5 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8 Thanks to Antonio Martin, Enterprise Secure Access Research and
9 Development, Cisco Systems, Inc. for discovering this bug and
10 preparing a fix. (CVE-2012-0050)
13 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
15 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
16 of the Vaudenay padding oracle attack on CBC mode encryption
17 which enables an efficient plaintext recovery attack against
18 the OpenSSL implementation of DTLS. Their attack exploits timing
19 differences arising during decryption processing. A research
20 paper describing this attack can be found at:
21 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
22 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
23 Security Group at Royal Holloway, University of London
24 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
25 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
26 for preparing the fix. (CVE-2011-4108)
27 [Robin Seggelmann, Michael Tuexen]
29 *) Clear bytes used for block padding of SSL 3.0 records.
31 [Adam Langley (Google)]
33 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
34 Kadianakis <desnacked@gmail.com> for discovering this issue and
35 Adam Langley for preparing the fix. (CVE-2011-4619)
36 [Adam Langley (Google)]
38 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
39 [Andrey Kulikov <amdeich@gmail.com>]
41 *) Prevent malformed RFC3779 data triggering an assertion failure.
42 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
43 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
44 [Rob Austein <sra@hactrn.net>]
46 *) Improved PRNG seeding for VOS.
47 [Paul Green <Paul.Green@stratus.com>]
49 *) Fix ssl_ciph.c set-up race.
50 [Adam Langley (Google)]
52 *) Fix spurious failures in ecdsatest.c.
53 [Emilia Käsper (Google)]
55 *) Fix the BIO_f_buffer() implementation (which was mixing different
56 interpretations of the '..._len' fields).
57 [Adam Langley (Google)]
59 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
60 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
61 threads won't reuse the same blinding coefficients.
63 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
64 lock to call BN_BLINDING_invert_ex, and avoids one use of
65 BN_BLINDING_update for each BN_BLINDING structure (previously,
66 the last update always remained unused).
67 [Emilia Käsper (Google)]
69 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
70 [Bob Buckholz (Google)]
72 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
74 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
75 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
76 [Kaspar Brand <ossl@velox.ch>]
78 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
79 for multi-threaded use of ECDH. (CVE-2011-3210)
80 [Adam Langley (Google)]
82 *) Fix x509_name_ex_d2i memory leak on bad inputs.
85 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
86 signature public key algorithm by using OID xref utilities instead.
87 Before this you could only use some ECC ciphersuites with SHA1 only.
90 *) Add protection against ECDSA timing attacks as mentioned in the paper
91 by Billy Bob Brumley and Nicola Tuveri, see:
93 http://eprint.iacr.org/2011/232.pdf
95 [Billy Bob Brumley and Nicola Tuveri]
97 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
99 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
100 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
102 *) Fix bug in string printing code: if *any* escaping is enabled we must
103 escape the escape character (backslash) or the resulting string is
107 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
109 *) Disable code workaround for ancient and obsolete Netscape browsers
110 and servers: an attacker can use it in a ciphersuite downgrade attack.
111 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
114 *) Fixed J-PAKE implementation error, originally discovered by
115 Sebastien Martini, further info and confirmation from Stefan
116 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
119 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
121 *) Fix extension code to avoid race conditions which can result in a buffer
122 overrun vulnerability: resumed sessions must not be modified as they can
123 be shared by multiple threads. CVE-2010-3864
126 *) Fix WIN32 build system to correctly link an ENGINE directory into
130 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
132 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
134 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
136 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
138 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
139 context. The operation can be customised via the ctrl mechanism in
140 case ENGINEs want to include additional functionality.
143 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
146 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
147 output hashes compatible with older versions of OpenSSL.
148 [Willy Weisz <weisz@vcpc.univie.ac.at>]
150 *) Fix compression algorithm handling: if resuming a session use the
151 compression algorithm of the resumed session instead of determining
152 it from client hello again. Don't allow server to change algorithm.
155 *) Add load_crls() function to apps tidying load_certs() too. Add option
156 to verify utility to allow additional CRLs to be included.
159 *) Update OCSP request code to permit adding custom headers to the request:
160 some responders need this.
163 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
165 [Julia Lawall <julia@diku.dk>]
167 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
168 needlessly dereferenced structures, used obsolete functions and
169 didn't handle all updated verify codes correctly.
172 *) Disable MD2 in the default configuration.
175 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
176 indicate the initial BIO being pushed or popped. This makes it possible
177 to determine whether the BIO is the one explicitly called or as a result
178 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
179 it handles reference counts correctly and doesn't zero out the I/O bio
180 when it is not being explicitly popped. WARNING: applications which
181 included workarounds for the old buggy behaviour will need to be modified
182 or they could free up already freed BIOs.
185 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
186 renaming to all platforms (within the 0.9.8 branch, this was
187 done conditionally on Netware platforms to avoid a name clash).
188 [Guenter <lists@gknw.net>]
190 *) Add ECDHE and PSK support to DTLS.
191 [Michael Tuexen <tuexen@fh-muenster.de>]
193 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
197 *) Add "missing" function EVP_MD_flags() (without this the only way to
198 retrieve a digest flags is by accessing the structure directly. Update
199 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
200 or cipher is registered as in the "from" argument. Print out all
201 registered digests in the dgst usage message instead of manually
202 attempting to work them out.
205 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
206 this allows the use of compression and extensions. Change default cipher
207 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
208 by default unless an application cipher string requests it.
211 *) Alter match criteria in PKCS12_parse(). It used to try to use local
212 key ids to find matching certificates and keys but some PKCS#12 files
213 don't follow the (somewhat unwritten) rules and this strategy fails.
214 Now just gather all certificates together and the first private key
215 then look for the first certificate that matches the key.
218 *) Support use of registered digest and cipher names for dgst and cipher
219 commands instead of having to add each one as a special case. So now
226 openssl dgst -sha256 foo
228 and this works for ENGINE based algorithms too.
232 *) Update Gost ENGINE to support parameter files.
233 [Victor B. Wagner <vitus@cryptocom.ru>]
235 *) Support GeneralizedTime in ca utility.
236 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
238 *) Enhance the hash format used for certificate directory links. The new
239 form uses the canonical encoding (meaning equivalent names will work
240 even if they aren't identical) and uses SHA1 instead of MD5. This form
241 is incompatible with the older format and as a result c_rehash should
242 be used to rebuild symbolic links.
245 *) Make PKCS#8 the default write format for private keys, replacing the
246 traditional format. This form is standardised, more secure and doesn't
247 include an implicit MD5 dependency.
250 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
251 committed to OpenSSL should pass this lot as a minimum.
254 *) Add session ticket override functionality for use by EAP-FAST.
255 [Jouni Malinen <j@w1.fi>]
257 *) Modify HMAC functions to return a value. Since these can be implemented
258 in an ENGINE errors can occur.
261 *) Type-checked OBJ_bsearch_ex.
264 *) Type-checked OBJ_bsearch. Also some constification necessitated
265 by type-checking. Still to come: TXT_DB, bsearch(?),
266 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
270 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
271 seconds to a tm structure directly, instead of going through OS
272 specific date routines. This avoids any issues with OS routines such
273 as the year 2038 bug. New *_adj() functions for ASN1 time structures
274 and X509_time_adj_ex() to cover the extended range. The existing
275 X509_time_adj() is still usable and will no longer have any date issues.
278 *) Delta CRL support. New use deltas option which will attempt to locate
279 and search any appropriate delta CRLs available.
281 This work was sponsored by Google.
284 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
285 code and add additional score elements. Validate alternate CRL paths
286 as part of the CRL checking and indicate a new error "CRL path validation
287 error" in this case. Applications wanting additional details can use
288 the verify callback and check the new "parent" field. If this is not
289 NULL CRL path validation is taking place. Existing applications wont
290 see this because it requires extended CRL support which is off by
293 This work was sponsored by Google.
296 *) Support for freshest CRL extension.
298 This work was sponsored by Google.
301 *) Initial indirect CRL support. Currently only supported in the CRLs
302 passed directly and not via lookup. Process certificate issuer
303 CRL entry extension and lookup CRL entries by bother issuer name
304 and serial number. Check and process CRL issuer entry in IDP extension.
306 This work was sponsored by Google.
309 *) Add support for distinct certificate and CRL paths. The CRL issuer
310 certificate is validated separately in this case. Only enabled if
311 an extended CRL support flag is set: this flag will enable additional
312 CRL functionality in future.
314 This work was sponsored by Google.
317 *) Add support for policy mappings extension.
319 This work was sponsored by Google.
322 *) Fixes to pathlength constraint, self issued certificate handling,
323 policy processing to align with RFC3280 and PKITS tests.
325 This work was sponsored by Google.
328 *) Support for name constraints certificate extension. DN, email, DNS
329 and URI types are currently supported.
331 This work was sponsored by Google.
334 *) To cater for systems that provide a pointer-based thread ID rather
335 than numeric, deprecate the current numeric thread ID mechanism and
336 replace it with a structure and associated callback type. This
337 mechanism allows a numeric "hash" to be extracted from a thread ID in
338 either case, and on platforms where pointers are larger than 'long',
339 mixing is done to help ensure the numeric 'hash' is usable even if it
340 can't be guaranteed unique. The default mechanism is to use "&errno"
341 as a pointer-based thread ID to distinguish between threads.
343 Applications that want to provide their own thread IDs should now use
344 CRYPTO_THREADID_set_callback() to register a callback that will call
345 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
347 Note that ERR_remove_state() is now deprecated, because it is tied
348 to the assumption that thread IDs are numeric. ERR_remove_state(0)
349 to free the current thread's error state should be replaced by
350 ERR_remove_thread_state(NULL).
352 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
353 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
354 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
355 application was previously providing a numeric thread callback that
356 was inappropriate for distinguishing threads, then uniqueness might
357 have been obtained with &errno that happened immediately in the
358 intermediate development versions of OpenSSL; this is no longer the
359 case, the numeric thread callback will now override the automatic use
361 [Geoff Thorpe, with help from Bodo Moeller]
363 *) Initial support for different CRL issuing certificates. This covers a
364 simple case where the self issued certificates in the chain exist and
365 the real CRL issuer is higher in the existing chain.
367 This work was sponsored by Google.
370 *) Removed effectively defunct crypto/store from the build.
373 *) Revamp of STACK to provide stronger type-checking. Still to come:
374 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
375 ASN1_STRING, CONF_VALUE.
378 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
379 RAM on SSL connections. This option can save about 34k per idle SSL.
382 *) Revamp of LHASH to provide stronger type-checking. Still to come:
383 STACK, TXT_DB, bsearch, qsort.
386 *) Initial support for Cryptographic Message Syntax (aka CMS) based
387 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
388 support for data, signedData, compressedData, digestedData and
389 encryptedData, envelopedData types included. Scripts to check against
390 RFC4134 examples draft and interop and consistency checks of many
391 content types and variants.
394 *) Add options to enc utility to support use of zlib compression BIO.
397 *) Extend mk1mf to support importing of options and assembly language
398 files from Configure script, currently only included in VC-WIN32.
399 The assembly language rules can now optionally generate the source
400 files from the associated perl scripts.
403 *) Implement remaining functionality needed to support GOST ciphersuites.
404 Interop testing has been performed using CryptoPro implementations.
405 [Victor B. Wagner <vitus@cryptocom.ru>]
407 *) s390x assembler pack.
410 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
414 *) Implement Opaque PRF Input TLS extension as specified in
415 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
416 official specification yet and no extension type assignment by
417 IANA exists, this extension (for now) will have to be explicitly
418 enabled when building OpenSSL by providing the extension number
419 to use. For example, specify an option
421 -DTLSEXT_TYPE_opaque_prf_input=0x9527
423 to the "config" or "Configure" script to enable the extension,
424 assuming extension number 0x9527 (which is a completely arbitrary
425 and unofficial assignment based on the MD5 hash of the Internet
426 Draft). Note that by doing so, you potentially lose
427 interoperability with other TLS implementations since these might
428 be using the same extension number for other purposes.
430 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
431 opaque PRF input value to use in the handshake. This will create
432 an interal copy of the length-'len' string at 'src', and will
433 return non-zero for success.
435 To get more control and flexibility, provide a callback function
438 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
439 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
443 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
446 Callback function 'cb' will be called in handshakes, and is
447 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
448 Argument 'arg' is for application purposes (the value as given to
449 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
450 be provided to the callback function). The callback function
451 has to return non-zero to report success: usually 1 to use opaque
452 PRF input just if possible, or 2 to enforce use of the opaque PRF
453 input. In the latter case, the library will abort the handshake
454 if opaque PRF input is not successfully negotiated.
456 Arguments 'peerinput' and 'len' given to the callback function
457 will always be NULL and 0 in the case of a client. A server will
458 see the client's opaque PRF input through these variables if
459 available (NULL and 0 otherwise). Note that if the server
460 provides an opaque PRF input, the length must be the same as the
461 length of the client's opaque PRF input.
463 Note that the callback function will only be called when creating
464 a new session (session resumption can resume whatever was
465 previously negotiated), and will not be called in SSL 2.0
466 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
467 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
468 for applications that need to enforce opaque PRF input.
472 *) Update ssl code to support digests other than SHA1+MD5 for handshake
475 [Victor B. Wagner <vitus@cryptocom.ru>]
477 *) Add RFC4507 support to OpenSSL. This includes the corrections in
478 RFC4507bis. The encrypted ticket format is an encrypted encoded
479 SSL_SESSION structure, that way new session features are automatically
482 If a client application caches session in an SSL_SESSION structure
483 support is transparent because tickets are now stored in the encoded
486 The SSL_CTX structure automatically generates keys for ticket
487 protection in servers so again support should be possible
488 with no application modification.
490 If a client or server wishes to disable RFC4507 support then the option
491 SSL_OP_NO_TICKET can be set.
493 Add a TLS extension debugging callback to allow the contents of any client
494 or server extensions to be examined.
496 This work was sponsored by Google.
499 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
500 OpenSSL should now compile cleanly on gcc 4.2
501 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
503 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
504 support including streaming MAC support: this is required for GOST
506 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
508 *) Add option -stream to use PKCS#7 streaming in smime utility. New
509 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
510 to output in BER and PEM format.
513 *) Experimental support for use of HMAC via EVP_PKEY interface. This
514 allows HMAC to be handled via the EVP_DigestSign*() interface. The
515 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
516 ENGINE support for HMAC keys which are unextractable. New -mac and
517 -macopt options to dgst utility.
520 *) New option -sigopt to dgst utility. Update dgst to use
521 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
522 alternative signing paramaters such as X9.31 or PSS in the dgst
526 *) Change ssl_cipher_apply_rule(), the internal function that does
527 the work each time a ciphersuite string requests enabling
528 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
529 removing ("!foo+bar") a class of ciphersuites: Now it maintains
530 the order of disabled ciphersuites such that those ciphersuites
531 that most recently went from enabled to disabled not only stay
532 in order with respect to each other, but also have higher priority
533 than other disabled ciphersuites the next time ciphersuites are
536 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
537 the same ciphersuites as with "HIGH" alone, but in a specific
538 order where the PSK ciphersuites come first (since they are the
539 most recently disabled ciphersuites when "HIGH" is parsed).
541 Also, change ssl_create_cipher_list() (using this new
542 funcionality) such that between otherwise identical
543 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
547 *) Change ssl_create_cipher_list() so that it automatically
548 arranges the ciphersuites in reasonable order before starting
549 to process the rule string. Thus, the definition for "DEFAULT"
550 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
551 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
552 This makes it much easier to arrive at a reasonable default order
553 in applications for which anonymous ciphers are OK (meaning
554 that you can't actually use DEFAULT).
555 [Bodo Moeller; suggested by Victor Duchovni]
557 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
558 processing) into multiple integers instead of setting
559 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
560 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
561 (These masks as well as the individual bit definitions are hidden
562 away into the non-exported interface ssl/ssl_locl.h, so this
563 change to the definition of the SSL_CIPHER structure shouldn't
564 affect applications.) This give us more bits for each of these
565 categories, so there is no longer a need to coagulate AES128 and
566 AES256 into a single algorithm bit, and to coagulate Camellia128
567 and Camellia256 into a single algorithm bit, which has led to all
570 Thus, among other things, the kludge introduced in 0.9.7m and
571 0.9.8e for masking out AES256 independently of AES128 or masking
572 out Camellia256 independently of AES256 is not needed here in 0.9.9.
574 With the change, we also introduce new ciphersuite aliases that
575 so far were missing: "AES128", "AES256", "CAMELLIA128", and
579 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
580 Use the leftmost N bytes of the signature input if the input is
581 larger than the prime q (with N being the size in bytes of q).
584 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
585 it yet and it is largely untested.
588 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
591 *) Initial incomplete changes to avoid need for function casts in OpenSSL
592 some compilers (gcc 4.2 and later) reject their use. Safestack is
593 reimplemented. Update ASN1 to avoid use of legacy functions.
596 *) Win32/64 targets are linked with Winsock2.
599 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
600 to external functions. This can be used to increase CRL handling
601 efficiency especially when CRLs are very large by (for example) storing
602 the CRL revoked certificates in a database.
605 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
606 new CRLs added to a directory can be used. New command line option
607 -verify_return_error to s_client and s_server. This causes real errors
608 to be returned by the verify callback instead of carrying on no matter
609 what. This reflects the way a "real world" verify callback would behave.
612 *) GOST engine, supporting several GOST algorithms and public key formats.
613 Kindly donated by Cryptocom.
616 *) Partial support for Issuing Distribution Point CRL extension. CRLs
617 partitioned by DP are handled but no indirect CRL or reason partitioning
618 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
619 selected via a scoring technique which handles IDP and AKID in CRLs.
622 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
623 will ultimately be used for all verify operations: this will remove the
624 X509_STORE dependency on certificate verification and allow alternative
625 lookup methods. X509_STORE based implementations of these two callbacks.
628 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
629 Modify get_crl() to find a valid (unexpired) CRL if possible.
632 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
633 this would be called X509_CRL_cmp() but that name is already used by
634 a function that just compares CRL issuer names. Cache several CRL
635 extensions in X509_CRL structure and cache CRLDP in X509.
638 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
639 this maps equivalent X509_NAME structures into a consistent structure.
640 Name comparison can then be performed rapidly using memcmp().
643 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
647 *) Allow digests to supply their own micalg string for S/MIME type using
648 the ctrl EVP_MD_CTRL_MICALG.
651 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
652 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
653 ctrl. It can then customise the structure before and/or after signing
657 *) New function OBJ_add_sigid() to allow application defined signature OIDs
658 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
659 to free up any added signature OIDs.
662 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
663 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
664 digest and cipher tables. New options added to openssl utility:
665 list-message-digest-algorithms and list-cipher-algorithms.
668 *) Change the array representation of binary polynomials: the list
669 of degrees of non-zero coefficients is now terminated with -1.
670 Previously it was terminated with 0, which was also part of the
671 value; thus, the array representation was not applicable to
672 polynomials where t^0 has coefficient zero. This change makes
673 the array representation useful in a more general context.
676 *) Various modifications and fixes to SSL/TLS cipher string
677 handling. For ECC, the code now distinguishes between fixed ECDH
678 with RSA certificates on the one hand and with ECDSA certificates
679 on the other hand, since these are separate ciphersuites. The
680 unused code for Fortezza ciphersuites has been removed.
682 For consistency with EDH, ephemeral ECDH is now called "EECDH"
683 (not "ECDHE"). For consistency with the code for DH
684 certificates, use of ECDH certificates is now considered ECDH
685 authentication, not RSA or ECDSA authentication (the latter is
686 merely the CA's signing algorithm and not actively used in the
689 The temporary ciphersuite alias "ECCdraft" is no longer
690 available, and ECC ciphersuites are no longer excluded from "ALL"
691 and "DEFAULT". The following aliases now exist for RFC 4492
692 ciphersuites, most of these by analogy with the DH case:
694 kECDHr - ECDH cert, signed with RSA
695 kECDHe - ECDH cert, signed with ECDSA
696 kECDH - ECDH cert (signed with either RSA or ECDSA)
697 kEECDH - ephemeral ECDH
698 ECDH - ECDH cert or ephemeral ECDH
704 AECDH - anonymous ECDH
705 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
709 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
710 Use correct micalg parameters depending on digest(s) in signed message.
713 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
714 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
717 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
718 an engine to register a method. Add ENGINE lookups for methods and
719 functional reference processing.
722 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
723 EVP_{Sign,Verify}* which allow an application to customise the signature
727 *) New -resign option to smime utility. This adds one or more signers
728 to an existing PKCS#7 signedData structure. Also -md option to use an
729 alternative message digest algorithm for signing.
732 *) Tidy up PKCS#7 routines and add new functions to make it easier to
733 create PKCS7 structures containing multiple signers. Update smime
734 application to support multiple signers.
737 *) New -macalg option to pkcs12 utility to allow setting of an alternative
741 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
742 Reorganize PBE internals to lookup from a static table using NIDs,
743 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
744 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
745 PRF which will be automatically used with PBES2.
748 *) Replace the algorithm specific calls to generate keys in "req" with the
752 *) Update PKCS#7 enveloped data routines to use new API. This is now
753 supported by any public key method supporting the encrypt operation. A
754 ctrl is added to allow the public key algorithm to examine or modify
755 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
759 *) Add a ctrl to asn1 method to allow a public key algorithm to express
760 a default digest type to use. In most cases this will be SHA1 but some
761 algorithms (such as GOST) need to specify an alternative digest. The
762 return value indicates how strong the prefernce is 1 means optional and
763 2 is mandatory (that is it is the only supported type). Modify
764 ASN1_item_sign() to accept a NULL digest argument to indicate it should
765 use the default md. Update openssl utilities to use the default digest
766 type for signing if it is not explicitly indicated.
769 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
770 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
771 signing method from the key type. This effectively removes the link
772 between digests and public key types.
775 *) Add an OID cross reference table and utility functions. Its purpose is to
776 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
777 rsaEncryption. This will allow some of the algorithm specific hackery
778 needed to use the correct OID to be removed.
781 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
782 structures for PKCS7_sign(). They are now set up by the relevant public
786 *) Add provisional EC pkey method with support for ECDSA and ECDH.
789 *) Add support for key derivation (agreement) in the API, DH method and
793 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
794 public and private key formats. As a side effect these add additional
795 command line functionality not previously available: DSA signatures can be
796 generated and verified using pkeyutl and DH key support and generation in
801 [Oliver Tappe <zooey@hirschkaefer.de>]
803 *) New make target "install_html_docs" installs HTML renditions of the
805 [Oliver Tappe <zooey@hirschkaefer.de>]
807 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
808 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
809 support key and parameter generation and add initial key generation
810 functionality for RSA.
