *) Generate an error if given an empty string as a cert directory. Also
9303 generate an error if handed NULL (previously returned 0 to indicate an
9304 error, but didn't set one).
9305 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
9307 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
9310 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
9311 parameters. This was causing a warning which killed off the Win32 compile.
9314 *) Remove C++ style comments from crypto/bn/bn_local.h.
9315 [Neil Costigan <neil.costigan@celocom.com>]
9317 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
9318 based on a text string, looking up short and long names and finally
9319 "dot" format. The "dot" format stuff didn't work. Added new function
9320 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
9321 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
9322 OID is not part of the table.
9325 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
9326 X509_LOOKUP_by_alias().
9329 *) Sort openssl functions by name.
9332 *) Get the gendsa program working (hopefully) and add it to app list. Remove
9333 encryption from sample DSA keys (in case anyone is interested the password
9337 *) Make _all_ *_free functions accept a NULL pointer.
9338 [Frans Heymans <fheymans@isaserver.be>]
9340 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
9342 [Anonymous <nobody@replay.com>]
9344 *) s_server should send the CAfile as acceptable CAs, not its own cert.
9345 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9347 *) Don't blow it for numeric -newkey arguments to apps/req.
9348 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9350 *) Temp key "for export" tests were wrong in s3_srvr.c.
9351 [Anonymous <nobody@replay.com>]
9353 *) Add prototype for temp key callback functions
9354 SSL_CTX_set_tmp_{rsa,dh}_callback().
9357 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
9358 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
9361 *) X509_name_add_entry() freed the wrong thing after an error.
9362 [Arne Ansper <arne@ats.cyber.ee>]
9364 *) rsa_eay.c would attempt to free a NULL context.
9365 [Arne Ansper <arne@ats.cyber.ee>]
9367 *) BIO_s_socket() had a broken should_retry() on Windoze.
9368 [Arne Ansper <arne@ats.cyber.ee>]
9370 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
9371 [Arne Ansper <arne@ats.cyber.ee>]
9373 *) Make sure the already existing X509_STORE->depth variable is initialized
9374 in X509_STORE_new(), but document the fact that this variable is still
9375 unused in the certificate verification process.
9376 [Ralf S. Engelschall]
9378 *) Fix the various library and apps files to free up pkeys obtained from
9379 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
9382 *) Fix reference counting in X509_PUBKEY_get(). This makes
9383 demos/maurice/example2.c work, amongst others, probably.
9384 [Steve Henson and Ben Laurie]
9386 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
9387 `openssl' and second, the shortcut symlinks for the `openssl <command>'
9388 are no longer created. This way we have a single and consistent command
9389 line interface `openssl <command>', similar to `cvs <command>'.
9390 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
9392 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
9393 BIT STRING wrapper always have zero unused bits.
9396 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
9399 *) Make the top-level INSTALL documentation easier to understand.
9402 *) Makefiles updated to exit if an error occurs in a sub-directory
9403 make (including if user presses ^C) [Paul Sutton]
9405 *) Make Montgomery context stuff explicit in RSA data structure.
9408 *) Fix build order of pem and err to allow for generated pem.h.
9411 *) Fix renumbering bug in X509_NAME_delete_entry().
9414 *) Enhanced the err-ins.pl script so it makes the error library number
9415 global and can add a library name. This is needed for external ASN1 and
9416 other error libraries.
9419 *) Fixed sk_insert which never worked properly.
9422 *) Fix ASN1 macros so they can handle indefinite length construted
9423 EXPLICIT tags. Some non standard certificates use these: they can now
9427 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
9428 into a single doc/ssleay.txt bundle. This way the information is still
9429 preserved but no longer messes up this directory. Now it's new room for
9430 the new set of documenation files.
9431 [Ralf S. Engelschall]
9433 *) SETs were incorrectly DER encoded. This was a major pain, because they
9434 shared code with SEQUENCEs, which aren't coded the same. This means that
9435 almost everything to do with SETs or SEQUENCEs has either changed name or
9436 number of arguments.