813 *) Add functions for main EVP_PKEY_method operations. The undocumented
814 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
815 EVP_PKEY_{encrypt,decrypt}_old.
818 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
819 key API, doesn't do much yet.
822 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
823 public key algorithms. New option to openssl utility:
824 "list-public-key-algorithms" to print out info.
827 *) Implement the Supported Elliptic Curves Extension for
828 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
831 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
832 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
835 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
836 utilities such as rsa, dsa, dsaparam etc except they process any key
840 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
841 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
842 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
846 *) Initial support for pluggable public key ASN1.
847 De-spaghettify the public key ASN1 handling. Move public and private
848 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
849 algorithm specific handling to a single module within the relevant
850 algorithm directory. Add functions to allow (near) opaque processing
851 of public and private key structures.
854 *) Implement the Supported Point Formats Extension for
855 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
858 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
859 for the psk identity [hint] and the psk callback functions to the
860 SSL_SESSION, SSL and SSL_CTX structure.
863 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
867 SSL_CTX_use_psk_identity_hint
868 SSL_get_psk_identity_hint
870 SSL_use_psk_identity_hint
872 [Mika Kousa and Pasi Eronen of Nokia Corporation]
874 *) Add RFC 3161 compliant time stamp request creation, response generation
875 and response verification functionality.
876 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
878 *) Add initial support for TLS extensions, specifically for the server_name
879 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
880 have new members for a host name. The SSL data structure has an
881 additional member SSL_CTX *initial_ctx so that new sessions can be
882 stored in that context to allow for session resumption, even after the
883 SSL has been switched to a new SSL_CTX in reaction to a client's
884 server_name extension.
886 New functions (subject to change):
889 SSL_get_servername_type()
892 New CTRL codes and macros (subject to change):
894 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
895 - SSL_CTX_set_tlsext_servername_callback()
896 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
897 - SSL_CTX_set_tlsext_servername_arg()
898 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
900 openssl s_client has a new '-servername ...' option.
902 openssl s_server has new options '-servername_host ...', '-cert2 ...',
903 '-key2 ...', '-servername_fatal' (subject to change). This allows
904 testing the HostName extension for a specific single host name ('-cert'
905 and '-key' remain fallbacks for handshakes without HostName
906 negotiation). If the unrecogninzed_name alert has to be sent, this by
907 default is a warning; it becomes fatal with the '-servername_fatal'
910 [Peter Sylvester, Remy Allais, Christophe Renou]
912 *) Whirlpool hash implementation is added.
915 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
916 bn(64,32). Because of instruction set limitations it doesn't have
917 any negative impact on performance. This was done mostly in order
918 to make it possible to share assembler modules, such as bn_mul_mont
919 implementations, between 32- and 64-bit builds without hassle.
922 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
923 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
927 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
928 dedicated Montgomery multiplication procedure, is introduced.
929 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
930 "64-bit" performance on certain 32-bit targets.
933 *) New option SSL_OP_NO_COMP to disable use of compression selectively
934 in SSL structures. New SSL ctrl to set maximum send fragment size.
935 Save memory by seeting the I/O buffer sizes dynamically instead of
936 using the maximum available value.
939 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
940 in addition to the text details.
943 *) Very, very preliminary EXPERIMENTAL support for printing of general
944 ASN1 structures. This currently produces rather ugly output and doesn't
945 handle several customised structures at all.
948 *) Integrated support for PVK file format and some related formats such
949 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
950 these in the 'rsa' and 'dsa' utilities.
953 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
956 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
957 place for the (very old) "NETSCAPE" format certificates which are now
958 handled using new ASN1 code equivalents.
961 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
962 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
963 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
966 *) Modify CRL distribution points extension code to print out previously
967 unsupported fields. Enhance extension setting code to allow setting of
971 *) Add print and set support for Issuing Distribution Point CRL extension.
974 *) Change 'Configure' script to enable Camellia by default.
977 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
979 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
980 Thanks to Antonio Martin, Enterprise Secure Access Research and
981 Development, Cisco Systems, Inc. for discovering this bug and
982 preparing a fix. (CVE-2012-0050)
985 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
987 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
988 of the Vaudenay padding oracle attack on CBC mode encryption
989 which enables an efficient plaintext recovery attack against
990 the OpenSSL implementation of DTLS. Their attack exploits timing
991 differences arising during decryption processing. A research
992 paper describing this attack can be found at:
993 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
994 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
995 Security Group at Royal Holloway, University of London
996 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
997 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
998 for preparing the fix. (CVE-2011-4108)
999 [Robin Seggelmann, Michael Tuexen]
1001 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
1002 [Ben Laurie, Kasper <ekasper@google.com>]
1004 *) Clear bytes used for block padding of SSL 3.0 records.
1006 [Adam Langley (Google)]
1008 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1009 Kadianakis <desnacked@gmail.com> for discovering this issue and
1010 Adam Langley for preparing the fix. (CVE-2011-4619)
1011 [Adam Langley (Google)]
1013 *) Prevent malformed RFC3779 data triggering an assertion failure.
1014 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1015 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1016 [Rob Austein <sra@hactrn.net>]
1018 *) Fix ssl_ciph.c set-up race.
1019 [Adam Langley (Google)]
1021 *) Fix spurious failures in ecdsatest.c.
1022 [Emilia Käsper (Google)]
1024 *) Fix the BIO_f_buffer() implementation (which was mixing different
1025 interpretations of the '..._len' fields).
1026 [Adam Langley (Google)]
1028 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1029 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1030 threads won't reuse the same blinding coefficients.
1032 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1033 lock to call BN_BLINDING_invert_ex, and avoids one use of
1034 BN_BLINDING_update for each BN_BLINDING structure (previously,
1035 the last update always remained unused).
1036 [Emilia Käsper (Google)]
1038 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1039 for multi-threaded use of ECDH.
1040 [Adam Langley (Google)]
1042 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1045 *) Add protection against ECDSA timing attacks as mentioned in the paper
1046 by Billy Bob Brumley and Nicola Tuveri, see:
1048 http://eprint.iacr.org/2011/232.pdf
1050 [Billy Bob Brumley and Nicola Tuveri]
1052 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
1054 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1055 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1057 *) Fix bug in string printing code: if *any* escaping is enabled we must
1058 escape the escape character (backslash) or the resulting string is
1062 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
1064 *) Disable code workaround for ancient and obsolete Netscape browsers
1065 and servers: an attacker can use it in a ciphersuite downgrade attack.
1066 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1069 *) Fixed J-PAKE implementation error, originally discovered by
1070 Sebastien Martini, further info and confirmation from Stefan
1071 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1074 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
1076 *) Fix extension code to avoid race conditions which can result in a buffer
1077 overrun vulnerability: resumed sessions must not be modified as they can
1078 be shared by multiple threads. CVE-2010-3864
1081 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
1084 *) Don't reencode certificate when calculating signature: cache and use
1085 the original encoding instead. This makes signature verification of
1086 some broken encodings work correctly.
1089 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
1090 is also one of the inputs.
1091 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1093 *) Don't repeatedly append PBE algorithms to table if they already exist.
1094 Sort table on each new add. This effectively makes the table read only
1095 after all algorithms are added and subsequent calls to PKCS12_pbe_add
1099 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
1101 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
1104 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
1105 access or freeing data twice (CVE-2010-0742)
1106 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
1108 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
1109 common in certificates and some applications which only call
1110 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
1114 Reduce copying into .apps and .test in makevms.com
1115 Don't try to use blank CA certificate in CA.com
1116 Allow use of C files from original directories in maketests.com
1117 [Steven M. Schweda" <sms@antinode.info>]
1119 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1121 *) When rejecting SSL/TLS records due to an incorrect version number, never
1122 update s->server with a new major version number. As of
1123 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1124 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1125 the previous behavior could result in a read attempt at NULL when
1126 receiving specific incorrect SSL/TLS records once record payload
1127 protection is active. (CVE-2010-0740)
1128 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1130 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1131 could be crashed if the relevant tables were not present (e.g. chrooted).
1132 [Tomas Hoger <thoger@redhat.com>]
1134 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1136 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1137 [Martin Olsson, Neel Mehta]
1139 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1140 accommodate for stack sorting, always a write lock!).
1143 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1144 excessive delays in the RAND_poll(): over a minute. As a workaround
1145 include a time check in the inner Heap32Next loop too.
1148 *) The code that handled flushing of data in SSL/TLS originally used the
1149 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1150 the problem outlined in PR#1949. The fix suggested there however can
1151 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1152 of Apache). So instead simplify the code to flush unconditionally.
1153 This should be fine since flushing with no data to flush is a no op.
1156 *) Handle TLS versions 2.0 and later properly and correctly use the
1157 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1158 off ancient servers have a habit of sticking around for a while...
1161 *) Modify compression code so it frees up structures without using the
1162 ex_data callbacks. This works around a problem where some applications
1163 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1164 restarting) then use compression (e.g. SSL with compression) later.
1165 This results in significant per-connection memory leaks and
1166 has caused some security issues including CVE-2008-1678 and
1170 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1171 change when encrypting or decrypting.
1174 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1175 connect and renegotiate with servers which do not support RI.
1176 Until RI is more widely deployed this option is enabled by default.
1179 *) Add "missing" ssl ctrls to clear options and mode.
1182 *) If client attempts to renegotiate and doesn't support RI respond with
1183 a no_renegotiation alert as required by RFC5746. Some renegotiating
1184 TLS clients will continue a connection gracefully when they receive
1185 the alert. Unfortunately OpenSSL mishandled this alert and would hang
1186 waiting for a server hello which it will never receive. Now we treat a
1187 received no_renegotiation alert as a fatal error. This is because
1188 applications requesting a renegotiation might well expect it to succeed
1189 and would have no code in place to handle the server denying it so the
1190 only safe thing to do is to terminate the connection.
1193 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
1194 peer supports secure renegotiation and 0 otherwise. Print out peer
1195 renegotiation support in s_client/s_server.
1198 *) Replace the highly broken and deprecated SPKAC certification method with
1199 the updated NID creation version. This should correctly handle UTF8.
1202 *) Implement RFC5746. Re-enable renegotiation but require the extension
1203 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
1204 turns out to be a bad idea. It has been replaced by
1205 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
1206 SSL_CTX_set_options(). This is really not recommended unless you
1207 know what you are doing.
1208 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
1210 *) Fixes to stateless session resumption handling. Use initial_ctx when
1211 issuing and attempting to decrypt tickets in case it has changed during
1212 servername handling. Use a non-zero length session ID when attempting
1213 stateless session resumption: this makes it possible to determine if
1214 a resumption has occurred immediately after receiving server hello
1215 (several places in OpenSSL subtly assume this) instead of later in
1219 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
1220 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
1221 fixes for a few places where the return code is not checked
1223 [Julia Lawall <julia@diku.dk>]
1225 *) Add --strict-warnings option to Configure script to include devteam
1226 warnings in other configurations.
1229 *) Add support for --libdir option and LIBDIR variable in makefiles. This
1230 makes it possible to install openssl libraries in locations which
1231 have names other than "lib", for example "/usr/lib64" which some
1233 [Steve Henson, based on patch from Jeremy Utley]
1235 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
1236 X690 8.9.12 and can produce some misleading textual output of OIDs.
1237 [Steve Henson, reported by Dan Kaminsky]
1239 *) Delete MD2 from algorithm tables. This follows the recommendation in
1240 several standards that it is not used in new applications due to
1241 several cryptographic weaknesses. For binary compatibility reasons
1242 the MD2 API is still compiled in by default.
1245 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
1249 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
1250 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
1252 [Guenter <lists@gknw.net>]
1254 *) Fix the server certificate chain building code to use X509_verify_cert(),
1255 it used to have an ad-hoc builder which was unable to cope with anything
1256 other than a simple chain.
1257 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
1259 *) Don't check self signed certificate signatures in X509_verify_cert()
1260 by default (a flag can override this): it just wastes time without
1261 adding any security. As a useful side effect self signed root CAs
1262 with non-FIPS digests are now usable in FIPS mode.
1265 *) In dtls1_process_out_of_seq_message() the check if the current message
1266 is already buffered was missing. For every new message was memory
1267 allocated, allowing an attacker to perform an denial of service attack
1268 with sending out of seq handshake messages until there is no memory
1269 left. Additionally every future messege was buffered, even if the
1270 sequence number made no sense and would be part of another handshake.
1271 So only messages with sequence numbers less than 10 in advance will be
1272 buffered. (CVE-2009-1378)
1273 [Robin Seggelmann, discovered by Daniel Mentz]
1275 *) Records are buffered if they arrive with a future epoch to be
1276 processed after finishing the corresponding handshake. There is
1277 currently no limitation to this buffer allowing an attacker to perform
1278 a DOS attack with sending records with future epochs until there is no
1279 memory left. This patch adds the pqueue_size() function to detemine
1280 the size of a buffer and limits the record buffer to 100 entries.
1282 [Robin Seggelmann, discovered by Daniel Mentz]
1284 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
1285 parent structure is freed. (CVE-2009-1379)
1288 *) Handle non-blocking I/O properly in SSL_shutdown() call.
1289 [Darryl Miles <darryl-mailinglists@netbauds.net>]
1292 [Ilya O. <vrghost@gmail.com>]
1294 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
1296 *) Disable renegotiation completely - this fixes a severe security
1297 problem (CVE-2009-3555) at the cost of breaking all
1298 renegotiation. Renegotiation can be re-enabled by setting
1299 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
1300 run-time. This is really not recommended unless you know what
1304 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
1306 *) Don't set val to NULL when freeing up structures, it is freed up by
1307 underlying code. If sizeof(void *) > sizeof(long) this can result in
1308 zeroing past the valid field. (CVE-2009-0789)
1309 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
1311 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
1312 checked correctly. This would allow some invalid signed attributes to
1313 appear to verify correctly. (CVE-2009-0591)
1314 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1316 *) Reject UniversalString and BMPString types with invalid lengths. This
1317 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
1318 a legal length. (CVE-2009-0590)
1321 *) Set S/MIME signing as the default purpose rather than setting it
1322 unconditionally. This allows applications to override it at the store
1326 *) Permit restricted recursion of ASN1 strings. This is needed in practice
1327 to handle some structures.
1330 *) Improve efficiency of mem_gets: don't search whole buffer each time
1332 [Jeremy Shapiro <jnshapir@us.ibm.com>]
1334 *) New -hex option for openssl rand.
1337 *) Print out UTF8String and NumericString when parsing ASN1.
1340 *) Support NumericString type for name components.
1343 *) Allow CC in the environment to override the automatically chosen
1344 compiler. Note that nothing is done to ensure flags work with the
1348 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
1350 *) Properly check EVP_VerifyFinal() and similar return values
1352 [Ben Laurie, Bodo Moeller, Google Security Team]
1354 *) Enable TLS extensions by default.
1357 *) Allow the CHIL engine to be loaded, whether the application is
1358 multithreaded or not. (This does not release the developer from the
1359 obligation to set up the dynamic locking callbacks.)
1360 [Sander Temme <sander@temme.net>]
1362 *) Use correct exit code if there is an error in dgst command.
1363 [Steve Henson; problem pointed out by Roland Dirlewanger]
1365 *) Tweak Configure so that you need to say "experimental-jpake" to enable
1366 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
1369 *) Add experimental JPAKE support, including demo authentication in
1370 s_client and s_server.
1373 *) Set the comparison function in v3_addr_canonize().
1374 [Rob Austein <sra@hactrn.net>]
1376 *) Add support for XMPP STARTTLS in s_client.
1377 [Philip Paeps <philip@freebsd.org>]
1379 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
1380 to ensure that even with this option, only ciphersuites in the
1381 server's preference list will be accepted. (Note that the option
1382 applies only when resuming a session, so the earlier behavior was
1383 just about the algorithm choice for symmetric cryptography.)
1386 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
1388 *) Fix NULL pointer dereference if a DTLS server received
1389 ChangeCipherSpec as first record (CVE-2009-1386).
1392 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
1393 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
1396 *) The fix in 0.9.8c that supposedly got rid of unsafe
1397 double-checked locking was incomplete for RSA blinding,
1398 addressing just one layer of what turns out to have been
1399 doubly unsafe triple-checked locking.
1401 So now fix this for real by retiring the MONT_HELPER macro
1402 in crypto/rsa/rsa_eay.c.
1404 [Bodo Moeller; problem pointed out by Marius Schilder]
1406 *) Various precautionary measures:
1408 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
1410 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
1411 (NB: This would require knowledge of the secret session ticket key
1412 to exploit, in which case you'd be SOL either way.)
1414 - Change bn_nist.c so that it will properly handle input BIGNUMs
1415 outside the expected range.
1417 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
1420 [Neel Mehta, Bodo Moeller]
1422 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
1423 the load fails. Useful for distros.
1424 [Ben Laurie and the FreeBSD team]
1426 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
1429 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
1432 *) Expand ENGINE to support engine supplied SSL client certificate functions.
1434 This work was sponsored by Logica.
1437 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
1438 keystores. Support for SSL/TLS client authentication too.
1439 Not compiled unless enable-capieng specified to Configure.
1441 This work was sponsored by Logica.
1444 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
1445 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
1446 attribute creation routines such as certifcate requests and PKCS#12
1450 Changes between 0.9.8g and 0.9.8h [28 May 2008]
1452 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
1453 handshake which could lead to a cilent crash as found using the
1454 Codenomicon TLS test suite (CVE-2008-1672)
1455 [Steve Henson, Mark Cox]
1457 *) Fix double free in TLS server name extensions which could lead to
1458 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
1461 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
1463 Clear the error queue to ensure that error entries left from
1464 older function calls do not interfere with the correct operation.
1465 [Lutz Jaenicke, Erik de Castro Lopo]
1467 *) Remove root CA certificates of commercial CAs:
1469 The OpenSSL project does not recommend any specific CA and does not
1470 have any policy with respect to including or excluding any CA.
1471 Therefore it does not make any sense to ship an arbitrary selection
1472 of root CA certificates with the OpenSSL software.
1475 *) RSA OAEP patches to fix two separate invalid memory reads.
1476 The first one involves inputs when 'lzero' is greater than
1477 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
1478 before the beginning of from). The second one involves inputs where
1479 the 'db' section contains nothing but zeroes (there is a one-byte
1480 invalid read after the end of 'db').
1481 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1483 *) Partial backport from 0.9.9-dev:
1485 Introduce bn_mul_mont (dedicated Montgomery multiplication
1486 procedure) as a candidate for BIGNUM assembler implementation.
1487 While 0.9.9-dev uses assembler for various architectures, only
1488 x86_64 is available by default here in the 0.9.8 branch, and
1489 32-bit x86 is available through a compile-time setting.
1491 To try the 32-bit x86 assembler implementation, use Configure
1492 option "enable-montasm" (which exists only for this backport).
1494 As "enable-montasm" for 32-bit x86 disclaims code stability
1495 anyway, in this constellation we activate additional code
1496 backported from 0.9.9-dev for further performance improvements,
1497 namely BN_from_montgomery_word. (To enable this otherwise,
1498 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
1500 [Andy Polyakov (backport partially by Bodo Moeller)]
1502 *) Add TLS session ticket callback. This allows an application to set
1503 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
1504 values. This is useful for key rollover for example where several key
1505 sets may exist with different names.
1508 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
1509 This was broken until now in 0.9.8 releases, such that the only way
1510 a registered ENGINE could be used (assuming it initialises
1511 successfully on the host) was to explicitly set it as the default
1512 for the relevant algorithms. This is in contradiction with 0.9.7
1513 behaviour and the documentation. With this fix, when an ENGINE is
1514 registered into a given algorithm's table of implementations, the
1515 'uptodate' flag is reset so that auto-discovery will be used next
1516 time a new context for that algorithm attempts to select an
1518 [Ian Lister (tweaked by Geoff Thorpe)]
1520 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
1521 implemention in the following ways:
1523 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
1526 Lack of BER streaming support means one pass streaming processing is
1527 only supported if data is detached: setting the streaming flag is
1528 ignored for embedded content.
1530 CMS support is disabled by default and must be explicitly enabled
1531 with the enable-cms configuration option.
1534 *) Update the GMP engine glue to do direct copies between BIGNUM and
1535 mpz_t when openssl and GMP use the same limb size. Otherwise the
1536 existing "conversion via a text string export" trick is still used.
1537 [Paul Sheer <paulsheer@gmail.com>]
1539 *) Zlib compression BIO. This is a filter BIO which compressed and
1540 uncompresses any data passed through it.
1543 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
1544 RFC3394 compatible AES key wrapping.
1547 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
1548 sets string data without copying. X509_ALGOR_set0() and
1549 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
1550 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
1551 from an X509_ATTRIBUTE structure optionally checking it occurs only
1552 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
1556 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
1557 to get the expected BN_FLG_CONSTTIME behavior.
1558 [Bodo Moeller (Google)]
1562 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
1563 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
1564 - added some more tests to do_tests.pl
1565 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
1566 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
1567 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
1568 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
1569 - various changes to netware.pl to enable gcc-cross builds on Win32
1571 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
1572 - various changes to fix missing prototype warnings
1573 - fixed x86nasm.pl to create correct asm files for NASM COFF output
1574 - added AES, WHIRLPOOL and CPUID assembler code to build files
1575 - added missing AES assembler make rules to mk1mf.pl
1576 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
1577 [Guenter Knauf <eflash@gmx.net>]
1579 *) Implement certificate status request TLS extension defined in RFC3546.
1580 A client can set the appropriate parameters and receive the encoded
1581 OCSP response via a callback. A server can query the supplied parameters
1582 and set the encoded OCSP response in the callback. Add simplified examples
1583 to s_client and s_server.
1586 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
1588 *) Fix various bugs:
1589 + Binary incompatibility of ssl_ctx_st structure
1590 + DTLS interoperation with non-compliant servers
1591 + Don't call get_session_cb() without proposed session
1592 + Fix ia64 assembler code
1593 [Andy Polyakov, Steve Henson]
1595 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
1597 *) DTLS Handshake overhaul. There were longstanding issues with
1598 OpenSSL DTLS implementation, which were making it impossible for
1599 RFC 4347 compliant client to communicate with OpenSSL server.
1600 Unfortunately just fixing these incompatibilities would "cut off"
1601 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
1602 server keeps tolerating non RFC compliant syntax. The opposite is
1603 not true, 0.9.8f client can not communicate with earlier server.
1604 This update even addresses CVE-2007-4995.
1607 *) Changes to avoid need for function casts in OpenSSL: some compilers
1608 (gcc 4.2 and later) reject their use.
1609 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
1612 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1613 RFC4507bis. The encrypted ticket format is an encrypted encoded
1614 SSL_SESSION structure, that way new session features are automatically
1617 If a client application caches session in an SSL_SESSION structure
1618 support is transparent because tickets are now stored in the encoded
1621 The SSL_CTX structure automatically generates keys for ticket
1622 protection in servers so again support should be possible
1623 with no application modification.
1625 If a client or server wishes to disable RFC4507 support then the option
1626 SSL_OP_NO_TICKET can be set.
1628 Add a TLS extension debugging callback to allow the contents of any client
1629 or server extensions to be examined.
1631 This work was sponsored by Google.
1634 *) Add initial support for TLS extensions, specifically for the server_name
1635 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1636 have new members for a host name. The SSL data structure has an
1637 additional member SSL_CTX *initial_ctx so that new sessions can be
1638 stored in that context to allow for session resumption, even after the
1639 SSL has been switched to a new SSL_CTX in reaction to a client's
1640 server_name extension.
1642 New functions (subject to change):
1644 SSL_get_servername()
1645 SSL_get_servername_type()
1648 New CTRL codes and macros (subject to change):
1650 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1651 - SSL_CTX_set_tlsext_servername_callback()
1652 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1653 - SSL_CTX_set_tlsext_servername_arg()
1654 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1656 openssl s_client has a new '-servername ...' option.
1658 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1659 '-key2 ...', '-servername_fatal' (subject to change). This allows
1660 testing the HostName extension for a specific single host name ('-cert'
1661 and '-key' remain fallbacks for handshakes without HostName
1662 negotiation). If the unrecogninzed_name alert has to be sent, this by
1663 default is a warning; it becomes fatal with the '-servername_fatal'
1666 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
1668 *) Add AES and SSE2 assembly language support to VC++ build.
1671 *) Mitigate attack on final subtraction in Montgomery reduction.
1674 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
1675 (which previously caused an internal error).
1678 *) Squeeze another 10% out of IGE mode when in != out.
1681 *) AES IGE mode speedup.
1682 [Dean Gaudet (Google)]
1684 *) Add the Korean symmetric 128-bit cipher SEED (see
1685 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
1686 add SEED ciphersuites from RFC 4162:
1688 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
1689 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
1690 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
1691 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
1693 To minimize changes between patchlevels in the OpenSSL 0.9.8
1694 series, SEED remains excluded from compilation unless OpenSSL
1695 is configured with 'enable-seed'.
1696 [KISA, Bodo Moeller]
1698 *) Mitigate branch prediction attacks, which can be practical if a
1699 single processor is shared, allowing a spy process to extract
1700 information. For detailed background information, see
1701 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
1702 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
1703 and Necessary Software Countermeasures"). The core of the change
1704 are new versions BN_div_no_branch() and
1705 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
1706 respectively, which are slower, but avoid the security-relevant
1707 conditional branches. These are automatically called by BN_div()
1708 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
1709 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
1710 remove a conditional branch.
1712 BN_FLG_CONSTTIME is the new name for the previous
1713 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
1714 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
1715 in the exponent causes BN_mod_exp_mont() to use the alternative
1716 implementation in BN_mod_exp_mont_consttime().) The old name
1717 remains as a deprecated alias.
1719 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
1720 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
1721 constant-time implementations for more than just exponentiation.
1722 Here too the old name is kept as a deprecated alias.
1724 BN_BLINDING_new() will now use BN_dup() for the modulus so that
1725 the BN_BLINDING structure gets an independent copy of the
1726 modulus. This means that the previous "BIGNUM *m" argument to
1727 BN_BLINDING_new() and to BN_BLINDING_create_param() now
1728 essentially becomes "const BIGNUM *m", although we can't actually
1729 change this in the header file before 0.9.9. It allows
1730 RSA_setup_blinding() to use BN_with_flags() on the modulus to
1731 enable BN_FLG_CONSTTIME.
1733 [Matthew D Wood (Intel Corp)]
1735 *) In the SSL/TLS server implementation, be strict about session ID
1736 context matching (which matters if an application uses a single
1737 external cache for different purposes). Previously,
1738 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
1739 set. This did ensure strict client verification, but meant that,
1740 with applications using a single external cache for quite
1741 different requirements, clients could circumvent ciphersuite
1742 restrictions for a given session ID context by starting a session
1743 in a different context.
1746 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
1747 a ciphersuite string such as "DEFAULT:RSA" cannot enable
1748 authentication-only ciphersuites.
1751 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
1752 not complete and could lead to a possible single byte overflow
1753 (CVE-2007-5135) [Ben Laurie]
1755 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
1757 *) Since AES128 and AES256 (and similarly Camellia128 and
1758 Camellia256) share a single mask bit in the logic of
1759 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
1760 kludge to work properly if AES128 is available and AES256 isn't
1761 (or if Camellia128 is available and Camellia256 isn't).
1764 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
1765 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
1766 When a point or a seed is encoded in a BIT STRING, we need to
1767 prevent the removal of trailing zero bits to get the proper DER
1768 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
1769 of a NamedBitList, for which trailing 0 bits need to be removed.)
1772 *) Have SSL/TLS server implementation tolerate "mismatched" record
1773 protocol version while receiving ClientHello even if the
1774 ClientHello is fragmented. (The server can't insist on the
1775 particular protocol version it has chosen before the ServerHello
1776 message has informed the client about his choice.)
1779 *) Add RFC 3779 support.
1780 [Rob Austein for ARIN, Ben Laurie]
1782 *) Load error codes if they are not already present instead of using a
1783 static variable. This allows them to be cleanly unloaded and reloaded.
1784 Improve header file function name parsing.
1787 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
1788 or CAPABILITY handshake as required by RFCs.
1791 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
1793 *) Introduce limits to prevent malicious keys being able to
1794 cause a denial of service. (CVE-2006-2940)
1795 [Steve Henson, Bodo Moeller]
1797 *) Fix ASN.1 parsing of certain invalid structures that can result
1798 in a denial of service. (CVE-2006-2937) [Steve Henson]
1800 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
1801 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
1803 *) Fix SSL client code which could crash if connecting to a
1804 malicious SSLv2 server. (CVE-2006-4343)
1805 [Tavis Ormandy and Will Drewry, Google Security Team]
1807 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
1808 match only those. Before that, "AES256-SHA" would be interpreted
1809 as a pattern and match "AES128-SHA" too (since AES128-SHA got
1810 the same strength classification in 0.9.7h) as we currently only
1811 have a single AES bit in the ciphersuite description bitmap.
1812 That change, however, also applied to ciphersuite strings such as
1813 "RC4-MD5" that intentionally matched multiple ciphersuites --
1814 namely, SSL 2.0 ciphersuites in addition to the more common ones
1815 from SSL 3.0/TLS 1.0.
1817 So we change the selection algorithm again: Naming an explicit
1818 ciphersuite selects this one ciphersuite, and any other similar
1819 ciphersuite (same bitmap) from *other* protocol versions.
1820 Thus, "RC4-MD5" again will properly select both the SSL 2.0
1821 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
1823 Since SSL 2.0 does not have any ciphersuites for which the
1824 128/256 bit distinction would be relevant, this works for now.
1825 The proper fix will be to use different bits for AES128 and
1826 AES256, which would have avoided the problems from the beginning;
1827 however, bits are scarce, so we can only do this in a new release
1828 (not just a patchlevel) when we can change the SSL_CIPHER
1829 definition to split the single 'unsigned long mask' bitmap into
1830 multiple values to extend the available space.
1834 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
1836 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
1837 (CVE-2006-4339) [Ben Laurie and Google Security Team]
1839 *) Add AES IGE and biIGE modes.
1842 *) Change the Unix randomness entropy gathering to use poll() when
1843 possible instead of select(), since the latter has some
1844 undesirable limitations.
1845 [Darryl Miles via Richard Levitte and Bodo Moeller]
1847 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
1848 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
1849 cannot be implicitly activated as part of, e.g., the "AES" alias.
1850 However, please upgrade to OpenSSL 0.9.9[-dev] for
1851 non-experimental use of the ECC ciphersuites to get TLS extension
1852 support, which is required for curve and point format negotiation
1853 to avoid potential handshake problems.
1856 *) Disable rogue ciphersuites:
1858 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
1859 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
1860 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
1862 The latter two were purportedly from
1863 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
1866 Also deactivate the remaining ciphersuites from
1867 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
1868 unofficial, and the ID has long expired.
1871 *) Fix RSA blinding Heisenbug (problems sometimes occured on
1872 dual-core machines) and other potential thread-safety issues.
1875 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
1876 versions), which is now available for royalty-free use
1877 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
1878 Also, add Camellia TLS ciphersuites from RFC 4132.
1880 To minimize changes between patchlevels in the OpenSSL 0.9.8
1881 series, Camellia remains excluded from compilation unless OpenSSL
1882 is configured with 'enable-camellia'.
1885 *) Disable the padding bug check when compression is in use. The padding
1886 bug check assumes the first packet is of even length, this is not
1887 necessarily true if compresssion is enabled and can result in false
1888 positives causing handshake failure. The actual bug test is ancient
1889 code so it is hoped that implementations will either have fixed it by
1890 now or any which still have the bug do not support compression.
1893 Changes between 0.9.8a and 0.9.8b [04 May 2006]
1895 *) When applying a cipher rule check to see if string match is an explicit
1896 cipher suite and only match that one cipher suite if it is.
1899 *) Link in manifests for VC++ if needed.
1900 [Austin Ziegler <halostatue@gmail.com>]
1902 *) Update support for ECC-based TLS ciphersuites according to
1903 draft-ietf-tls-ecc-12.txt with proposed changes (but without
1904 TLS extensions, which are supported starting with the 0.9.9
1905 branch, not in the OpenSSL 0.9.8 branch).
1908 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
1909 opaque EVP_CIPHER_CTX handling.
1912 *) Fixes and enhancements to zlib compression code. We now only use
1913 "zlib1.dll" and use the default __cdecl calling convention on Win32
1914 to conform with the standards mentioned here:
1915 http://www.zlib.net/DLL_FAQ.txt
1916 Static zlib linking now works on Windows and the new --with-zlib-include
1917 --with-zlib-lib options to Configure can be used to supply the location
1918 of the headers and library. Gracefully handle case where zlib library
1922 *) Several fixes and enhancements to the OID generation code. The old code
1923 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
1924 handle numbers larger than ULONG_MAX, truncated printing and had a
1925 non standard OBJ_obj2txt() behaviour.
1928 *) Add support for building of engines under engine/ as shared libraries
1929 under VC++ build system.
1932 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
1933 Hopefully, we will not see any false combination of paths any more.
1936 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
1938 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1939 (part of SSL_OP_ALL). This option used to disable the
1940 countermeasure against man-in-the-middle protocol-version
1941 rollback in the SSL 2.0 server implementation, which is a bad
1942 idea. (CVE-2005-2969)
1944 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
1945 for Information Security, National Institute of Advanced Industrial
1946 Science and Technology [AIST], Japan)]
1948 *) Add two function to clear and return the verify parameter flags.
1951 *) Keep cipherlists sorted in the source instead of sorting them at
1952 runtime, thus removing the need for a lock.
1955 *) Avoid some small subgroup attacks in Diffie-Hellman.
1956 [Nick Mathewson and Ben Laurie]
1958 *) Add functions for well-known primes.
1961 *) Extended Windows CE support.
1962 [Satoshi Nakamura and Andy Polyakov]
1964 *) Initialize SSL_METHOD structures at compile time instead of during
1965 runtime, thus removing the need for a lock.
1968 *) Make PKCS7_decrypt() work even if no certificate is supplied by
1969 attempting to decrypt each encrypted key in turn. Add support to
1973 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
1975 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
1978 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
1981 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
1982 key into the same file any more.
1985 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
1988 *) Add -utf8 command line and config file option to 'ca'.
1989 [Stefan <stf@udoma.org]
1991 *) Removed the macro des_crypt(), as it seems to conflict with some
1992 libraries. Use DES_crypt().
1995 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
1996 involves renaming the source and generated shared-libs for
1997 both. The engines will accept the corrected or legacy ids
1998 ('ncipher' and '4758_cca' respectively) when binding. NB,
1999 this only applies when building 'shared'.
2000 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2002 *) Add attribute functions to EVP_PKEY structure. Modify
2003 PKCS12_create() to recognize a CSP name attribute and
2004 use it. Make -CSP option work again in pkcs12 utility.
2007 *) Add new functionality to the bn blinding code:
2008 - automatic re-creation of the BN_BLINDING parameters after
2009 a fixed number of uses (currently 32)
2010 - add new function for parameter creation
2011 - introduce flags to control the update behaviour of the
2012 BN_BLINDING parameters
2013 - hide BN_BLINDING structure
2014 Add a second BN_BLINDING slot to the RSA structure to improve
2015 performance when a single RSA object is shared among several
2019 *) Add support for DTLS.
2020 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2022 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2023 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2026 *) Remove buggy and incompletet DH cert support from
2027 ssl/ssl_rsa.c and ssl/s3_both.c
2030 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2031 the apps/openssl applications.
2034 *) Compile clean with "-Wall -Wmissing-prototypes
2035 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2036 DEBUG_SAFESTACK must also be set.
2039 *) Change ./Configure so that certain algorithms can be disabled by default.
2040 The new counterpiece to "no-xxx" is "enable-xxx".
2042 The patented RC5 and MDC2 algorithms will now be disabled unless
2043 "enable-rc5" and "enable-mdc2", respectively, are specified.
2045 (IDEA remains enabled despite being patented. This is because IDEA
2046 is frequently required for interoperability, and there is no license
2047 fee for non-commercial use. As before, "no-idea" can be used to
2048 avoid this algorithm.)
2052 *) Add processing of proxy certificates (see RFC 3820). This work was
2053 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2054 EGEE (Enabling Grids for E-science in Europe).
2057 *) RC4 performance overhaul on modern architectures/implementations, such
2058 as Intel P4, IA-64 and AMD64.
2061 *) New utility extract-section.pl. This can be used specify an alternative
2062 section number in a pod file instead of having to treat each file as
2063 a separate case in Makefile. This can be done by adding two lines to the
2066 =for comment openssl_section:XXX
2068 The blank line is mandatory.
2072 *) New arguments -certform, -keyform and -pass for s_client and s_server
2073 to allow alternative format key and certificate files and passphrase
2077 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2078 update associated structures and add various utility functions.
2080 Add new policy related verify parameters, include policy checking in
2081 standard verify code. Enhance 'smime' application with extra parameters
2082 to support policy checking and print out.
2085 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2086 Nehemiah processors. These extensions support AES encryption in hardware
2087 as well as RNG (though RNG support is currently disabled).
2088 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2090 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2093 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2094 [Andy Polyakov and a number of other people]
2096 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2097 implementation contributed by IBM.
2098 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2100 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2101 exponent rather than 'unsigned long'. There is a corresponding change to
2102 the new 'rsa_keygen' element of the RSA_METHOD structure.
2103 [Jelte Jansen, Geoff Thorpe]
2105 *) Functionality for creating the initial serial number file is now
2106 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2108 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2109 number file to 1, which is bound to cause problems. To avoid
2110 the problems while respecting compatibility between different 0.9.7
2111 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2112 CA.pl for serial number initialization. With the new release 0.9.8,
2113 we can fix the problem directly in the 'ca' utility.)
2116 *) Reduced header interdepencies by declaring more opaque objects in
2117 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2118 give fewer recursive includes, which could break lazy source code - so
2119 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2120 developers should define this symbol when building and using openssl to
2121 ensure they track the recommended behaviour, interfaces, [etc], but
2122 backwards-compatible behaviour prevails when this isn't defined.
2125 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2128 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2129 This will generate a random key of the appropriate length based on the
2130 cipher context. The EVP_CIPHER can provide its own random key generation
2131 routine to support keys of a specific form. This is used in the des and
2132 3des routines to generate a key of the correct parity. Update S/MIME
2133 code to use new functions and hence generate correct parity DES keys.
2134 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2135 valid (weak or incorrect parity).
2138 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2139 as looking them up. This is useful when the verified structure may contain
2140 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2141 present unless the new PKCS7_NO_CRL flag is asserted.
2144 *) Extend ASN1 oid configuration module. It now additionally accepts the
2147 shortName = some long name, 1.2.3.4
2150 *) Reimplemented the BN_CTX implementation. There is now no more static
2151 limitation on the number of variables it can handle nor the depth of the
2152 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2153 information can now expand as required, and rather than having a single
2154 static array of bignums, BN_CTX now uses a linked-list of such arrays
2155 allowing it to expand on demand whilst maintaining the usefulness of
2156 BN_CTX's "bundling".
2159 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2160 to allow all RSA operations to function using a single BN_CTX.
2163 *) Preliminary support for certificate policy evaluation and checking. This
2164 is initially intended to pass the tests outlined in "Conformance Testing
2165 of Relying Party Client Certificate Path Processing Logic" v1.07.
2168 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2169 remained unused and not that useful. A variety of other little bignum
2170 tweaks and fixes have also been made continuing on from the audit (see
2174 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2175 associated ASN1, EVP and SSL functions and old ASN1 macros.
2178 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2179 and this should never fail. So the return value from the use of
2180 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2181 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2184 *) BN_CTX_get() should return zero-valued bignums, providing the same
2185 initialised value as BN_new().
2186 [Geoff Thorpe, suggested by Ulf Möller]
2188 *) Support for inhibitAnyPolicy certificate extension.
2191 *) An audit of the BIGNUM code is underway, for which debugging code is
2192 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
2193 is considered valid when processing BIGNUMs, and causes execution to
2194 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
2195 further steps are taken to deliberately pollute unused data in BIGNUM
2196 structures to try and expose faulty code further on. For now, openssl will
2197 (in its default mode of operation) continue to tolerate the inconsistent
2198 forms that it has tolerated in the past, but authors and packagers should
2199 consider trying openssl and their own applications when compiled with
2200 these debugging symbols defined. It will help highlight potential bugs in
2201 their own code, and will improve the test coverage for OpenSSL itself. At
2202 some point, these tighter rules will become openssl's default to improve
2203 maintainability, though the assert()s and other overheads will remain only
2204 in debugging configurations. See bn.h for more details.
2205 [Geoff Thorpe, Nils Larsch, Ulf Möller]
2207 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
2208 that can only be obtained through BN_CTX_new() (which implicitly
2209 initialises it). The presence of this function only made it possible
2210 to overwrite an existing structure (and cause memory leaks).
2213 *) Because of the callback-based approach for implementing LHASH as a
2214 template type, lh_insert() adds opaque objects to hash-tables and
2215 lh_doall() or lh_doall_arg() are typically used with a destructor callback
2216 to clean up those corresponding objects before destroying the hash table
2217 (and losing the object pointers). So some over-zealous constifications in
2218 LHASH have been relaxed so that lh_insert() does not take (nor store) the
2219 objects as "const" and the lh_doall[_arg] callback wrappers are not
2220 prototyped to have "const" restrictions on the object pointers they are
2221 given (and so aren't required to cast them away any more).
2224 *) The tmdiff.h API was so ugly and minimal that our own timing utility
2225 (speed) prefers to use its own implementation. The two implementations
2226 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
2227 its object type properly exposed (MS_TM) instead of casting to/from "char
2228 *". This may still change yet if someone realises MS_TM and "ms_time_***"
2229 aren't necessarily the greatest nomenclatures - but this is what was used
2230 internally to the implementation so I've used that for now.
2233 *) Ensure that deprecated functions do not get compiled when
2234 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
2235 the self-tests were still using deprecated key-generation functions so
2236 these have been updated also.
2239 *) Reorganise PKCS#7 code to separate the digest location functionality
2240 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
2241 New function PKCS7_set_digest() to set the digest type for PKCS#7
2242 digestedData type. Add additional code to correctly generate the
2243 digestedData type and add support for this type in PKCS7 initialization
2247 *) New function PKCS7_set0_type_other() this initializes a PKCS7
2248 structure of type "other".
2251 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
2252 sure the loop does correctly stop and breaking ("division by zero")
2253 modulus operations are not performed. The (pre-generated) prime
2254 table crypto/bn/bn_prime.h was already correct, but it could not be
2255 re-generated on some platforms because of the "division by zero"
2256 situation in the script.
2257 [Ralf S. Engelschall]
2259 *) Update support for ECC-based TLS ciphersuites according to
2260 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
2261 SHA-1 now is only used for "small" curves (where the
2262 representation of a field element takes up to 24 bytes); for
2263 larger curves, the field element resulting from ECDH is directly
2264 used as premaster secret.
2265 [Douglas Stebila (Sun Microsystems Laboratories)]
2267 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
2268 curve secp160r1 to the tests.
2269 [Douglas Stebila (Sun Microsystems Laboratories)]
2271 *) Add the possibility to load symbols globally with DSO.
2272 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
2274 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
2275 control of the error stack.
2278 *) Add support for STORE in ENGINE.
2281 *) Add the STORE type. The intention is to provide a common interface
2282 to certificate and key stores, be they simple file-based stores, or
2283 HSM-type store, or LDAP stores, or...
2284 NOTE: The code is currently UNTESTED and isn't really used anywhere.
2287 *) Add a generic structure called OPENSSL_ITEM. This can be used to
2288 pass a list of arguments to any function as well as provide a way
2289 for a function to pass data back to the caller.
2292 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
2293 works like BUF_strdup() but can be used to duplicate a portion of
2294 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
2298 *) Add the function sk_find_ex() which works like sk_find(), but will
2299 return an index to an element even if an exact match couldn't be
2300 found. The index is guaranteed to point at the element where the
2301 searched-for key would be inserted to preserve sorting order.
2304 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
2305 takes an extra flags argument for optional functionality. Currently,
2306 the following flags are defined:
2308 OBJ_BSEARCH_VALUE_ON_NOMATCH
2309 This one gets OBJ_bsearch_ex() to return a pointer to the first
2310 element where the comparing function returns a negative or zero
2313 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
2314 This one gets OBJ_bsearch_ex() to return a pointer to the first
2315 element where the comparing function returns zero. This is useful
2316 if there are more than one element where the comparing function
2320 *) Make it possible to create self-signed certificates with 'openssl ca'
2321 in such a way that the self-signed certificate becomes part of the
2322 CA database and uses the same mechanisms for serial number generation
2323 as all other certificate signing. The new flag '-selfsign' enables
2324 this functionality. Adapt CA.sh and CA.pl.in.
2327 *) Add functionality to check the public key of a certificate request
2328 against a given private. This is useful to check that a certificate
2329 request can be signed by that key (self-signing).
2332 *) Make it possible to have multiple active certificates with the same
2333 subject in the CA index file. This is done only if the keyword
2334 'unique_subject' is set to 'no' in the main CA section (default
2335 if 'CA_default') of the configuration file. The value is saved
2336 with the database itself in a separate index attribute file,
2337 named like the index file with '.attr' appended to the name.