9437 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
9439 *) Fix test data to work with the above.
9442 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
9443 was already fixed by Eric for 0.9.1 it seems.
9444 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
9446 *) Autodetect FreeBSD3.
9449 *) Fix various bugs in Configure. This affects the following platforms:
9453 unixware-2.0-pentium
9457 *) Eliminate generated files from CVS. Reorder tests to regenerate files
9458 before they are needed.
9461 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
9465 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
9467 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
9468 changed SSLeay to OpenSSL in version strings.
9469 [Ralf S. Engelschall]
9471 *) Some fixups to the top-level documents.
9474 *) Fixed the nasty bug where rsaref.h was not found under compile-time
9475 because the symlink to include/ was missing.
9476 [Ralf S. Engelschall]
9478 *) Incorporated the popular no-RSA/DSA-only patches
9479 which allow to compile a RSA-free SSLeay.
9480 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
9482 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
9483 when "ssleay" is still not found.
9484 [Ralf S. Engelschall]
9486 *) Added more platforms to Configure: Cray T3E, HPUX 11,
9487 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
9489 *) Updated the README file.
9490 [Ralf S. Engelschall]
9492 *) Added various .cvsignore files in the CVS repository subdirs
9493 to make a "cvs update" really silent.
9494 [Ralf S. Engelschall]
9496 *) Recompiled the error-definition header files and added
9497 missing symbols to the Win32 linker tables.
9498 [Ralf S. Engelschall]
9500 *) Cleaned up the top-level documents;
9501 o new files: CHANGES and LICENSE
9502 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
9503 o merged COPYRIGHT into LICENSE
9504 o removed obsolete TODO file
9505 o renamed MICROSOFT to INSTALL.W32
9506 [Ralf S. Engelschall]
9508 *) Removed dummy files from the 0.9.1b source tree:
9509 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
9510 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
9511 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
9512 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
9513 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
9514 [Ralf S. Engelschall]
9516 *) Added various platform portability fixes.
9519 *) The Genesis of the OpenSSL rpject:
9520 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
9521 Young and Tim J. Hudson created while they were working for C2Net until
9523 [The OpenSSL Project]
9526 Changes between 0.9.0b and 0.9.1b [not released]
9528 *) Updated a few CA certificates under certs/
9531 *) Changed some BIGNUM api stuff.
9534 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
9535 DGUX x86, Linux Alpha, etc.
9538 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
9539 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
9543 *) Add -strparse option to asn1pars program which parses nested
9545 [Dr Stephen Henson <shenson@bigfoot.com>]
9547 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
9550 *) DSA fix for "ca" program.
9553 *) Added "-genkey" option to "dsaparam" program.
9556 *) Added RIPE MD160 (rmd160) message digest.
9559 *) Added -a (all) option to "ssleay version" command.
9562 *) Added PLATFORM define which is the id given to Configure.
9565 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
9568 *) Extended the ASN.1 parser routines.
9571 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
9574 *) Added a BN_CTX to the BN library.
9577 *) Fixed the weak key values in DES library
9580 *) Changed API in EVP library for cipher aliases.
9583 *) Added support for RC2/64bit cipher.
9586 *) Converted the lhash library to the crypto/mem.c functions.
9589 *) Added more recognized ASN.1 object ids.
9592 *) Added more RSA padding checks for SSL/TLS.
9595 *) Added BIO proxy/filter functionality.
9598 *) Added extra_certs to SSL_CTX which can be used
9599 send extra CA certificates to the client in the CA cert chain sending
9600 process. It can be configured with SSL_CTX_add_extra_chain_cert().
9603 *) Now Fortezza is denied in the authentication phase because
9604 this is key exchange mechanism is not supported by SSLeay at all.
9607 *) Additional PKCS1 checks.
9610 *) Support the string "TLSv1" for all TLS v1 ciphers.
9613 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
9614 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
9617 *) Fixed a few memory leaks.
9620 *) Fixed various code and comment typos.
9623 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
9624 bytes sent in the client random.
9625 [Edward Bishop <ebishop@spyglass.com>]