2340 *) Generate muti valued AVAs using '+' notation in config files for
2344 *) Support for nameConstraints certificate extension.
2347 *) Support for policyConstraints certificate extension.
2350 *) Support for policyMappings certificate extension.
2353 *) Make sure the default DSA_METHOD implementation only uses its
2354 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
2355 and change its own handlers to be NULL so as to remove unnecessary
2356 indirection. This lets alternative implementations fallback to the
2357 default implementation more easily.
2360 *) Support for directoryName in GeneralName related extensions
2364 *) Make it possible to link applications using Makefile.shared.
2365 Make that possible even when linking against static libraries!
2368 *) Support for single pass processing for S/MIME signing. This now
2369 means that S/MIME signing can be done from a pipe, in addition
2370 cleartext signing (multipart/signed type) is effectively streaming
2371 and the signed data does not need to be all held in memory.
2373 This is done with a new flag PKCS7_STREAM. When this flag is set
2374 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
2375 is done after the data is output (and digests calculated) in
2376 SMIME_write_PKCS7().
2379 *) Add full support for -rpath/-R, both in shared libraries and
2380 applications, at least on the platforms where it's known how
2384 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
2385 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
2386 will now compute a table of multiples of the generator that
2387 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
2388 faster (notably in the case of a single point multiplication,
2389 scalar * generator).
2390 [Nils Larsch, Bodo Moeller]
2392 *) IPv6 support for certificate extensions. The various extensions
2393 which use the IP:a.b.c.d can now take IPv6 addresses using the
2394 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
2398 *) Added an ENGINE that implements RSA by performing private key
2399 exponentiations with the GMP library. The conversions to and from
2400 GMP's mpz_t format aren't optimised nor are any montgomery forms
2401 cached, and on x86 it appears OpenSSL's own performance has caught up.
2402 However there are likely to be other architectures where GMP could
2403 provide a boost. This ENGINE is not built in by default, but it can be
2404 specified at Configure time and should be accompanied by the necessary
2405 linker additions, eg;
2406 ./config -DOPENSSL_USE_GMP -lgmp
2409 *) "openssl engine" will not display ENGINE/DSO load failure errors when
2410 testing availability of engines with "-t" - the old behaviour is
2411 produced by increasing the feature's verbosity with "-tt".
2414 *) ECDSA routines: under certain error conditions uninitialized BN objects
2415 could be freed. Solution: make sure initialization is performed early
2416 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
2420 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
2421 and DH_METHOD (eg. by ENGINE implementations) to override the normal
2422 software implementations. For DSA and DH, parameter generation can
2423 also be overriden by providing the appropriate method callbacks.
2426 *) Change the "progress" mechanism used in key-generation and
2427 primality testing to functions that take a new BN_GENCB pointer in
2428 place of callback/argument pairs. The new API functions have "_ex"
2429 postfixes and the older functions are reimplemented as wrappers for
2430 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
2431 declarations of the old functions to help (graceful) attempts to
2432 migrate to the new functions. Also, the new key-generation API
2433 functions operate on a caller-supplied key-structure and return
2434 success/failure rather than returning a key or NULL - this is to
2435 help make "keygen" another member function of RSA_METHOD etc.
2437 Example for using the new callback interface:
2439 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
2443 BN_GENCB_set(&my_cb, my_callback, my_arg);
2445 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
2446 /* For the meaning of a, b in calls to my_callback(), see the
2447 * documentation of the function that calls the callback.
2448 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
2449 * my_callback should return 1 if it wants BN_is_prime_ex()
2450 * to continue, or 0 to stop.
2455 *) Change the ZLIB compression method to be stateful, and make it
2456 available to TLS with the number defined in
2457 draft-ietf-tls-compression-04.txt.
2460 *) Add the ASN.1 structures and functions for CertificatePair, which
2461 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
2463 CertificatePair ::= SEQUENCE {
2464 forward [0] Certificate OPTIONAL,
2465 reverse [1] Certificate OPTIONAL,
2466 -- at least one of the pair shall be present -- }
2468 Also implement the PEM functions to read and write certificate
2469 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
2471 This needed to be defined, mostly for the sake of the LDAP
2472 attribute crossCertificatePair, but may prove useful elsewhere as
2476 *) Make it possible to inhibit symlinking of shared libraries in
2477 Makefile.shared, for Cygwin's sake.
2480 *) Extend the BIGNUM API by creating a function
2481 void BN_set_negative(BIGNUM *a, int neg);
2482 and a macro that behave like
2483 int BN_is_negative(const BIGNUM *a);
2485 to avoid the need to access 'a->neg' directly in applications.
2488 *) Implement fast modular reduction for pseudo-Mersenne primes
2489 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
2490 EC_GROUP_new_curve_GFp() will now automatically use this
2492 [Nils Larsch <nla@trustcenter.de>]
2494 *) Add new lock type (CRYPTO_LOCK_BN).
2497 *) Change the ENGINE framework to automatically load engines
2498 dynamically from specific directories unless they could be
2499 found to already be built in or loaded. Move all the
2500 current engines except for the cryptodev one to a new
2502 The engines in engines/ are built as shared libraries if
2503 the "shared" options was given to ./Configure or ./config.
2504 Otherwise, they are inserted in libcrypto.a.
2505 /usr/local/ssl/engines is the default directory for dynamic
2506 engines, but that can be overriden at configure time through
2507 the usual use of --prefix and/or --openssldir, and at run
2508 time with the environment variable OPENSSL_ENGINES.
2509 [Geoff Thorpe and Richard Levitte]
2511 *) Add Makefile.shared, a helper makefile to build shared
2512 libraries. Addapt Makefile.org.
2515 *) Add version info to Win32 DLLs.
2516 [Peter 'Luna' Runestig" <peter@runestig.com>]
2518 *) Add new 'medium level' PKCS#12 API. Certificates and keys
2519 can be added using this API to created arbitrary PKCS#12
2520 files while avoiding the low level API.
2522 New options to PKCS12_create(), key or cert can be NULL and
2523 will then be omitted from the output file. The encryption
2524 algorithm NIDs can be set to -1 for no encryption, the mac
2525 iteration count can be set to 0 to omit the mac.
2527 Enhance pkcs12 utility by making the -nokeys and -nocerts
2528 options work when creating a PKCS#12 file. New option -nomac
2529 to omit the mac, NONE can be set for an encryption algorithm.
2530 New code is modified to use the enhanced PKCS12_create()
2531 instead of the low level API.
2534 *) Extend ASN1 encoder to support indefinite length constructed
2535 encoding. This can output sequences tags and octet strings in
2536 this form. Modify pk7_asn1.c to support indefinite length
2537 encoding. This is experimental and needs additional code to
2538 be useful, such as an ASN1 bio and some enhanced streaming
2541 Extend template encode functionality so that tagging is passed
2542 down to the template encoder.
2545 *) Let 'openssl req' fail if an argument to '-newkey' is not
2546 recognized instead of using RSA as a default.
2549 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
2550 As these are not official, they are not included in "ALL";
2551 the "ECCdraft" ciphersuite group alias can be used to select them.
2552 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
2554 *) Add ECDH engine support.
2555 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
2557 *) Add ECDH in new directory crypto/ecdh/.
2558 [Douglas Stebila (Sun Microsystems Laboratories)]
2560 *) Let BN_rand_range() abort with an error after 100 iterations
2561 without success (which indicates a broken PRNG).
2564 *) Change BN_mod_sqrt() so that it verifies that the input value
2565 is really the square of the return value. (Previously,
2566 BN_mod_sqrt would show GIGO behaviour.)
2569 *) Add named elliptic curves over binary fields from X9.62, SECG,
2570 and WAP/WTLS; add OIDs that were still missing.
2572 [Sheueling Chang Shantz and Douglas Stebila
2573 (Sun Microsystems Laboratories)]
2575 *) Extend the EC library for elliptic curves over binary fields
2576 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
2579 EC_GF2m_simple_method
2583 EC_GROUP_new_curve_GF2m
2584 EC_GROUP_set_curve_GF2m
2585 EC_GROUP_get_curve_GF2m
2586 EC_POINT_set_affine_coordinates_GF2m
2587 EC_POINT_get_affine_coordinates_GF2m
2588 EC_POINT_set_compressed_coordinates_GF2m
2590 Point compression for binary fields is disabled by default for
2591 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
2594 As binary polynomials are represented as BIGNUMs, various members
2595 of the EC_GROUP and EC_POINT data structures can be shared
2596 between the implementations for prime fields and binary fields;
2597 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
2598 are essentially identical to their ..._GFp counterparts.
2599 (For simplicity, the '..._GFp' prefix has been dropped from
2600 various internal method names.)
2602 An internal 'field_div' method (similar to 'field_mul' and
2603 'field_sqr') has been added; this is used only for binary fields.
2605 [Sheueling Chang Shantz and Douglas Stebila
2606 (Sun Microsystems Laboratories)]
2608 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
2609 through methods ('mul', 'precompute_mult').
2611 The generic implementations (now internally called 'ec_wNAF_mul'
2612 and 'ec_wNAF_precomputed_mult') remain the default if these
2613 methods are undefined.
2615 [Sheueling Chang Shantz and Douglas Stebila
2616 (Sun Microsystems Laboratories)]
2618 *) New function EC_GROUP_get_degree, which is defined through
2619 EC_METHOD. For curves over prime fields, this returns the bit
2620 length of the modulus.
2622 [Sheueling Chang Shantz and Douglas Stebila
2623 (Sun Microsystems Laboratories)]
2625 *) New functions EC_GROUP_dup, EC_POINT_dup.
2626 (These simply call ..._new and ..._copy).
2628 [Sheueling Chang Shantz and Douglas Stebila
2629 (Sun Microsystems Laboratories)]
2631 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
2632 Polynomials are represented as BIGNUMs (where the sign bit is not
2633 used) in the following functions [macros]:
2636 BN_GF2m_sub [= BN_GF2m_add]
2637 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
2638 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
2639 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
2641 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
2642 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
2643 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
2644 BN_GF2m_cmp [= BN_ucmp]
2646 (Note that only the 'mod' functions are actually for fields GF(2^m).
2647 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
2649 For some functions, an the irreducible polynomial defining a
2650 field can be given as an 'unsigned int[]' with strictly
2651 decreasing elements giving the indices of those bits that are set;
2652 i.e., p[] represents the polynomial
2653 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
2655 p[0] > p[1] > ... > p[k] = 0.
2656 This applies to the following functions:
2661 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
2662 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
2664 BN_GF2m_mod_sqrt_arr
2665 BN_GF2m_mod_solve_quad_arr
2669 Conversion can be performed by the following functions:
2674 bntest.c has additional tests for binary polynomial arithmetic.
2676 Two implementations for BN_GF2m_mod_div() are available.
2677 The default algorithm simply uses BN_GF2m_mod_inv() and
2678 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
2679 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
2680 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
2682 [Sheueling Chang Shantz and Douglas Stebila
2683 (Sun Microsystems Laboratories)]
2685 *) Add new error code 'ERR_R_DISABLED' that can be used when some
2686 functionality is disabled at compile-time.
2687 [Douglas Stebila <douglas.stebila@sun.com>]
2689 *) Change default behaviour of 'openssl asn1parse' so that more
2690 information is visible when viewing, e.g., a certificate:
2692 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
2693 mode the content of non-printable OCTET STRINGs is output in a
2694 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
2695 avoid the appearance of a printable string.
2696 [Nils Larsch <nla@trustcenter.de>]
2698 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
2700 EC_GROUP_set_asn1_flag()
2701 EC_GROUP_get_asn1_flag()
2702 EC_GROUP_set_point_conversion_form()
2703 EC_GROUP_get_point_conversion_form()
2704 These control ASN1 encoding details:
2705 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
2706 has been set to OPENSSL_EC_NAMED_CURVE.
2707 - Points are encoded in uncompressed form by default; options for
2708 asn1_for are as for point2oct, namely
2709 POINT_CONVERSION_COMPRESSED
2710 POINT_CONVERSION_UNCOMPRESSED
2711 POINT_CONVERSION_HYBRID
2713 Also add 'seed' and 'seed_len' members to EC_GROUP with access
2716 EC_GROUP_get0_seed()
2717 EC_GROUP_get_seed_len()
2718 This is used only for ASN1 purposes (so far).
2719 [Nils Larsch <nla@trustcenter.de>]
2721 *) Add 'field_type' member to EC_METHOD, which holds the NID
2722 of the appropriate field type OID. The new function
2723 EC_METHOD_get_field_type() returns this value.
2724 [Nils Larsch <nla@trustcenter.de>]
2729 EC_POINT_point2hex()
2730 EC_POINT_hex2point()
2731 providing useful interfaces to EC_POINT_point2oct() and
2732 EC_POINT_oct2point().
2733 [Nils Larsch <nla@trustcenter.de>]
2735 *) Change internals of the EC library so that the functions
2736 EC_GROUP_set_generator()
2737 EC_GROUP_get_generator()
2738 EC_GROUP_get_order()
2739 EC_GROUP_get_cofactor()
2740 are implemented directly in crypto/ec/ec_lib.c and not dispatched
2741 to methods, which would lead to unnecessary code duplication when
2742 adding different types of curves.
2743 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
2745 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
2746 arithmetic, and such that modified wNAFs are generated
2747 (which avoid length expansion in many cases).
2750 *) Add a function EC_GROUP_check_discriminant() (defined via
2751 EC_METHOD) that verifies that the curve discriminant is non-zero.
2753 Add a function EC_GROUP_check() that makes some sanity tests
2754 on a EC_GROUP, its generator and order. This includes
2755 EC_GROUP_check_discriminant().
2756 [Nils Larsch <nla@trustcenter.de>]
2758 *) Add ECDSA in new directory crypto/ecdsa/.
2760 Add applications 'openssl ecparam' and 'openssl ecdsa'
2761 (these are based on 'openssl dsaparam' and 'openssl dsa').
2763 ECDSA support is also included in various other files across the
2764 library. Most notably,
2765 - 'openssl req' now has a '-newkey ecdsa:file' option;
2766 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
2767 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
2768 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
2769 them suitable for ECDSA where domain parameters must be
2770 extracted before the specific public key;
2771 - ECDSA engine support has been added.
2772 [Nils Larsch <nla@trustcenter.de>]
2774 *) Include some named elliptic curves, and add OIDs from X9.62,
2775 SECG, and WAP/WTLS. Each curve can be obtained from the new
2777 EC_GROUP_new_by_curve_name(),
2778 and the list of available named curves can be obtained with
2779 EC_get_builtin_curves().
2780 Also add a 'curve_name' member to EC_GROUP objects, which can be
2782 EC_GROUP_set_curve_name()
2783 EC_GROUP_get_curve_name()
2784 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
2786 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2787 was actually never needed) and in BN_mul(). The removal in BN_mul()
2788 required a small change in bn_mul_part_recursive() and the addition
2789 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2790 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2791 bn_sub_words() and bn_add_words() except they take arrays with
2795 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
2797 *) Cleanse PEM buffers before freeing them since they may contain
2799 [Benjamin Bennett <ben@psc.edu>]
2801 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2802 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2803 authentication-only ciphersuites.
2806 *) Since AES128 and AES256 share a single mask bit in the logic of
2807 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2808 kludge to work properly if AES128 is available and AES256 isn't.
2811 *) Expand security boundary to match 1.1.1 module.
2814 *) Remove redundant features: hash file source, editing of test vectors
2815 modify fipsld to use external fips_premain.c signature.
2818 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
2819 run algorithm test programs.
2822 *) Make algorithm test programs more tolerant of whitespace.
2825 *) Have SSL/TLS server implementation tolerate "mismatched" record
2826 protocol version while receiving ClientHello even if the
2827 ClientHello is fragmented. (The server can't insist on the
2828 particular protocol version it has chosen before the ServerHello
2829 message has informed the client about his choice.)
2832 *) Load error codes if they are not already present instead of using a
2833 static variable. This allows them to be cleanly unloaded and reloaded.
2836 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
2838 *) Introduce limits to prevent malicious keys being able to
2839 cause a denial of service. (CVE-2006-2940)
2840 [Steve Henson, Bodo Moeller]
2842 *) Fix ASN.1 parsing of certain invalid structures that can result
2843 in a denial of service. (CVE-2006-2937) [Steve Henson]
2845 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2846 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2848 *) Fix SSL client code which could crash if connecting to a
2849 malicious SSLv2 server. (CVE-2006-4343)
2850 [Tavis Ormandy and Will Drewry, Google Security Team]
2852 *) Change ciphersuite string processing so that an explicit
2853 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
2854 will no longer include "AES128-SHA"), and any other similar
2855 ciphersuite (same bitmap) from *other* protocol versions (so that
2856 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
2857 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
2858 changes from 0.9.8b and 0.9.8d.
2861 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
2863 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2864 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2866 *) Change the Unix randomness entropy gathering to use poll() when
2867 possible instead of select(), since the latter has some
2868 undesirable limitations.
2869 [Darryl Miles via Richard Levitte and Bodo Moeller]
2871 *) Disable rogue ciphersuites:
2873 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2874 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2875 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2877 The latter two were purportedly from
2878 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2881 Also deactive the remaining ciphersuites from
2882 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2883 unofficial, and the ID has long expired.
2886 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2887 dual-core machines) and other potential thread-safety issues.
2890 Changes between 0.9.7i and 0.9.7j [04 May 2006]
2892 *) Adapt fipsld and the build system to link against the validated FIPS
2893 module in FIPS mode.
2896 *) Fixes for VC++ 2005 build under Windows.
2899 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
2900 from a Windows bash shell such as MSYS. It is autodetected from the
2901 "config" script when run from a VC++ environment. Modify standard VC++
2902 build to use fipscanister.o from the GNU make build.
2905 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
2907 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
2908 The value now differs depending on if you build for FIPS or not.
2909 BEWARE! A program linked with a shared FIPSed libcrypto can't be
2910 safely run with a non-FIPSed libcrypto, as it may crash because of
2911 the difference induced by this change.
2914 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
2916 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2917 (part of SSL_OP_ALL). This option used to disable the
2918 countermeasure against man-in-the-middle protocol-version
2919 rollback in the SSL 2.0 server implementation, which is a bad
2920 idea. (CVE-2005-2969)
2922 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2923 for Information Security, National Institute of Advanced Industrial
2924 Science and Technology [AIST], Japan)]
2926 *) Minimal support for X9.31 signatures and PSS padding modes. This is
2927 mainly for FIPS compliance and not fully integrated at this stage.
2930 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
2931 the exponentiation using a fixed-length exponent. (Otherwise,
2932 the information leaked through timing could expose the secret key
2933 after many signatures; cf. Bleichenbacher's attack on DSA with
2937 *) Make a new fixed-window mod_exp implementation the default for
2938 RSA, DSA, and DH private-key operations so that the sequence of
2939 squares and multiplies and the memory access pattern are
2940 independent of the particular secret key. This will mitigate
2941 cache-timing and potential related attacks.
2943 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
2944 and this is automatically used by BN_mod_exp_mont() if the new flag
2945 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
2946 will use this BN flag for private exponents unless the flag
2947 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
2948 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
2950 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
2952 *) Change the client implementation for SSLv23_method() and
2953 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
2954 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
2955 (Previously, the SSL 2.0 backwards compatible Client Hello
2956 message format would be used even with SSL_OP_NO_SSLv2.)
2959 *) Add support for smime-type MIME parameter in S/MIME messages which some
2963 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
2964 a threadsafe manner. Modify rsa code to use new function and add calls
2965 to dsa and dh code (which had race conditions before).
2968 *) Include the fixed error library code in the C error file definitions
2969 instead of fixing them up at runtime. This keeps the error code
2970 structures constant.
2973 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
2975 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
2978 *) Fixes for newer kerberos headers. NB: the casts are needed because
2979 the 'length' field is signed on one version and unsigned on another
2980 with no (?) obvious way to tell the difference, without these VC++
2981 complains. Also the "definition" of FAR (blank) is no longer included
2982 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
2983 some needed definitions.
2986 *) Undo Cygwin change.
2989 *) Added support for proxy certificates according to RFC 3820.
2990 Because they may be a security thread to unaware applications,
2991 they must be explicitely allowed in run-time. See
2992 docs/HOWTO/proxy_certificates.txt for further information.
2995 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
2997 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
2998 server and client random values. Previously
2999 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3000 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3002 This change has negligible security impact because:
3004 1. Server and client random values still have 24 bytes of pseudo random
3007 2. Server and client random values are sent in the clear in the initial
3010 3. The master secret is derived using the premaster secret (48 bytes in
3011 size for static RSA ciphersuites) as well as client server and random
3014 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3017 [Stephen Henson, reported by UK NISCC]
3019 *) Use Windows randomness collection on Cygwin.
3022 *) Fix hang in EGD/PRNGD query when communication socket is closed
3023 prematurely by EGD/PRNGD.
3024 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3026 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3029 *) Back-port of selected performance improvements from development
3030 branch, as well as improved support for PowerPC platforms.
3033 *) Add lots of checks for memory allocation failure, error codes to indicate
3034 failure and freeing up memory if a failure occurs.
3035 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3037 *) Add new -passin argument to dgst.
3040 *) Perform some character comparisons of different types in X509_NAME_cmp:
3041 this is needed for some certificates that reencode DNs into UTF8Strings
3042 (in violation of RFC3280) and can't or wont issue name rollover
3046 *) Make an explicit check during certificate validation to see that
3047 the CA setting in each certificate on the chain is correct. As a
3048 side effect always do the following basic checks on extensions,
3049 not just when there's an associated purpose to the check:
3051 - if there is an unhandled critical extension (unless the user
3052 has chosen to ignore this fault)
3053 - if the path length has been exceeded (if one is set at all)
3054 - that certain extensions fit the associated purpose (if one has
3058 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3060 *) Avoid a race condition when CRLs are checked in a multi threaded
3061 environment. This would happen due to the reordering of the revoked
3062 entries during signature checking and serial number lookup. Now the
3063 encoding is cached and the serial number sort performed under a lock.
3064 Add new STACK function sk_is_sorted().
3067 *) Add Delta CRL to the extension code.
3070 *) Various fixes to s3_pkt.c so alerts are sent properly.
3071 [David Holmes <d.holmes@f5.com>]
3073 *) Reduce the chances of duplicate issuer name and serial numbers (in
3074 violation of RFC3280) using the OpenSSL certificate creation utilities.
3075 This is done by creating a random 64 bit value for the initial serial
3076 number when a serial number file is created or when a self signed
3077 certificate is created using 'openssl req -x509'. The initial serial
3078 number file is created using 'openssl x509 -next_serial' in CA.pl
3079 rather than being initialized to 1.
3082 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3084 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3085 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3086 [Joe Orton, Steve Henson]
3088 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3090 [Joe Orton, Steve Henson]
3092 *) Make it possible to have multiple active certificates with the same
3093 subject in the CA index file. This is done only if the keyword
3094 'unique_subject' is set to 'no' in the main CA section (default
3095 if 'CA_default') of the configuration file. The value is saved
3096 with the database itself in a separate index attribute file,
3097 named like the index file with '.attr' appended to the name.
3100 *) X509 verify fixes. Disable broken certificate workarounds when
3101 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3102 keyUsage extension present. Don't accept CRLs with unhandled critical
3103 extensions: since verify currently doesn't process CRL extensions this
3104 rejects a CRL with *any* critical extensions. Add new verify error codes
3108 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
3109 A clarification of RFC2560 will require the use of OCTET STRINGs and
3110 some implementations cannot handle the current raw format. Since OpenSSL
3111 copies and compares OCSP nonces as opaque blobs without any attempt at
3112 parsing them this should not create any compatibility issues.
3115 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
3116 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
3117 this HMAC (and other) operations are several times slower than OpenSSL
3121 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
3122 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3124 *) Use the correct content when signing type "other".
3127 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
3129 *) Fix various bugs revealed by running the NISCC test suite:
3131 Stop out of bounds reads in the ASN1 code when presented with
3132 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3134 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
3136 If verify callback ignores invalid public key errors don't try to check
3137 certificate signature with the NULL public key.
3141 *) New -ignore_err option in ocsp application to stop the server
3142 exiting on the first error in a request.
3145 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3146 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3150 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3151 extra data after the compression methods not only for TLS 1.0
3152 but also for SSL 3.0 (as required by the specification).
3153 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3155 *) Change X509_certificate_type() to mark the key as exported/exportable
3156 when it's 512 *bits* long, not 512 bytes.
3159 *) Change AES_cbc_encrypt() so it outputs exact multiple of
3160 blocks during encryption.
3163 *) Various fixes to base64 BIO and non blocking I/O. On write
3164 flushes were not handled properly if the BIO retried. On read
3165 data was not being buffered properly and had various logic bugs.
3166 This also affects blocking I/O when the data being decoded is a
3170 *) Various S/MIME bugfixes and compatibility changes:
3171 output correct application/pkcs7 MIME type if
3172 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
3173 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
3174 of files as .eml work). Correctly handle very long lines in MIME
3178 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
3180 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3181 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3182 a protocol version number mismatch like a decryption error
3183 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3186 *) Turn on RSA blinding by default in the default implementation
3187 to avoid a timing attack. Applications that don't want it can call
3188 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3189 They would be ill-advised to do so in most cases.
3190 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3192 *) Change RSA blinding code so that it works when the PRNG is not
3193 seeded (in this case, the secret RSA exponent is abused as
3194 an unpredictable seed -- if it is not unpredictable, there
3195 is no point in blinding anyway). Make RSA blinding thread-safe
3196 by remembering the creator's thread ID in rsa->blinding and
3197 having all other threads use local one-time blinding factors
3198 (this requires more computation than sharing rsa->blinding, but
3199 avoids excessive locking; and if an RSA object is not shared
3200 between threads, blinding will still be very fast).
3203 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
3204 ENGINE as defaults for all supported algorithms irrespective of
3205 the 'flags' parameter. 'flags' is now honoured, so applications
3206 should make sure they are passing it correctly.
3209 *) Target "mingw" now allows native Windows code to be generated in
3210 the Cygwin environment as well as with the MinGW compiler.
3213 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
3215 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3216 via timing by performing a MAC computation even if incorrrect
3217 block cipher padding has been found. This is a countermeasure
3218 against active attacks where the attacker has to distinguish
3219 between bad padding and a MAC verification error. (CVE-2003-0078)
3221 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3222 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3223 Martin Vuagnoux (EPFL, Ilion)]
3225 *) Make the no-err option work as intended. The intention with no-err
3226 is not to have the whole error stack handling routines removed from
3227 libcrypto, it's only intended to remove all the function name and
3228 reason texts, thereby removing some of the footprint that may not
3229 be interesting if those errors aren't displayed anyway.
3231 NOTE: it's still possible for any application or module to have it's
3232 own set of error texts inserted. The routines are there, just not
3233 used by default when no-err is given.
3236 *) Add support for FreeBSD on IA64.
3237 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
3239 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
3240 Kerberos function mit_des_cbc_cksum(). Before this change,
3241 the value returned by DES_cbc_cksum() was like the one from
3242 mit_des_cbc_cksum(), except the bytes were swapped.
3243 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
3245 *) Allow an application to disable the automatic SSL chain building.
3246 Before this a rather primitive chain build was always performed in
3247 ssl3_output_cert_chain(): an application had no way to send the
3248 correct chain if the automatic operation produced an incorrect result.
3250 Now the chain builder is disabled if either:
3252 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
3254 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
3256 The reasoning behind this is that an application would not want the
3257 auto chain building to take place if extra chain certificates are
3258 present and it might also want a means of sending no additional
3259 certificates (for example the chain has two certificates and the
3263 *) Add the possibility to build without the ENGINE framework.
3264 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3266 *) Under Win32 gmtime() can return NULL: check return value in
3267 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
3270 *) DSA routines: under certain error conditions uninitialized BN objects
3271 could be freed. Solution: make sure initialization is performed early
3272 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
3273 Nils Larsch <nla@trustcenter.de> via PR#459)
3276 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
3277 checked on reconnect on the client side, therefore session resumption
3278 could still fail with a "ssl session id is different" error. This
3279 behaviour is masked when SSL_OP_ALL is used due to
3280 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
3281 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3282 followup to PR #377.
3285 *) IA-32 assembler support enhancements: unified ELF targets, support
3286 for SCO/Caldera platforms, fix for Cygwin shared build.
3289 *) Add support for FreeBSD on sparc64. As a consequence, support for
3290 FreeBSD on non-x86 processors is separate from x86 processors on
3291 the config script, much like the NetBSD support.
3292 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
3294 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
3296 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
3299 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
3300 code (06) was taken as the first octet of the session ID and the last
3301 octet was ignored consequently. As a result SSLv2 client side session
3302 caching could not have worked due to the session ID mismatch between
3304 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3308 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
3309 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
3313 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
3314 seems that in spite of existing for more than a year, many application
3315 author have done nothing to provide the necessary callbacks, which
3316 means that this particular engine will not work properly anywhere.
3317 This is a very unfortunate situation which forces us, in the name
3318 of usability, to give the hw_ncipher.c a static lock, which is part
3320 NOTE: This is for the 0.9.7 series ONLY. This hack will never
3321 appear in 0.9.8 or later. We EXPECT application authors to have
3322 dealt properly with this when 0.9.8 is released (unless we actually
3323 make such changes in the libcrypto locking code that changes will
3324 have to be made anyway).
3327 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
3328 octets have been read, EOF or an error occurs. Without this change
3329 some truncated ASN1 structures will not produce an error.
3332 *) Disable Heimdal support, since it hasn't been fully implemented.
3333 Still give the possibility to force the use of Heimdal, but with
3334 warnings and a request that patches get sent to openssl-dev.
3337 *) Add the VC-CE target, introduce the WINCE sysname, and add
3338 INSTALL.WCE and appropriate conditionals to make it build.
3339 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3341 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
3342 cygssl-x.y.z.dll, where x, y and z are the major, minor and
3343 edit numbers of the version.
3344 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3346 *) Introduce safe string copy and catenation functions
3347 (BUF_strlcpy() and BUF_strlcat()).
3348 [Ben Laurie (CHATS) and Richard Levitte]
3350 *) Avoid using fixed-size buffers for one-line DNs.
3351 [Ben Laurie (CHATS)]
3353 *) Add BUF_MEM_grow_clean() to avoid information leakage when
3354 resizing buffers containing secrets, and use where appropriate.
3355 [Ben Laurie (CHATS)]
3357 *) Avoid using fixed size buffers for configuration file location.
3358 [Ben Laurie (CHATS)]
3360 *) Avoid filename truncation for various CA files.
3361 [Ben Laurie (CHATS)]
3363 *) Use sizeof in preference to magic numbers.
3364 [Ben Laurie (CHATS)]
3366 *) Avoid filename truncation in cert requests.
3367 [Ben Laurie (CHATS)]
3369 *) Add assertions to check for (supposedly impossible) buffer
3371 [Ben Laurie (CHATS)]
3373 *) Don't cache truncated DNS entries in the local cache (this could
3374 potentially lead to a spoofing attack).
3375 [Ben Laurie (CHATS)]
3377 *) Fix various buffers to be large enough for hex/decimal
3378 representations in a platform independent manner.
3379 [Ben Laurie (CHATS)]
3381 *) Add CRYPTO_realloc_clean() to avoid information leakage when
3382 resizing buffers containing secrets, and use where appropriate.
3383 [Ben Laurie (CHATS)]
3385 *) Add BIO_indent() to avoid much slightly worrying code to do
3387 [Ben Laurie (CHATS)]
3389 *) Convert sprintf()/BIO_puts() to BIO_printf().
3390 [Ben Laurie (CHATS)]
3392 *) buffer_gets() could terminate with the buffer only half
3394 [Ben Laurie (CHATS)]
3396 *) Add assertions to prevent user-supplied crypto functions from
3397 overflowing internal buffers by having large block sizes, etc.
3398 [Ben Laurie (CHATS)]
3400 *) New OPENSSL_assert() macro (similar to assert(), but enabled
3402 [Ben Laurie (CHATS)]
3404 *) Eliminate unused copy of key in RC4.
3405 [Ben Laurie (CHATS)]
3407 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
3408 [Ben Laurie (CHATS)]
3410 *) Fix off-by-one error in EGD path.
3411 [Ben Laurie (CHATS)]
3413 *) If RANDFILE path is too long, ignore instead of truncating.
3414 [Ben Laurie (CHATS)]
3416 *) Eliminate unused and incorrectly sized X.509 structure
3418 [Ben Laurie (CHATS)]
3420 *) Eliminate unused and dangerous function knumber().
3421 [Ben Laurie (CHATS)]
3423 *) Eliminate unused and dangerous structure, KSSL_ERR.
3424 [Ben Laurie (CHATS)]
3426 *) Protect against overlong session ID context length in an encoded
3427 session object. Since these are local, this does not appear to be
3429 [Ben Laurie (CHATS)]
3431 *) Change from security patch (see 0.9.6e below) that did not affect
3432 the 0.9.6 release series:
3434 Remote buffer overflow in SSL3 protocol - an attacker could
3435 supply an oversized master key in Kerberos-enabled versions.
3437 [Ben Laurie (CHATS)]
3439 *) Change the SSL kerb5 codes to match RFC 2712.
3442 *) Make -nameopt work fully for req and add -reqopt switch.
3443 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
3445 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
3446 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
3448 *) Make sure tests can be performed even if the corresponding algorithms
3449 have been removed entirely. This was also the last step to make
3450 OpenSSL compilable with DJGPP under all reasonable conditions.
3451 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
3453 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
3454 to allow version independent disabling of normally unselected ciphers,
3455 which may be activated as a side-effect of selecting a single cipher.
3457 (E.g., cipher list string "RSA" enables ciphersuites that are left
3458 out of "ALL" because they do not provide symmetric encryption.
3459 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
3460 [Lutz Jaenicke, Bodo Moeller]
3462 *) Add appropriate support for separate platform-dependent build
3463 directories. The recommended way to make a platform-dependent
3464 build directory is the following (tested on Linux), maybe with
3467 # Place yourself outside of the OpenSSL source tree. In
3468 # this example, the environment variable OPENSSL_SOURCE
3469 # is assumed to contain the absolute OpenSSL source directory.
3470 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
3471 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
3472 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
3473 mkdir -p `dirname $F`
3474 ln -s $OPENSSL_SOURCE/$F $F
3477 To be absolutely sure not to disturb the source tree, a "make clean"
3478 is a good thing. If it isn't successfull, don't worry about it,
3479 it probably means the source directory is very clean.
3482 *) Make sure any ENGINE control commands make local copies of string
3483 pointers passed to them whenever necessary. Otherwise it is possible
3484 the caller may have overwritten (or deallocated) the original string
3485 data when a later ENGINE operation tries to use the stored values.
3486 [Götz Babin-Ebell <babinebell@trustcenter.de>]
3488 *) Improve diagnostics in file reading and command-line digests.
3489 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
3491 *) Add AES modes CFB and OFB to the object database. Correct an
3492 error in AES-CFB decryption.
3495 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
3496 allows existing EVP_CIPHER_CTX structures to be reused after
3497 calling EVP_*Final(). This behaviour is used by encryption
3498 BIOs and some applications. This has the side effect that
3499 applications must explicitly clean up cipher contexts with
3500 EVP_CIPHER_CTX_cleanup() or they will leak memory.
3503 *) Check the values of dna and dnb in bn_mul_recursive before calling
3504 bn_mul_comba (a non zero value means the a or b arrays do not contain
3505 n2 elements) and fallback to bn_mul_normal if either is not zero.
3508 *) Fix escaping of non-ASCII characters when using the -subj option
3509 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
3512 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
3513 form for "surname", serialNumber has no short form.
3514 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
3515 therefore remove "mail" short name for "internet 7".
3516 The OID for unique identifiers in X509 certificates is
3517 x500UniqueIdentifier, not uniqueIdentifier.
3518 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
3521 *) Add an "init" command to the ENGINE config module and auto initialize
3522 ENGINEs. Without any "init" command the ENGINE will be initialized
3523 after all ctrl commands have been executed on it. If init=1 the
3524 ENGINE is initailized at that point (ctrls before that point are run
3525 on the uninitialized ENGINE and after on the initialized one). If
3526 init=0 then the ENGINE will not be iniatialized at all.
3529 *) Fix the 'app_verify_callback' interface so that the user-defined
3530 argument is actually passed to the callback: In the
3531 SSL_CTX_set_cert_verify_callback() prototype, the callback
3532 declaration has been changed from
3535 int (*cb)(X509_STORE_CTX *,void *);
3536 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
3537 i=s->ctx->app_verify_callback(&ctx)
3538 has been changed into
3539 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
3541 To update applications using SSL_CTX_set_cert_verify_callback(),
3542 a dummy argument can be added to their callback functions.
3543 [D. K. Smetters <smetters@parc.xerox.com>]
3545 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
3546 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
3548 *) Add and OPENSSL_LOAD_CONF define which will cause
3549 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
3550 This allows older applications to transparently support certain
3551 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
3552 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
3553 load the config file and OPENSSL_add_all_algorithms_conf() which will
3554 always load it have also been added.
3557 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
3558 Adjust NIDs and EVP layer.
3559 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3561 *) Config modules support in openssl utility.
3563 Most commands now load modules from the config file,
3564 though in a few (such as version) this isn't done
3565 because it couldn't be used for anything.
3567 In the case of ca and req the config file used is
3568 the same as the utility itself: that is the -config
3569 command line option can be used to specify an
3573 *) Move default behaviour from OPENSSL_config(). If appname is NULL
3574 use "openssl_conf" if filename is NULL use default openssl config file.
3577 *) Add an argument to OPENSSL_config() to allow the use of an alternative
3578 config section name. Add a new flag to tolerate a missing config file
3579 and move code to CONF_modules_load_file().
3582 *) Support for crypto accelerator cards from Accelerated Encryption
3583 Processing, www.aep.ie. (Use engine 'aep')
3584 The support was copied from 0.9.6c [engine] and adapted/corrected
3585 to work with the new engine framework.
3586 [AEP Inc. and Richard Levitte]
3588 *) Support for SureWare crypto accelerator cards from Baltimore
3589 Technologies. (Use engine 'sureware')
3590 The support was copied from 0.9.6c [engine] and adapted
3591 to work with the new engine framework.
3594 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
3595 make the newer ENGINE framework commands for the CHIL engine work.
3596 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
3598 *) Make it possible to produce shared libraries on ReliantUNIX.
3599 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
3601 *) Add the configuration target debug-linux-ppro.
3602 Make 'openssl rsa' use the general key loading routines
3603 implemented in apps.c, and make those routines able to
3604 handle the key format FORMAT_NETSCAPE and the variant
3606 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3608 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3609 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3611 *) Add -keyform to rsautl, and document -engine.
3612 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
3614 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
3615 BIO_R_NO_SUCH_FILE error code rather than the generic
3616 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
3619 *) Add new functions
3621 ERR_peek_last_error_line
3622 ERR_peek_last_error_line_data.
3623 These are similar to
3626 ERR_peek_error_line_data,
3627 but report on the latest error recorded rather than the first one
3628 still in the error queue.
3629 [Ben Laurie, Bodo Moeller]
3631 *) default_algorithms option in ENGINE config module. This allows things
3633 default_algorithms = ALL
3634 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
3637 *) Prelminary ENGINE config module.
3640 *) New experimental application configuration code.
3643 *) Change the AES code to follow the same name structure as all other
3644 symmetric ciphers, and behave the same way. Move everything to
3645 the directory crypto/aes, thereby obsoleting crypto/rijndael.
3646 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3648 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
3649 [Ben Laurie and Theo de Raadt]
3651 *) Add option to output public keys in req command.
3652 [Massimiliano Pala madwolf@openca.org]
3654 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
3655 (up to about 10% better than before for P-192 and P-224).
3658 *) New functions/macros
3660 SSL_CTX_set_msg_callback(ctx, cb)
3661 SSL_CTX_set_msg_callback_arg(ctx, arg)
3662 SSL_set_msg_callback(ssl, cb)
3663 SSL_set_msg_callback_arg(ssl, arg)
3665 to request calling a callback function
3667 void cb(int write_p, int version, int content_type,
3668 const void *buf, size_t len, SSL *ssl, void *arg)
3670 whenever a protocol message has been completely received
3671 (write_p == 0) or sent (write_p == 1). Here 'version' is the
3672 protocol version according to which the SSL library interprets
3673 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
3674 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
3675 the content type as defined in the SSL 3.0/TLS 1.0 protocol
3676 specification (change_cipher_spec(20), alert(21), handshake(22)).
3677 'buf' and 'len' point to the actual message, 'ssl' to the
3678 SSL object, and 'arg' is the application-defined value set by
3679 SSL[_CTX]_set_msg_callback_arg().
3681 'openssl s_client' and 'openssl s_server' have new '-msg' options
3682 to enable a callback that displays all protocol messages.
3685 *) Change the shared library support so shared libraries are built as
3686 soon as the corresponding static library is finished, and thereby get
3687 openssl and the test programs linked against the shared library.
3688 This still only happens when the keyword "shard" has been given to
3689 the configuration scripts.
3691 NOTE: shared library support is still an experimental thing, and
3692 backward binary compatibility is still not guaranteed.
3693 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
3695 *) Add support for Subject Information Access extension.
3696 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3698 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
3699 additional bytes when new memory had to be allocated, not just
3700 when reusing an existing buffer.
3703 *) New command line and configuration option 'utf8' for the req command.
3704 This allows field values to be specified as UTF8 strings.
3707 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
3708 runs for the former and machine-readable output for the latter.
3711 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
3712 of the e-mail address in the DN (i.e., it will go into a certificate
3713 extension only). The new configuration file option 'email_in_dn = no'
3714 has the same effect.
3715 [Massimiliano Pala madwolf@openca.org]
3717 *) Change all functions with names starting with des_ to be starting
3718 with DES_ instead. Add wrappers that are compatible with libdes,
3719 but are named _ossl_old_des_*. Finally, add macros that map the
3720 des_* symbols to the corresponding _ossl_old_des_* if libdes
3721 compatibility is desired. If OpenSSL 0.9.6c compatibility is
3722 desired, the des_* symbols will be mapped to DES_*, with one
3725 Since we provide two compatibility mappings, the user needs to
3726 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
3727 compatibility is desired. The default (i.e., when that macro
3728 isn't defined) is OpenSSL 0.9.6c compatibility.
3730 There are also macros that enable and disable the support of old
3731 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
3732 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
3733 are defined, the default will apply: to support the old des routines.
3735 In either case, one must include openssl/des.h to get the correct
3736 definitions. Do not try to just include openssl/des_old.h, that
3739 NOTE: This is a major break of an old API into a new one. Software
3740 authors are encouraged to switch to the DES_ style functions. Some
3741 time in the future, des_old.h and the libdes compatibility functions
3742 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
3743 default), and then completely removed.
3746 *) Test for certificates which contain unsupported critical extensions.
3747 If such a certificate is found during a verify operation it is
3748 rejected by default: this behaviour can be overridden by either
3749 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
3750 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
3751 X509_supported_extension() has also been added which returns 1 if a
3752 particular extension is supported.
3755 *) Modify the behaviour of EVP cipher functions in similar way to digests
3756 to retain compatibility with existing code.
3759 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
3760 compatibility with existing code. In particular the 'ctx' parameter does
3761 not have to be to be initialized before the call to EVP_DigestInit() and
3762 it is tidied up after a call to EVP_DigestFinal(). New function
3763 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
3764 EVP_MD_CTX_copy() changed to not require the destination to be
3765 initialized valid and new function EVP_MD_CTX_copy_ex() added which
3766 requires the destination to be valid.
3768 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
3769 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
3772 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
3773 so that complete 'Handshake' protocol structures are kept in memory
3774 instead of overwriting 'msg_type' and 'length' with 'body' data.
3777 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
3778 [Massimo Santin via Richard Levitte]
3780 *) Major restructuring to the underlying ENGINE code. This includes
3781 reduction of linker bloat, separation of pure "ENGINE" manipulation
3782 (initialisation, etc) from functionality dealing with implementations
3783 of specific crypto iterfaces. This change also introduces integrated
3784 support for symmetric ciphers and digest implementations - so ENGINEs
3785 can now accelerate these by providing EVP_CIPHER and EVP_MD
3786 implementations of their own. This is detailed in crypto/engine/README
3787 as it couldn't be adequately described here. However, there are a few
3788 API changes worth noting - some RSA, DSA, DH, and RAND functions that
3789 were changed in the original introduction of ENGINE code have now
3790 reverted back - the hooking from this code to ENGINE is now a good
3791 deal more passive and at run-time, operations deal directly with
3792 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
3793 dereferencing through an ENGINE pointer any more. Also, the ENGINE
3794 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
3795 they were not being used by the framework as there is no concept of a
3796 BIGNUM_METHOD and they could not be generalised to the new
3797 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
3798 ENGINE_cpy() has been removed as it cannot be consistently defined in
3802 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
3805 *) Change mkdef.pl to sort symbols that get the same entry number,
3806 and make sure the automatically generated functions ERR_load_*
3807 become part of libeay.num as well.
3810 *) New function SSL_renegotiate_pending(). This returns true once
3811 renegotiation has been requested (either SSL_renegotiate() call
3812 or HelloRequest/ClientHello receveived from the peer) and becomes
3813 false once a handshake has been completed.
3814 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
3815 sends a HelloRequest, but does not ensure that a handshake takes
3816 place. SSL_renegotiate_pending() is useful for checking if the
3817 client has followed the request.)
3820 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
3821 By default, clients may request session resumption even during
3822 renegotiation (if session ID contexts permit); with this option,
3823 session resumption is possible only in the first handshake.
3825 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
3826 more bits available for options that should not be part of
3827 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
3830 *) Add some demos for certificate and certificate request creation.
3833 *) Make maximum certificate chain size accepted from the peer application
3834 settable (SSL*_get/set_max_cert_list()), as proposed by
3835 "Douglas E. Engert" <deengert@anl.gov>.
3838 *) Add support for shared libraries for Unixware-7
3839 (Boyd Lynn Gerber <gerberb@zenez.com>).
3842 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
3843 be done prior to destruction. Use this to unload error strings from
3844 ENGINEs that load their own error strings. NB: This adds two new API
3845 functions to "get" and "set" this destroy handler in an ENGINE.
3848 *) Alter all existing ENGINE implementations (except "openssl" and
3849 "openbsd") to dynamically instantiate their own error strings. This
3850 makes them more flexible to be built both as statically-linked ENGINEs
3851 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
3852 Also, add stub code to each that makes building them as self-contained
3853 shared-libraries easier (see README.ENGINE).
3856 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
3857 implementations into applications that are completely implemented in
3858 self-contained shared-libraries. The "dynamic" ENGINE exposes control
3859 commands that can be used to configure what shared-library to load and
3860 to control aspects of the way it is handled. Also, made an update to
3861 the README.ENGINE file that brings its information up-to-date and
3862 provides some information and instructions on the "dynamic" ENGINE
3863 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
3866 *) Make it possible to unload ranges of ERR strings with a new
3867 "ERR_unload_strings" function.
3870 *) Add a copy() function to EVP_MD.
3873 *) Make EVP_MD routines take a context pointer instead of just the
3874 md_data void pointer.
3877 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
3878 that the digest can only process a single chunk of data
3879 (typically because it is provided by a piece of
3880 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
3881 is only going to provide a single chunk of data, and hence the
3882 framework needn't accumulate the data for oneshot drivers.
3885 *) As with "ERR", make it possible to replace the underlying "ex_data"
3886 functions. This change also alters the storage and management of global
3887 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
3888 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
3889 index counters. The API functions that use this state have been changed
3890 to take a "class_index" rather than pointers to the class's local STACK
3891 and counter, and there is now an API function to dynamically create new
3892 classes. This centralisation allows us to (a) plug a lot of the
3893 thread-safety problems that existed, and (b) makes it possible to clean
3894 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
3895 such data would previously have always leaked in application code and
3896 workarounds were in place to make the memory debugging turn a blind eye
3897 to it. Application code that doesn't use this new function will still
3898 leak as before, but their memory debugging output will announce it now
3899 rather than letting it slide.
3901 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
3902 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
3903 has a return value to indicate success or failure.
3906 *) Make it possible to replace the underlying "ERR" functions such that the
3907 global state (2 LHASH tables and 2 locks) is only used by the "default"
3908 implementation. This change also adds two functions to "get" and "set"
3909 the implementation prior to it being automatically set the first time
3910 any other ERR function takes place. Ie. an application can call "get",
3911 pass the return value to a module it has just loaded, and that module
3912 can call its own "set" function using that value. This means the
3913 module's "ERR" operations will use (and modify) the error state in the
3914 application and not in its own statically linked copy of OpenSSL code.
3917 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
3918 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
3919 the operation, and provides a more encapsulated way for external code
3920 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
3921 to use these functions rather than manually incrementing the counts.
3923 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
3926 *) Add EVP test program.
3929 *) Add symmetric cipher support to ENGINE. Expect the API to change!
3932 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
3933 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
3934 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
3935 These allow a CRL to be built without having to access X509_CRL fields
3936 directly. Modify 'ca' application to use new functions.
3939 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
3940 bug workarounds. Rollback attack detection is a security feature.
3941 The problem will only arise on OpenSSL servers when TLSv1 is not
3942 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
3943 Software authors not wanting to support TLSv1 will have special reasons
3944 for their choice and can explicitly enable this option.
3945 [Bodo Moeller, Lutz Jaenicke]
3947 *) Rationalise EVP so it can be extended: don't include a union of
3948 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
3949 (similar to those existing for EVP_CIPHER_CTX).
3954 EVP_MD_CTX_init(&md); /* new function call */
3955 EVP_DigestInit(&md, EVP_sha1());
3956 EVP_DigestUpdate(&md, in, len);
3957 EVP_DigestFinal(&md, out, NULL);
3958 EVP_MD_CTX_cleanup(&md); /* new function call */
3962 *) Make DES key schedule conform to the usual scheme, as well as
3963 correcting its structure. This means that calls to DES functions
3964 now have to pass a pointer to a des_key_schedule instead of a
3965 plain des_key_schedule (which was actually always a pointer
3968 des_key_schedule ks;
3970 des_set_key_checked(..., &ks);
3971 des_ncbc_encrypt(..., &ks, ...);
3973 (Note that a later change renames 'des_...' into 'DES_...'.)
3976 *) Initial reduction of linker bloat: the use of some functions, such as
3977 PEM causes large amounts of unused functions to be linked in due to
3978 poor organisation. For example pem_all.c contains every PEM function
3979 which has a knock on effect of linking in large amounts of (unused)
3980 ASN1 code. Grouping together similar functions and splitting unrelated
3981 functions prevents this.
3984 *) Cleanup of EVP macros.
3987 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
3988 correct _ecb suffix.
3991 *) Add initial OCSP responder support to ocsp application. The
3992 revocation information is handled using the text based index
3993 use by the ca application. The responder can either handle
3994 requests generated internally, supplied in files (for example
3995 via a CGI script) or using an internal minimal server.
3998 *) Add configuration choices to get zlib compression for TLS.
4001 *) Changes to Kerberos SSL for RFC 2712 compliance:
4002 1. Implemented real KerberosWrapper, instead of just using
4003 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4004 2. Implemented optional authenticator field of KerberosWrapper.
4006 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4007 and authenticator structs; see crypto/krb5/.
4009 Generalized Kerberos calls to support multiple Kerberos libraries.
4010 [Vern Staats <staatsvr@asc.hpc.mil>,
4011 Jeffrey Altman <jaltman@columbia.edu>
4012 via Richard Levitte]
4014 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4015 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4016 values for each of the key sizes rather than having just
4017 parameters (and 'speed' generating keys each time).
4020 *) Speed up EVP routines.
4023 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4024 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4025 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4026 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4028 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4029 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4030 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4033 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4035 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4038 *) Added the OS2-EMX target.
4039 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4041 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4042 to support NCONF routines in extension code. New function CONF_set_nconf()
4043 to allow functions which take an NCONF to also handle the old LHASH
4044 structure: this means that the old CONF compatible routines can be
4045 retained (in particular wrt extensions) without having to duplicate the
4046 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4049 *) Enhance the general user interface with mechanisms for inner control
4050 and with possibilities to have yes/no kind of prompts.
4053 *) Change all calls to low level digest routines in the library and
4054 applications to use EVP. Add missing calls to HMAC_cleanup() and
4055 don't assume HMAC_CTX can be copied using memcpy().
4056 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4058 *) Add the possibility to control engines through control names but with
4059 arbitrary arguments instead of just a string.
4060 Change the key loaders to take a UI_METHOD instead of a callback
4061 function pointer. NOTE: this breaks binary compatibility with earlier
4062 versions of OpenSSL [engine].
4063 Adapt the nCipher code for these new conditions and add a card insertion
4067 *) Enhance the general user interface with mechanisms to better support
4068 dialog box interfaces, application-defined prompts, the possibility
4069 to use defaults (for example default passwords from somewhere else)
4070 and interrupts/cancellations.
4073 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4074 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4077 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4078 tidy up some unnecessarily weird code in 'sk_new()').
4079 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4081 *) Change the key loading routines for ENGINEs to use the same kind
4082 callback (pem_password_cb) as all other routines that need this
4086 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4087 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4088 than this minimum value is recommended.
4091 *) New random seeder for OpenVMS, using the system process statistics
4092 that are easily reachable.
4095 *) Windows apparently can't transparently handle global
4096 variables defined in DLLs. Initialisations such as:
4098 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4100 wont compile. This is used by the any applications that need to
4101 declare their own ASN1 modules. This was fixed by adding the option
4102 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4103 needed for static libraries under Win32.
4106 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
4107 setting of purpose and trust fields. New X509_STORE trust and
4108 purpose functions and tidy up setting in other SSL functions.
4111 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
4112 structure. These are inherited by X509_STORE_CTX when it is
4113 initialised. This allows various defaults to be set in the
4114 X509_STORE structure (such as flags for CRL checking and custom
4115 purpose or trust settings) for functions which only use X509_STORE_CTX
4116 internally such as S/MIME.
4118 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
4119 trust settings if they are not set in X509_STORE. This allows X509_STORE
4120 purposes and trust (in S/MIME for example) to override any set by default.
4122 Add command line options for CRL checking to smime, s_client and s_server
4126 *) Initial CRL based revocation checking. If the CRL checking flag(s)
4127 are set then the CRL is looked up in the X509_STORE structure and
4128 its validity and signature checked, then if the certificate is found
4129 in the CRL the verify fails with a revoked error.
4131 Various new CRL related callbacks added to X509_STORE_CTX structure.
4133 Command line options added to 'verify' application to support this.
4135 This needs some additional work, such as being able to handle multiple
4136 CRLs with different times, extension based lookup (rather than just
4137 by subject name) and ultimately more complete V2 CRL extension
4141 *) Add a general user interface API (crypto/ui/). This is designed
4142 to replace things like des_read_password and friends (backward
4143 compatibility functions using this new API are provided).
4144 The purpose is to remove prompting functions from the DES code
4145 section as well as provide for prompting through dialog boxes in
4146 a window system and the like.
4149 *) Add "ex_data" support to ENGINE so implementations can add state at a
4150 per-structure level rather than having to store it globally.
4153 *) Make it possible for ENGINE structures to be copied when retrieved by
4154 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
4155 This causes the "original" ENGINE structure to act like a template,
4156 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
4157 operational state can be localised to each ENGINE structure, despite the
4158 fact they all share the same "methods". New ENGINE structures returned in
4159 this case have no functional references and the return value is the single
4160 structural reference. This matches the single structural reference returned
4161 by ENGINE_by_id() normally, when it is incremented on the pre-existing
4165 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
4166 needs to match any other type at all we need to manually clear the
4170 *) Changes to the "openssl engine" utility to include;
4171 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
4172 about an ENGINE's available control commands.
4173 - executing control commands from command line arguments using the
4174 '-pre' and '-post' switches. '-post' is only used if '-t' is
4175 specified and the ENGINE is successfully initialised. The syntax for
4176 the individual commands are colon-separated, for example;
4177 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
4180 *) New dynamic control command support for ENGINEs. ENGINEs can now
4181 declare their own commands (numbers), names (strings), descriptions,
4182 and input types for run-time discovery by calling applications. A
4183 subset of these commands are implicitly classed as "executable"
4184 depending on their input type, and only these can be invoked through
4185 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
4186 can be based on user input, config files, etc). The distinction is
4187 that "executable" commands cannot return anything other than a boolean
4188 result and can only support numeric or string input, whereas some
4189 discoverable commands may only be for direct use through
4190 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
4191 pointers, or other custom uses. The "executable" commands are to
4192 support parameterisations of ENGINE behaviour that can be
4193 unambiguously defined by ENGINEs and used consistently across any
4194 OpenSSL-based application. Commands have been added to all the
4195 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
4196 control over shared-library paths without source code alterations.
4199 *) Changed all ENGINE implementations to dynamically allocate their
4200 ENGINEs rather than declaring them statically. Apart from this being
4201 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
4202 this also allows the implementations to compile without using the
4203 internal engine_int.h header.
4206 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
4207 'const' value. Any code that should be able to modify a RAND_METHOD
4208 should already have non-const pointers to it (ie. they should only
4209 modify their own ones).
4212 *) Made a variety of little tweaks to the ENGINE code.
4213 - "atalla" and "ubsec" string definitions were moved from header files
4214 to C code. "nuron" string definitions were placed in variables
4215 rather than hard-coded - allowing parameterisation of these values
4216 later on via ctrl() commands.
4217 - Removed unused "#if 0"'d code.
4218 - Fixed engine list iteration code so it uses ENGINE_free() to release
4219 structural references.
4220 - Constified the RAND_METHOD element of ENGINE structures.
4221 - Constified various get/set functions as appropriate and added
4222 missing functions (including a catch-all ENGINE_cpy that duplicates
4223 all ENGINE values onto a new ENGINE except reference counts/state).
4224 - Removed NULL parameter checks in get/set functions. Setting a method
4225 or function to NULL is a way of cancelling out a previously set
4226 value. Passing a NULL ENGINE parameter is just plain stupid anyway
4227 and doesn't justify the extra error symbols and code.
4228 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
4229 flags from engine_int.h to engine.h.
4230 - Changed prototypes for ENGINE handler functions (init(), finish(),
4231 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
4234 *) Implement binary inversion algorithm for BN_mod_inverse in addition
4235 to the algorithm using long division. The binary algorithm can be
4236 used only if the modulus is odd. On 32-bit systems, it is faster
4237 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
4238 roughly 5-15% for 256-bit moduli), so we use it only for moduli
4239 up to 450 bits. In 64-bit environments, the binary algorithm
4240 appears to be advantageous for much longer moduli; here we use it
4241 for moduli up to 2048 bits.
4244 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
4245 could not support the combine flag in choice fields.
4248 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
4249 extensions from a certificate request to the certificate.
4252 *) Allow multiple 'certopt' and 'nameopt' options to be separated
4253 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
4254 file: this allows the display of the certificate about to be
4255 signed to be customised, to allow certain fields to be included
4256 or excluded and extension details. The old system didn't display
4257 multicharacter strings properly, omitted fields not in the policy
4258 and couldn't display additional details such as extensions.
4261 *) Function EC_POINTs_mul for multiple scalar multiplication
4262 of an arbitrary number of elliptic curve points
4263 \sum scalars[i]*points[i],
4264 optionally including the generator defined for the EC_GROUP:
4265 scalar*generator + \sum scalars[i]*points[i].
4267 EC_POINT_mul is a simple wrapper function for the typical case
4268 that the point list has just one item (besides the optional
4272 *) First EC_METHODs for curves over GF(p):
4274 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
4275 operations and provides various method functions that can also
4276 operate with faster implementations of modular arithmetic.
4278 EC_GFp_mont_method() reuses most functions that are part of
4279 EC_GFp_simple_method, but uses Montgomery arithmetic.
4281 [Bodo Moeller; point addition and point doubling
4282 implementation directly derived from source code provided by
4283 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
4285 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
4286 crypto/ec/ec_lib.c):
4288 Curves are EC_GROUP objects (with an optional group generator)
4289 based on EC_METHODs that are built into the library.
4291 Points are EC_POINT objects based on EC_GROUP objects.
4293 Most of the framework would be able to handle curves over arbitrary
4294 finite fields, but as there are no obvious types for fields other
4295 than GF(p), some functions are limited to that for now.
4298 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
4299 that the file contains a complete HTTP response.
4302 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
4303 change the def and num file printf format specifier from "%-40sXXX"
4304 to "%-39s XXX". The latter will always guarantee a space after the
4305 field while the former will cause them to run together if the field
4306 is 40 of more characters long.
4309 *) Constify the cipher and digest 'method' functions and structures
4310 and modify related functions to take constant EVP_MD and EVP_CIPHER
4314 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
4315 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
4318 *) Modify EVP_Digest*() routines so they now return values. Although the
4319 internal software routines can never fail additional hardware versions
4323 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
4325 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
4326 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
4329 ERR_R_NESTED_ASN1_ERROR
4331 ERR_R_MISSING_ASN1_EOS
4332 were 4 .. 9, conflicting with
4333 ERR_LIB_RSA (= ERR_R_RSA_LIB)
4335 ERR_LIB_PEM (= ERR_R_PEM_LIB).
4336 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
4338 Add new error code 'ERR_R_INTERNAL_ERROR'.
4341 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
4345 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
4346 sets the subject name for a new request or supersedes the
4347 subject name in a given request. Formats that can be parsed are
4348 'CN=Some Name, OU=myOU, C=IT'
4350 'CN=Some Name/OU=myOU/C=IT'.
4352 Add options '-batch' and '-verbose' to 'openssl req'.
4353 [Massimiliano Pala <madwolf@hackmasters.net>]
4355 *) Introduce the possibility to access global variables through
4356 functions on platform were that's the best way to handle exporting
4357 global variables in shared libraries. To enable this functionality,
4358 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
4359 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
4360 is normally done by Configure or something similar).
4362 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
4363 in the source file (foo.c) like this:
4365 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
4366 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
4368 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
4369 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
4371 OPENSSL_DECLARE_GLOBAL(int,foo);
4372 #define foo OPENSSL_GLOBAL_REF(foo)
4373 OPENSSL_DECLARE_GLOBAL(double,bar);
4374 #define bar OPENSSL_GLOBAL_REF(bar)
4376 The #defines are very important, and therefore so is including the
4377 header file everywhere where the defined globals are used.
4379 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
4380 of ASN.1 items, but that structure is a bit different.
4382 The largest change is in util/mkdef.pl which has been enhanced with
4383 better and easier to understand logic to choose which symbols should
4384 go into the Windows .def files as well as a number of fixes and code
4385 cleanup (among others, algorithm keywords are now sorted
4386 lexicographically to avoid constant rewrites).
4389 *) In BN_div() keep a copy of the sign of 'num' before writing the
4390 result to 'rm' because if rm==num the value will be overwritten
4391 and produce the wrong result if 'num' is negative: this caused
4392 problems with BN_mod() and BN_nnmod().
4395 *) Function OCSP_request_verify(). This checks the signature on an
4396 OCSP request and verifies the signer certificate. The signer
4397 certificate is just checked for a generic purpose and OCSP request
4401 *) Add OCSP_check_validity() function to check the validity of OCSP
4402 responses. OCSP responses are prepared in real time and may only
4403 be a few seconds old. Simply checking that the current time lies
4404 between thisUpdate and nextUpdate max reject otherwise valid responses
4405 caused by either OCSP responder or client clock inaccuracy. Instead
4406 we allow thisUpdate and nextUpdate to fall within a certain period of
4407 the current time. The age of the response can also optionally be
4408 checked. Two new options -validity_period and -status_age added to
4412 *) If signature or public key algorithm is unrecognized print out its
4413 OID rather that just UNKNOWN.
4416 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
4417 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
4418 ID to be generated from the issuer certificate alone which can then be
4419 passed to OCSP_id_issuer_cmp().
4422 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
4423 ASN1 modules to export functions returning ASN1_ITEM pointers
4424 instead of the ASN1_ITEM structures themselves. This adds several
4425 new macros which allow the underlying ASN1 function/structure to
4426 be accessed transparently. As a result code should not use ASN1_ITEM
4427 references directly (such as &X509_it) but instead use the relevant
4428 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
4429 use of the new ASN1 code on platforms where exporting structures
4430 is problematical (for example in shared libraries) but exporting
4431 functions returning pointers to structures is not.
4434 *) Add support for overriding the generation of SSL/TLS session IDs.
4435 These callbacks can be registered either in an SSL_CTX or per SSL.
4436 The purpose of this is to allow applications to control, if they wish,
4437 the arbitrary values chosen for use as session IDs, particularly as it
4438 can be useful for session caching in multiple-server environments. A
4439 command-line switch for testing this (and any client code that wishes
4440 to use such a feature) has been added to "s_server".
4441 [Geoff Thorpe, Lutz Jaenicke]
4443 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
4444 of the form '#if defined(...) || defined(...) || ...' and
4445 '#if !defined(...) && !defined(...) && ...'. This also avoids
4446 the growing number of special cases it was previously handling.
4449 *) Make all configuration macros available for application by making
4450 sure they are available in opensslconf.h, by giving them names starting
4451 with "OPENSSL_" to avoid conflicts with other packages and by making
4452 sure e_os2.h will cover all platform-specific cases together with
4454 Additionally, it is now possible to define configuration/platform-
4455 specific names (called "system identities"). In the C code, these
4456 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
4457 macro with the name beginning with "OPENSSL_SYS_", which is determined
4458 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
4462 *) New option -set_serial to 'req' and 'x509' this allows the serial
4463 number to use to be specified on the command line. Previously self
4464 signed certificates were hard coded with serial number 0 and the
4465 CA options of 'x509' had to use a serial number in a file which was
4469 *) New options to 'ca' utility to support V2 CRL entry extensions.
4470 Currently CRL reason, invalidity date and hold instruction are
4471 supported. Add new CRL extensions to V3 code and some new objects.
4474 *) New function EVP_CIPHER_CTX_set_padding() this is used to
4475 disable standard block padding (aka PKCS#5 padding) in the EVP
4476 API, which was previously mandatory. This means that the data is
4477 not padded in any way and so the total length much be a multiple
4478 of the block size, otherwise an error occurs.
4481 *) Initial (incomplete) OCSP SSL support.
4484 *) New function OCSP_parse_url(). This splits up a URL into its host,
4485 port and path components: primarily to parse OCSP URLs. New -url
4486 option to ocsp utility.
4489 *) New nonce behavior. The return value of OCSP_check_nonce() now
4490 reflects the various checks performed. Applications can decide
4491 whether to tolerate certain situations such as an absent nonce
4492 in a response when one was present in a request: the ocsp application
4493 just prints out a warning. New function OCSP_add1_basic_nonce()
4494 this is to allow responders to include a nonce in a response even if
4495 the request is nonce-less.
4498 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
4499 skipped when using openssl x509 multiple times on a single input file,
4500 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
4503 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
4504 set string type: to handle setting ASN1_TIME structures. Fix ca
4505 utility to correctly initialize revocation date of CRLs.
4508 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
4509 the clients preferred ciphersuites and rather use its own preferences.
4510 Should help to work around M$ SGC (Server Gated Cryptography) bug in
4511 Internet Explorer by ensuring unchanged hash method during stepup.
4512 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
4515 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
4516 to aes and add a new 'exist' option to print out symbols that don't
4520 *) Additional options to ocsp utility to allow flags to be set and
4521 additional certificates supplied.
4524 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
4525 OCSP client a number of certificate to only verify the response
4529 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
4530 handle the new API. Currently only ECB, CBC modes supported. Add new
4533 Add TLS AES ciphersuites as described in RFC3268, "Advanced
4534 Encryption Standard (AES) Ciphersuites for Transport Layer
4535 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
4536 not enabled by default and were not part of the "ALL" ciphersuite
4537 alias because they were not yet official; they could be
4538 explicitly requested by specifying the "AESdraft" ciphersuite
4539 group alias. In the final release of OpenSSL 0.9.7, the group
4540 alias is called "AES" and is part of "ALL".)
4541 [Ben Laurie, Steve Henson, Bodo Moeller]
4543 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
4544 request to response.
4547 *) Functions for OCSP responders. OCSP_request_onereq_count(),
4548 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
4549 extract information from a certificate request. OCSP_response_create()
4550 creates a response and optionally adds a basic response structure.
4551 OCSP_basic_add1_status() adds a complete single response to a basic
4552 response and returns the OCSP_SINGLERESP structure just added (to allow
4553 extensions to be included for example). OCSP_basic_add1_cert() adds a
4554 certificate to a basic response and OCSP_basic_sign() signs a basic
4555 response with various flags. New helper functions ASN1_TIME_check()
4556 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
4557 (converts ASN1_TIME to GeneralizedTime).
4560 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
4561 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
4562 structure from a certificate. X509_pubkey_digest() digests the public_key
4563 contents: this is used in various key identifiers.
4566 *) Make sk_sort() tolerate a NULL argument.
4567 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
4569 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
4570 passed by the function are trusted implicitly. If any of them signed the
4571 response then it is assumed to be valid and is not verified.
4574 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
4575 to data. This was previously part of the PKCS7 ASN1 code. This
4576 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
4577 [Steve Henson, reported by Kenneth R. Robinette
4578 <support@securenetterm.com>]
4580 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
4581 routines: without these tracing memory leaks is very painful.
4582 Fix leaks in PKCS12 and PKCS7 routines.
4585 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
4586 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
4587 effectively meant GeneralizedTime would never be used. Now it
4588 is initialised to -1 but X509_time_adj() now has to check the value
4589 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
4590 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
4591 [Steve Henson, reported by Kenneth R. Robinette
4592 <support@securenetterm.com>]
4594 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
4595 result in a zero length in the ASN1_INTEGER structure which was
4596 not consistent with the structure when d2i_ASN1_INTEGER() was used
4597 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
4598 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
4599 where it did not print out a minus for negative ASN1_INTEGER.
4602 *) Add summary printout to ocsp utility. The various functions which
4603 convert status values to strings have been renamed to:
4604 OCSP_response_status_str(), OCSP_cert_status_str() and
4605 OCSP_crl_reason_str() and are no longer static. New options
4606 to verify nonce values and to disable verification. OCSP response
4607 printout format cleaned up.
4610 *) Add additional OCSP certificate checks. These are those specified
4611 in RFC2560. This consists of two separate checks: the CA of the
4612 certificate being checked must either be the OCSP signer certificate
4613 or the issuer of the OCSP signer certificate. In the latter case the
4614 OCSP signer certificate must contain the OCSP signing extended key
4615 usage. This check is performed by attempting to match the OCSP
4616 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
4617 in the OCSP_CERTID structures of the response.
4620 *) Initial OCSP certificate verification added to OCSP_basic_verify()
4621 and related routines. This uses the standard OpenSSL certificate
4622 verify routines to perform initial checks (just CA validity) and
4623 to obtain the certificate chain. Then additional checks will be
4624 performed on the chain. Currently the root CA is checked to see
4625 if it is explicitly trusted for OCSP signing. This is used to set
4626 a root CA as a global signing root: that is any certificate that
4627 chains to that CA is an acceptable OCSP signing certificate.
4630 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
4631 extensions from a separate configuration file.
4632 As when reading extensions from the main configuration file,
4633 the '-extensions ...' option may be used for specifying the
4635 [Massimiliano Pala <madwolf@comune.modena.it>]
4637 *) New OCSP utility. Allows OCSP requests to be generated or
4638 read. The request can be sent to a responder and the output
4639 parsed, outputed or printed in text form. Not complete yet:
4640 still needs to check the OCSP response validity.
4643 *) New subcommands for 'openssl ca':
4644 'openssl ca -status <serial>' prints the status of the cert with
4645 the given serial number (according to the index file).
4646 'openssl ca -updatedb' updates the expiry status of certificates
4648 [Massimiliano Pala <madwolf@comune.modena.it>]
4650 *) New '-newreq-nodes' command option to CA.pl. This is like
4651 '-newreq', but calls 'openssl req' with the '-nodes' option
4652 so that the resulting key is not encrypted.
4653 [Damien Miller <djm@mindrot.org>]
4655 *) New configuration for the GNU Hurd.
4656 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
4658 *) Initial code to implement OCSP basic response verify. This
4659 is currently incomplete. Currently just finds the signer's
4660 certificate and verifies the signature on the response.
4663 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
4664 value of OPENSSLDIR. This is available via the new '-d' option
4665 to 'openssl version', and is also included in 'openssl version -a'.
4668 *) Allowing defining memory allocation callbacks that will be given
4669 file name and line number information in additional arguments
4670 (a const char* and an int). The basic functionality remains, as
4671 well as the original possibility to just replace malloc(),
4672 realloc() and free() by functions that do not know about these
4673 additional arguments. To register and find out the current
4674 settings for extended allocation functions, the following
4675 functions are provided:
4677 CRYPTO_set_mem_ex_functions
4678 CRYPTO_set_locked_mem_ex_functions
4679 CRYPTO_get_mem_ex_functions
4680 CRYPTO_get_locked_mem_ex_functions
4682 These work the same way as CRYPTO_set_mem_functions and friends.
4683 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
4684 extended allocation function is enabled.
4685 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
4686 a conventional allocation function is enabled.
4687 [Richard Levitte, Bodo Moeller]
4689 *) Finish off removing the remaining LHASH function pointer casts.
4690 There should no longer be any prototype-casting required when using
4691 the LHASH abstraction, and any casts that remain are "bugs". See
4692 the callback types and macros at the head of lhash.h for details
4693 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
4696 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
4697 If /dev/[u]random devices are not available or do not return enough
4698 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
4700 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
4701 /etc/entropy will be queried once each in this sequence, quering stops
4702 when enough entropy was collected without querying more sockets.
4705 *) Change the Unix RAND_poll() variant to be able to poll several
4706 random devices, as specified by DEVRANDOM, until a sufficient amount
4707 of data has been collected. We spend at most 10 ms on each file
4708 (select timeout) and read in non-blocking mode. DEVRANDOM now
4709 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
4710 (previously it was just the string "/dev/urandom"), so on typical
4711 platforms the 10 ms delay will never occur.
4712 Also separate out the Unix variant to its own file, rand_unix.c.
4713 For VMS, there's a currently-empty rand_vms.c.
4716 *) Move OCSP client related routines to ocsp_cl.c. These
4717 provide utility functions which an application needing
4718 to issue a request to an OCSP responder and analyse the
4719 response will typically need: as opposed to those which an
4720 OCSP responder itself would need which will be added later.
4722 OCSP_request_sign() signs an OCSP request with an API similar
4723 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
4724 response. OCSP_response_get1_basic() extracts basic response
4725 from response. OCSP_resp_find_status(): finds and extracts status
4726 information from an OCSP_CERTID structure (which will be created
4727 when the request structure is built). These are built from lower
4728 level functions which work on OCSP_SINGLERESP structures but
4729 wont normally be used unless the application wishes to examine
4730 extensions in the OCSP response for example.
4732 Replace nonce routines with a pair of functions.
4733 OCSP_request_add1_nonce() adds a nonce value and optionally
4734 generates a random value. OCSP_check_nonce() checks the
4735 validity of the nonce in an OCSP response.
4738 *) Change function OCSP_request_add() to OCSP_request_add0_id().
4739 This doesn't copy the supplied OCSP_CERTID and avoids the
4740 need to free up the newly created id. Change return type
4741 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
4742 This can then be used to add extensions to the request.
4743 Deleted OCSP_request_new(), since most of its functionality
4744 is now in OCSP_REQUEST_new() (and the case insensitive name
4745 clash) apart from the ability to set the request name which
4746 will be added elsewhere.
4749 *) Update OCSP API. Remove obsolete extensions argument from
4750 various functions. Extensions are now handled using the new
4751 OCSP extension code. New simple OCSP HTTP function which
4752 can be used to send requests and parse the response.
4755 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
4756 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
4757 uses the special reorder version of SET OF to sort the attributes
4758 and reorder them to match the encoded order. This resolves a long
4759 standing problem: a verify on a PKCS7 structure just after signing
4760 it used to fail because the attribute order did not match the
4761 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
4762 it uses the received order. This is necessary to tolerate some broken
4763 software that does not order SET OF. This is handled by encoding
4764 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
4765 to produce the required SET OF.
4768 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
4769 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
4770 files to get correct declarations of the ASN.1 item variables.
4773 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
4774 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
4775 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
4776 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
4777 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
4778 ASN1_ITEM and no wrapper functions.
4781 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
4782 replace the old function pointer based I/O routines. Change most of
4783 the *_d2i_bio() and *_d2i_fp() functions to use these.
4786 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
4787 lines, recognice more "algorithms" that can be deselected, and make
4788 it complain about algorithm deselection that isn't recognised.
4791 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
4792 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
4793 to use new functions. Add NO_ASN1_OLD which can be set to remove
4794 some old style ASN1 functions: this can be used to determine if old
4795 code will still work when these eventually go away.
4798 *) New extension functions for OCSP structures, these follow the
4799 same conventions as certificates and CRLs.
4802 *) New function X509V3_add1_i2d(). This automatically encodes and
4803 adds an extension. Its behaviour can be customised with various
4804 flags to append, replace or delete. Various wrappers added for
4805 certifcates and CRLs.
4808 *) Fix to avoid calling the underlying ASN1 print routine when
4809 an extension cannot be parsed. Correct a typo in the
4810 OCSP_SERVICELOC extension. Tidy up print OCSP format.
4813 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
4814 entries for variables.
4817 *) Add functionality to apps/openssl.c for detecting locking
4818 problems: As the program is single-threaded, all we have
4819 to do is register a locking callback using an array for
4820 storing which locks are currently held by the program.
4823 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
4824 SSL_get_ex_data_X509_STORE_idx(), which is used in
4825 ssl_verify_cert_chain() and thus can be called at any time
4826 during TLS/SSL handshakes so that thread-safety is essential.
4827 Unfortunately, the ex_data design is not at all suited
4828 for multi-threaded use, so it probably should be abolished.
4831 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
4832 [Broadcom, tweaked and integrated by Geoff Thorpe]
4834 *) Move common extension printing code to new function
4835 X509V3_print_extensions(). Reorganise OCSP print routines and
4836 implement some needed OCSP ASN1 functions. Add OCSP extensions.
4839 *) New function X509_signature_print() to remove duplication in some
4843 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
4844 set (this was treated exactly the same as SET OF previously). This
4845 is used to reorder the STACK representing the structure to match the
4846 encoding. This will be used to get round a problem where a PKCS7
4847 structure which was signed could not be verified because the STACK
4848 order did not reflect the encoded order.
4851 *) Reimplement the OCSP ASN1 module using the new code.
4854 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
4855 for its ASN1 operations. The old style function pointers still exist
4856 for now but they will eventually go away.
4859 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
4860 completely replaces the old ASN1 functionality with a table driven
4861 encoder and decoder which interprets an ASN1_ITEM structure describing
4862 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
4863 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
4864 has also been converted to the new form.
4867 *) Change BN_mod_exp_recp so that negative moduli are tolerated
4868 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
4869 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
4870 for negative moduli.
4873 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
4874 of not touching the result's sign bit.
4877 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
4881 *) Changed the LHASH code to use prototypes for callbacks, and created
4882 macros to declare and implement thin (optionally static) functions
4883 that provide type-safety and avoid function pointer casting for the
4884 type-specific callbacks.
4887 *) Added Kerberos Cipher Suites to be used with TLS, as written in
4889 [Veers Staats <staatsvr@asc.hpc.mil>,
4890 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
4892 *) Reformat the FAQ so the different questions and answers can be divided
4893 in sections depending on the subject.
4896 *) Have the zlib compression code load ZLIB.DLL dynamically under
4900 *) New function BN_mod_sqrt for computing square roots modulo a prime
4901 (using the probabilistic Tonelli-Shanks algorithm unless
4902 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
4903 be handled deterministically).
4904 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
4906 *) Make BN_mod_inverse faster by explicitly handling small quotients
4907 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
4908 512 bits], about 30% for larger ones [1024 or 2048 bits].)
4911 *) New function BN_kronecker.
4914 *) Fix BN_gcd so that it works on negative inputs; the result is
4915 positive unless both parameters are zero.
4916 Previously something reasonably close to an infinite loop was
4917 possible because numbers could be growing instead of shrinking
4918 in the implementation of Euclid's algorithm.
4921 *) Fix BN_is_word() and BN_is_one() macros to take into account the
4922 sign of the number in question.
4924 Fix BN_is_word(a,w) to work correctly for w == 0.
4926 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
4927 because its test if the absolute value of 'a' equals 'w'.
4928 Note that BN_abs_is_word does *not* handle w == 0 reliably;
4929 it exists mostly for use in the implementations of BN_is_zero(),
4930 BN_is_one(), and BN_is_word().
4933 *) New function BN_swap.
4936 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
4937 the exponentiation functions are more likely to produce reasonable
4938 results on negative inputs.
4941 *) Change BN_mod_mul so that the result is always non-negative.
4942 Previously, it could be negative if one of the factors was negative;
4943 I don't think anyone really wanted that behaviour.
4946 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
4947 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
4948 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
4949 and add new functions:
4958 BN_mod_lshift1_quick
4962 These functions always generate non-negative results.
4964 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
4965 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
4967 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
4968 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
4969 be reduced modulo m.
4970 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
4973 The following entry accidentily appeared in the CHANGES file
4974 distributed with OpenSSL 0.9.7. The modifications described in
4975 it do *not* apply to OpenSSL 0.9.7.
4977 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4978 was actually never needed) and in BN_mul(). The removal in BN_mul()
4979 required a small change in bn_mul_part_recursive() and the addition
4980 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4981 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4982 bn_sub_words() and bn_add_words() except they take arrays with
4987 *) In 'openssl passwd', verify passwords read from the terminal
4988 unless the '-salt' option is used (which usually means that
4989 verification would just waste user's time since the resulting
4990 hash is going to be compared with some given password hash)
4991 or the new '-noverify' option is used.
4993 This is an incompatible change, but it does not affect
4994 non-interactive use of 'openssl passwd' (passwords on the command
4995 line, '-stdin' option, '-in ...' option) and thus should not
4999 *) Remove all references to RSAref, since there's no more need for it.
5002 *) Make DSO load along a path given through an environment variable
5003 (SHLIB_PATH) with shl_load().
5006 *) Constify the ENGINE code as a result of BIGNUM constification.
5007 Also constify the RSA code and most things related to it. In a
5008 few places, most notable in the depth of the ASN.1 code, ugly
5009 casts back to non-const were required (to be solved at a later
5013 *) Make it so the openssl application has all engines loaded by default.
5016 *) Constify the BIGNUM routines a little more.
5019 *) Add the following functions:
5021 ENGINE_load_cswift()
5023 ENGINE_load_atalla()
5025 ENGINE_load_builtin_engines()
5027 That way, an application can itself choose if external engines that
5028 are built-in in OpenSSL shall ever be used or not. The benefit is
5029 that applications won't have to be linked with libdl or other dso
5030 libraries unless it's really needed.
5032 Changed 'openssl engine' to load all engines on demand.
5033 Changed the engine header files to avoid the duplication of some
5034 declarations (they differed!).
5037 *) 'openssl engine' can now list capabilities.
5040 *) Better error reporting in 'openssl engine'.
5043 *) Never call load_dh_param(NULL) in s_server.
5046 *) Add engine application. It can currently list engines by name and
5047 identity, and test if they are actually available.
5050 *) Improve RPM specification file by forcing symbolic linking and making
5051 sure the installed documentation is also owned by root.root.
5052 [Damien Miller <djm@mindrot.org>]
5054 *) Give the OpenSSL applications more possibilities to make use of
5055 keys (public as well as private) handled by engines.
5058 *) Add OCSP code that comes from CertCo.
5061 *) Add VMS support for the Rijndael code.
5064 *) Added untested support for Nuron crypto accelerator.
5067 *) Add support for external cryptographic devices. This code was
5068 previously distributed separately as the "engine" branch.
5069 [Geoff Thorpe, Richard Levitte]
5071 *) Rework the filename-translation in the DSO code. It is now possible to
5072 have far greater control over how a "name" is turned into a filename
5073 depending on the operating environment and any oddities about the
5074 different shared library filenames on each system.
5077 *) Support threads on FreeBSD-elf in Configure.
5080 *) Fix for SHA1 assembly problem with MASM: it produces
5081 warnings about corrupt line number information when assembling
5082 with debugging information. This is caused by the overlapping
5084 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5087 NCONF_get_number() has no error checking at all. As a replacement,
5088 NCONF_get_number_e() is defined (_e for "error checking") and is
5089 promoted strongly. The old NCONF_get_number is kept around for
5090 binary backward compatibility.
5091 Make it possible for methods to load from something other than a BIO,
5092 by providing a function pointer that is given a name instead of a BIO.
5093 For example, this could be used to load configuration data from an
5097 *) Fix for non blocking accept BIOs. Added new I/O special reason
5098 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5099 with non blocking I/O was not possible because no retry code was
5100 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5104 *) Added the beginnings of Rijndael support.
5107 *) Fix for bug in DirectoryString mask setting. Add support for
5108 X509_NAME_print_ex() in 'req' and X509_print_ex() function
5109 to allow certificate printing to more controllable, additional
5110 'certopt' option to 'x509' to allow new printing options to be
5114 *) Clean old EAY MD5 hack from e_os.h.
5117 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5119 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5120 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5121 [Joe Orton, Steve Henson]
5123 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5125 *) Fix additional bug revealed by the NISCC test suite:
5127 Stop bug triggering large recursion when presented with
5128 certain ASN.1 tags (CVE-2003-0851)
5131 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5133 *) Fix various bugs revealed by running the NISCC test suite:
5135 Stop out of bounds reads in the ASN1 code when presented with
5136 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5138 If verify callback ignores invalid public key errors don't try to check
5139 certificate signature with the NULL public key.
5143 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5144 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5148 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5149 extra data after the compression methods not only for TLS 1.0
5150 but also for SSL 3.0 (as required by the specification).
5151 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5153 *) Change X509_certificate_type() to mark the key as exported/exportable
5154 when it's 512 *bits* long, not 512 bytes.
5157 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5159 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5160 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5161 a protocol version number mismatch like a decryption error
5162 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5165 *) Turn on RSA blinding by default in the default implementation
5166 to avoid a timing attack. Applications that don't want it can call
5167 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5168 They would be ill-advised to do so in most cases.
5169 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5171 *) Change RSA blinding code so that it works when the PRNG is not
5172 seeded (in this case, the secret RSA exponent is abused as
5173 an unpredictable seed -- if it is not unpredictable, there
5174 is no point in blinding anyway). Make RSA blinding thread-safe
5175 by remembering the creator's thread ID in rsa->blinding and
5176 having all other threads use local one-time blinding factors
5177 (this requires more computation than sharing rsa->blinding, but
5178 avoids excessive locking; and if an RSA object is not shared
5179 between threads, blinding will still be very fast).
5182 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5184 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5185 via timing by performing a MAC computation even if incorrrect
5186 block cipher padding has been found. This is a countermeasure
5187 against active attacks where the attacker has to distinguish
5188 between bad padding and a MAC verification error. (CVE-2003-0078)
5190 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5191 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5192 Martin Vuagnoux (EPFL, Ilion)]
5194 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
5196 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
5197 memory from it's contents. This is done with a counter that will
5198 place alternating values in each byte. This can be used to solve
5199 two issues: 1) the removal of calls to memset() by highly optimizing
5200 compilers, and 2) cleansing with other values than 0, since those can
5201 be read through on certain media, for example a swap space on disk.
5204 *) Bugfix: client side session caching did not work with external caching,
5205 because the session->cipher setting was not restored when reloading
5206 from the external cache. This problem was masked, when
5207 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
5208 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
5211 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
5212 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
5213 [Zeev Lieber <zeev-l@yahoo.com>]
5215 *) Undo an undocumented change introduced in 0.9.6e which caused
5216 repeated calls to OpenSSL_add_all_ciphers() and
5217 OpenSSL_add_all_digests() to be ignored, even after calling
5221 *) Change the default configuration reader to deal with last line not
5222 being properly terminated.
5225 *) Change X509_NAME_cmp() so it applies the special rules on handling
5226 DN values that are of type PrintableString, as well as RDNs of type
5227 emailAddress where the value has the type ia5String.
5228 [stefank@valicert.com via Richard Levitte]
5230 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
5231 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
5232 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
5233 the bitwise-OR of the two for use by the majority of applications
5234 wanting this behaviour, and update the docs. The documented
5235 behaviour and actual behaviour were inconsistent and had been
5236 changing anyway, so this is more a bug-fix than a behavioural
5238 [Geoff Thorpe, diagnosed by Nadav Har'El]
5240 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
5241 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
5244 *) Fix initialization code race conditions in
5245 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
5246 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
5247 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
5248 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
5249 ssl2_get_cipher_by_char(),
5250 ssl3_get_cipher_by_char().
5251 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
5253 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
5254 the cached sessions are flushed, as the remove_cb() might use ex_data
5255 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
5256 (see [openssl.org #212]).
5257 [Geoff Thorpe, Lutz Jaenicke]
5259 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
5260 length, instead of the encoding length to d2i_ASN1_OBJECT.
5263 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
5265 *) [In 0.9.6g-engine release:]
5266 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
5267 [Lynn Gazis <lgazis@rainbow.com>]
5269 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
5271 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
5272 and get fix the header length calculation.
5273 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
5274 Alon Kantor <alonk@checkpoint.com> (and others),
5277 *) Use proper error handling instead of 'assertions' in buffer
5278 overflow checks added in 0.9.6e. This prevents DoS (the
5279 assertions could call abort()).
5280 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
5282 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
5284 *) Add various sanity checks to asn1_get_length() to reject
5285 the ASN1 length bytes if they exceed sizeof(long), will appear
5286 negative or the content length exceeds the length of the
5288 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5290 *) Fix cipher selection routines: ciphers without encryption had no flags
5291 for the cipher strength set and where therefore not handled correctly
5292 by the selection routines (PR #130).
5295 *) Fix EVP_dsa_sha macro.
5299 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
5300 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
5301 that was added in OpenSSL 0.9.6d.
5303 As the countermeasure turned out to be incompatible with some
5304 broken SSL implementations, the new option is part of SSL_OP_ALL.
5305 SSL_OP_ALL is usually employed when compatibility with weird SSL
5306 implementations is desired (e.g. '-bugs' option to 's_client' and
5307 's_server'), so the new option is automatically set in many
5311 *) Changes in security patch:
5313 Changes marked "(CHATS)" were sponsored by the Defense Advanced
5314 Research Projects Agency (DARPA) and Air Force Research Laboratory,
5315 Air Force Materiel Command, USAF, under agreement number
5318 *) Add various sanity checks to asn1_get_length() to reject
5319 the ASN1 length bytes if they exceed sizeof(long), will appear
5320 negative or the content length exceeds the length of the
5321 supplied buffer. (CVE-2002-0659)
5322 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5324 *) Assertions for various potential buffer overflows, not known to
5326 [Ben Laurie (CHATS)]
5328 *) Various temporary buffers to hold ASCII versions of integers were
5329 too small for 64 bit platforms. (CVE-2002-0655)
5330 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
5332 *) Remote buffer overflow in SSL3 protocol - an attacker could
5333 supply an oversized session ID to a client. (CVE-2002-0656)
5334 [Ben Laurie (CHATS)]
5336 *) Remote buffer overflow in SSL2 protocol - an attacker could
5337 supply an oversized client master key. (CVE-2002-0656)
5338 [Ben Laurie (CHATS)]
5340 Changes between 0.9.6c and 0.9.6d [9 May 2002]
5342 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
5343 encoded as NULL) with id-dsa-with-sha1.
5344 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
5346 *) Check various X509_...() return values in apps/req.c.
5347 [Nils Larsch <nla@trustcenter.de>]
5349 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
5350 an end-of-file condition would erronously be flagged, when the CRLF
5351 was just at the end of a processed block. The bug was discovered when
5352 processing data through a buffering memory BIO handing the data to a
5353 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
5354 <ptsekov@syntrex.com> and Nedelcho Stanev.
5357 *) Implement a countermeasure against a vulnerability recently found
5358 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
5359 before application data chunks to avoid the use of known IVs
5360 with data potentially chosen by the attacker.
5363 *) Fix length checks in ssl3_get_client_hello().
5366 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
5367 to prevent ssl3_read_internal() from incorrectly assuming that
5368 ssl3_read_bytes() found application data while handshake
5369 processing was enabled when in fact s->s3->in_read_app_data was
5370 merely automatically cleared during the initial handshake.
5371 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
5373 *) Fix object definitions for Private and Enterprise: they were not
5374 recognized in their shortname (=lowercase) representation. Extend
5375 obj_dat.pl to issue an error when using undefined keywords instead
5376 of silently ignoring the problem (Svenning Sorensen
5377 <sss@sss.dnsalias.net>).
5380 *) Fix DH_generate_parameters() so that it works for 'non-standard'
5381 generators, i.e. generators other than 2 and 5. (Previously, the
5382 code did not properly initialise the 'add' and 'rem' values to
5383 BN_generate_prime().)
5385 In the new general case, we do not insist that 'generator' is
5386 actually a primitive root: This requirement is rather pointless;
5387 a generator of the order-q subgroup is just as good, if not
5391 *) Map new X509 verification errors to alerts. Discovered and submitted by
5392 Tom Wu <tom@arcot.com>.
5395 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
5396 returning non-zero before the data has been completely received
5397 when using non-blocking I/O.
5398 [Bodo Moeller; problem pointed out by John Hughes]
5400 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
5401 [Ben Laurie, Lutz Jaenicke]
5403 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
5404 Yoram Zahavi <YoramZ@gilian.com>).
5407 *) Add information about CygWin 1.3 and on, and preserve proper
5408 configuration for the versions before that.
5409 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5411 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
5412 check whether we deal with a copy of a session and do not delete from
5413 the cache in this case. Problem reported by "Izhar Shoshani Levi"
5414 <izhar@checkpoint.com>.
5417 *) Do not store session data into the internal session cache, if it
5418 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
5419 flag is set). Proposed by Aslam <aslam@funk.com>.
5422 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
5426 *) [In 0.9.6d-engine release:]
5427 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5428 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5430 *) Add the configuration target linux-s390x.
5431 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
5433 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
5434 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
5435 variable as an indication that a ClientHello message has been
5436 received. As the flag value will be lost between multiple
5437 invocations of ssl3_accept when using non-blocking I/O, the
5438 function may not be aware that a handshake has actually taken
5439 place, thus preventing a new session from being added to the
5442 To avoid this problem, we now set s->new_session to 2 instead of
5443 using a local variable.
5444 [Lutz Jaenicke, Bodo Moeller]
5446 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
5447 if the SSL_R_LENGTH_MISMATCH error is detected.
5448 [Geoff Thorpe, Bodo Moeller]
5450 *) New 'shared_ldflag' column in Configure platform table.
5453 *) Fix EVP_CIPHER_mode macro.
5454 ["Dan S. Camper" <dan@bti.net>]
5456 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
5457 type, we must throw them away by setting rr->length to 0.
5458 [D P Chang <dpc@qualys.com>]
5460 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
5462 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
5463 <Dominikus.Scherkl@biodata.com>. (The previous implementation
5464 worked incorrectly for those cases where range = 10..._2 and
5465 3*range is two bits longer than range.)
5468 *) Only add signing time to PKCS7 structures if it is not already
5472 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
5473 OBJ_ld_ce should be OBJ_id_ce.
5474 Also some ip-pda OIDs in crypto/objects/objects.txt were
5475 incorrect (cf. RFC 3039).
5476 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
5478 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
5479 returns early because it has nothing to do.
5480 [Andy Schneider <andy.schneider@bjss.co.uk>]
5482 *) [In 0.9.6c-engine release:]
5483 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
5484 [Andy Schneider <andy.schneider@bjss.co.uk>]
5486 *) [In 0.9.6c-engine release:]
5487 Add support for Cryptographic Appliance's keyserver technology.
5488 (Use engine 'keyclient')
5489 [Cryptographic Appliances and Geoff Thorpe]
5491 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
5492 is called via tools/c89.sh because arguments have to be
5493 rearranged (all '-L' options must appear before the first object
5495 [Richard Shapiro <rshapiro@abinitio.com>]
5497 *) [In 0.9.6c-engine release:]
5498 Add support for Broadcom crypto accelerator cards, backported
5500 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
5502 *) [In 0.9.6c-engine release:]
5503 Add support for SureWare crypto accelerator cards from
5504 Baltimore Technologies. (Use engine 'sureware')
5505 [Baltimore Technologies and Mark Cox]
5507 *) [In 0.9.6c-engine release:]
5508 Add support for crypto accelerator cards from Accelerated
5509 Encryption Processing, www.aep.ie. (Use engine 'aep')
5510 [AEP Inc. and Mark Cox]
5512 *) Add a configuration entry for gcc on UnixWare.
5513 [Gary Benson <gbenson@redhat.com>]
5515 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
5516 messages are stored in a single piece (fixed-length part and
5517 variable-length part combined) and fix various bugs found on the way.
5520 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
5521 instead. BIO_gethostbyname() does not know what timeouts are
5522 appropriate, so entries would stay in cache even when they have
5524 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
5526 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
5527 faced with a pathologically small ClientHello fragment that does
5528 not contain client_version: Instead of aborting with an error,
5529 simply choose the highest available protocol version (i.e.,
5530 TLS 1.0 unless it is disabled). In practice, ClientHello
5531 messages are never sent like this, but this change gives us
5532 strictly correct behaviour at least for TLS.
5535 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
5536 never resets s->method to s->ctx->method when called from within
5537 one of the SSL handshake functions.
5538 [Bodo Moeller; problem pointed out by Niko Baric]
5540 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
5541 (sent using the client's version number) if client_version is
5542 smaller than the protocol version in use. Also change
5543 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
5544 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
5545 the client will at least see that alert.
5548 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
5552 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
5553 client receives HelloRequest while in a handshake.
5554 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
5556 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
5557 should end in 'break', not 'goto end' which circuments various
5558 cleanups done in state SSL_ST_OK. But session related stuff
5559 must be disabled for SSL_ST_OK in the case that we just sent a
5562 Also avoid some overhead by not calling ssl_init_wbio_buffer()
5563 before just sending a HelloRequest.
5564 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
5566 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
5567 reveal whether illegal block cipher padding was found or a MAC
5568 verification error occured. (Neither SSLerr() codes nor alerts
5569 are directly visible to potential attackers, but the information
5570 may leak via logfiles.)
5572 Similar changes are not required for the SSL 2.0 implementation
5573 because the number of padding bytes is sent in clear for SSL 2.0,
5574 and the extra bytes are just ignored. However ssl/s2_pkt.c
5575 failed to verify that the purported number of padding bytes is in
5579 *) Add OpenUNIX-8 support including shared libraries
5580 (Boyd Lynn Gerber <gerberb@zenez.com>).
5583 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
5584 'wristwatch attack' using huge encoding parameters (cf.
5585 James H. Manger's CRYPTO 2001 paper). Note that the
5586 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
5587 encoding parameters and hence was not vulnerable.
5590 *) BN_sqr() bug fix.
5591 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
5593 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
5594 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
5595 followed by modular reduction.
5596 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
5598 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
5599 equivalent based on BN_pseudo_rand() instead of BN_rand().
5602 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
5603 This function was broken, as the check for a new client hello message
5604 to handle SGC did not allow these large messages.
5605 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
5608 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
5611 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
5612 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
5615 *) Rework the configuration and shared library support for Tru64 Unix.
5616 The configuration part makes use of modern compiler features and
5617 still retains old compiler behavior for those that run older versions
5618 of the OS. The shared library support part includes a variant that
5619 uses the RPATH feature, and is available through the special
5620 configuration target "alpha-cc-rpath", which will never be selected
5622 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
5624 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
5625 with the same message size as in ssl3_get_certificate_request().
5626 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
5627 messages might inadvertently be reject as too long.
5628 [Petr Lampa <lampa@fee.vutbr.cz>]
5630 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
5633 *) Modified SSL library such that the verify_callback that has been set
5634 specificly for an SSL object with SSL_set_verify() is actually being
5635 used. Before the change, a verify_callback set with this function was
5636 ignored and the verify_callback() set in the SSL_CTX at the time of
5637 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
5638 to allow the necessary settings.
5641 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
5642 explicitly to NULL, as at least on Solaris 8 this seems not always to be
5643 done automatically (in contradiction to the requirements of the C
5644 standard). This made problems when used from OpenSSH.
5647 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
5648 dh->length and always used
5650 BN_rand_range(priv_key, dh->p).
5652 BN_rand_range() is not necessary for Diffie-Hellman, and this
5653 specific range makes Diffie-Hellman unnecessarily inefficient if
5654 dh->length (recommended exponent length) is much smaller than the
5655 length of dh->p. We could use BN_rand_range() if the order of
5656 the subgroup was stored in the DH structure, but we only have
5661 BN_rand(priv_key, l, ...)
5663 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
5669 RSA_eay_public_encrypt
5670 RSA_eay_private_decrypt
5671 RSA_eay_private_encrypt (signing)
5672 RSA_eay_public_decrypt (signature verification)
5674 (default implementations for RSA_public_encrypt,
5675 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
5676 always reject numbers >= n.
5679 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
5680 to synchronize access to 'locking_thread'. This is necessary on
5681 systems where access to 'locking_thread' (an 'unsigned long'
5682 variable) is not atomic.
5685 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
5686 *before* setting the 'crypto_lock_rand' flag. The previous code had
5687 a race condition if 0 is a valid thread ID.
5688 [Travis Vitek <vitek@roguewave.com>]
5690 *) Add support for shared libraries under Irix.
5691 [Albert Chin-A-Young <china@thewrittenword.com>]
5693 *) Add configuration option to build on Linux on both big-endian and
5695 [Ralf Baechle <ralf@uni-koblenz.de>]
5697 *) Add the possibility to create shared libraries on HP-UX.
5700 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
5702 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
5703 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
5704 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
5705 PRNG state recovery was possible based on the output of
5706 one PRNG request appropriately sized to gain knowledge on
5707 'md' followed by enough consecutive 1-byte PRNG requests
5708 to traverse all of 'state'.
5710 1. When updating 'md_local' (the current thread's copy of 'md')
5711 during PRNG output generation, hash all of the previous
5712 'md_local' value, not just the half used for PRNG output.
5714 2. Make the number of bytes from 'state' included into the hash
5715 independent from the number of PRNG bytes requested.
5717 The first measure alone would be sufficient to avoid
5718 Markku-Juhani's attack. (Actually it had never occurred
5719 to me that the half of 'md_local' used for chaining was the
5720 half from which PRNG output bytes were taken -- I had always
5721 assumed that the secret half would be used.) The second
5722 measure makes sure that additional data from 'state' is never
5723 mixed into 'md_local' in small portions; this heuristically
5724 further strengthens the PRNG.
5727 *) Fix crypto/bn/asm/mips3.s.
5730 *) When only the key is given to "enc", the IV is undefined. Print out
5731 an error message in this case.
5734 *) Handle special case when X509_NAME is empty in X509 printing routines.
5737 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
5738 positive and less than q.
5741 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
5742 used: it isn't thread safe and the add_lock_callback should handle
5744 [Paul Rose <Paul.Rose@bridge.com>]
5746 *) Verify that incoming data obeys the block size in
5747 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
5751 [Ulf Möller, Bodo Möller]
5753 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
5754 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
5755 when fixing the server behaviour for backwards-compatible 'client
5756 hello' messages. (Note that the attack is impractical against
5757 SSL 3.0 and TLS 1.0 anyway because length and version checking
5758 means that the probability of guessing a valid ciphertext is
5759 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
5762 Before 0.9.5, the countermeasure (hide the error by generating a
5763 random 'decryption result') did not work properly because
5764 ERR_clear_error() was missing, meaning that SSL_get_error() would
5765 detect the supposedly ignored error.
5767 Both problems are now fixed.
5770 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
5771 (previously it was 1024).
5774 *) Fix for compatibility mode trust settings: ignore trust settings
5775 unless some valid trust or reject settings are present.
5778 *) Fix for blowfish EVP: its a variable length cipher.
5781 *) Fix various bugs related to DSA S/MIME verification. Handle missing
5782 parameters in DSA public key structures and return an error in the
5783 DSA routines if parameters are absent.
5786 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
5787 in the current directory if neither $RANDFILE nor $HOME was set.
5788 RAND_file_name() in 0.9.6a returned NULL in this case. This has
5789 caused some confusion to Windows users who haven't defined $HOME.
5790 Thus RAND_file_name() is changed again: e_os.h can define a
5791 DEFAULT_HOME, which will be used if $HOME is not set.
5792 For Windows, we use "C:"; on other platforms, we still require
5793 environment variables.
5795 *) Move 'if (!initialized) RAND_poll()' into regions protected by
5796 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
5797 having multiple threads call RAND_poll() concurrently.
5800 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
5801 combination of a flag and a thread ID variable.
5802 Otherwise while one thread is in ssleay_rand_bytes (which sets the
5803 flag), *other* threads can enter ssleay_add_bytes without obeying
5804 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
5805 that they do not hold after the first thread unsets add_do_not_lock).
5808 *) Change bctest again: '-x' expressions are not available in all
5812 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
5814 *) Fix a couple of memory leaks in PKCS7_dataDecode()
5815 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
5817 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
5818 the default extension for executables, if any. Also, make the perl
5819 scripts that use symlink() to test if it really exists and use "cp"
5820 if it doesn't. All this made OpenSSL compilable and installable in
5824 *) Fix for asn1_GetSequence() for indefinite length constructed data.
5825 If SEQUENCE is length is indefinite just set c->slen to the total
5826 amount of data available.
5827 [Steve Henson, reported by shige@FreeBSD.org]
5828 [This change does not apply to 0.9.7.]
5830 *) Change bctest to avoid here-documents inside command substitution
5831 (workaround for FreeBSD /bin/sh bug).
5832 For compatibility with Ultrix, avoid shell functions (introduced
5833 in the bctest version that searches along $PATH).
5836 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
5837 with des_encrypt() defined on some operating systems, like Solaris
5841 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
5842 On the Importance of Eliminating Errors in Cryptographic
5843 Computations, J. Cryptology 14 (2001) 2, 101-119,
5844 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
5847 *) MIPS assembler BIGNUM division bug fix.
5850 *) Disabled incorrect Alpha assembler code.
5853 *) Fix PKCS#7 decode routines so they correctly update the length
5854 after reading an EOC for the EXPLICIT tag.
5856 [This change does not apply to 0.9.7.]
5858 *) Fix bug in PKCS#12 key generation routines. This was triggered
5859 if a 3DES key was generated with a 0 initial byte. Include
5860 PKCS12_BROKEN_KEYGEN compilation option to retain the old
5861 (but broken) behaviour.
5864 *) Enhance bctest to search for a working bc along $PATH and print
5866 [Tim Rice <tim@multitalents.net> via Richard Levitte]
5868 *) Fix memory leaks in err.c: free err_data string if necessary;
5869 don't write to the wrong index in ERR_set_error_data.
5872 *) Implement ssl23_peek (analogous to ssl23_read), which previously
5876 *) Replace rdtsc with _emit statements for VC++ version 5.
5877 [Jeremy Cooper <jeremy@baymoo.org>]
5879 *) Make it possible to reuse SSLv2 sessions.
5882 *) In copy_email() check for >= 0 as a return value for
5883 X509_NAME_get_index_by_NID() since 0 is a valid index.
5884 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
5886 *) Avoid coredump with unsupported or invalid public keys by checking if
5887 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
5888 PKCS7_verify() fails with non detached data.
5891 *) Don't use getenv in library functions when run as setuid/setgid.
5892 New function OPENSSL_issetugid().
5895 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
5896 due to incorrect handling of multi-threading:
5898 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
5900 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
5902 3. Count how many times MemCheck_off() has been called so that
5903 nested use can be treated correctly. This also avoids
5904 inband-signalling in the previous code (which relied on the
5905 assumption that thread ID 0 is impossible).
5908 *) Add "-rand" option also to s_client and s_server.
5911 *) Fix CPU detection on Irix 6.x.
5912 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
5913 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
5915 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
5918 [This change does not apply to 0.9.7.]
5920 *) Use the cached encoding of an X509_NAME structure rather than
5921 copying it. This is apparently the reason for the libsafe "errors"
5922 but the code is actually correct.
5925 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
5926 Bleichenbacher's DSA attack.
5927 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
5928 to be set and top=0 forces the highest bit to be set; top=-1 is new
5929 and leaves the highest bit random.
5930 [Ulf Moeller, Bodo Moeller]
5932 *) In the NCONF_...-based implementations for CONF_... queries
5933 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
5934 a temporary CONF structure with the data component set to NULL
5935 (which gives segmentation faults in lh_retrieve).
5936 Instead, use NULL for the CONF pointer in CONF_get_string and
5937 CONF_get_number (which may use environment variables) and directly
5938 return NULL from CONF_get_section.
5941 *) Fix potential buffer overrun for EBCDIC.
5944 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
5945 keyUsage if basicConstraints absent for a CA.
5948 *) Make SMIME_write_PKCS7() write mail header values with a format that
5949 is more generally accepted (no spaces before the semicolon), since
5950 some programs can't parse those values properly otherwise. Also make
5951 sure BIO's that break lines after each write do not create invalid
5955 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
5956 macros previously used would not encode an empty SEQUENCE OF
5957 and break the signature.
5959 [This change does not apply to 0.9.7.]
5961 *) Zero the premaster secret after deriving the master secret in
5965 *) Add some EVP_add_digest_alias registrations (as found in
5966 OpenSSL_add_all_digests()) to SSL_library_init()
5967 aka OpenSSL_add_ssl_algorithms(). This provides improved
5968 compatibility with peers using X.509 certificates
5969 with unconventional AlgorithmIdentifier OIDs.
5972 *) Fix for Irix with NO_ASM.
5973 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
5975 *) ./config script fixes.
5976 [Ulf Moeller, Richard Levitte]
5978 *) Fix 'openssl passwd -1'.
5981 *) Change PKCS12_key_gen_asc() so it can cope with non null
5982 terminated strings whose length is passed in the passlen
5983 parameter, for example from PEM callbacks. This was done
5984 by adding an extra length parameter to asc2uni().
5985 [Steve Henson, reported by <oddissey@samsung.co.kr>]
5987 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
5988 call failed, free the DSA structure.
5991 *) Fix to uni2asc() to cope with zero length Unicode strings.
5992 These are present in some PKCS#12 files.
5995 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
5996 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
5997 when writing a 32767 byte record.
5998 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6000 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6001 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6003 (RSA objects have a reference count access to which is protected
6004 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6005 so they are meant to be shared between threads.)
6006 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6007 "Reddie, Steven" <Steven.Reddie@ca.com>]
6009 *) Fix a deadlock in CRYPTO_mem_leaks().
6012 *) Use better test patterns in bntest.
6015 *) rand_win.c fix for Borland C.
6018 *) BN_rshift bugfix for n == 0.
6021 *) Add a 'bctest' script that checks for some known 'bc' bugs
6022 so that 'make test' does not abort just because 'bc' is broken.
6025 *) Store verify_result within SSL_SESSION also for client side to
6026 avoid potential security hole. (Re-used sessions on the client side
6027 always resulted in verify_result==X509_V_OK, not using the original
6028 result of the server certificate verification.)
6031 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6032 SSL3_RT_APPLICATION_DATA, return 0.
6033 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6037 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6038 releases, have been re-implemented by renaming the previous
6039 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6040 and ssl3_read_internal, respectively, and adding 'peek' parameters
6041 to them. The new ssl[23]_{read,peek} functions are calls to
6042 ssl[23]_read_internal with the 'peek' flag set appropriately.
6043 A 'peek' parameter has also been added to ssl3_read_bytes, which
6044 does the actual work for ssl3_read_internal.
6047 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6048 the method-specific "init()" handler. Also clean up ex_data after
6049 calling the method-specific "finish()" handler. Previously, this was
6050 happening the other way round.
6053 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6054 The previous value, 12, was not always sufficient for BN_mod_exp().
6057 *) Make sure that shared libraries get the internal name engine with
6058 the full version number and not just 0. This should mark the
6059 shared libraries as not backward compatible. Of course, this should
6060 be changed again when we can guarantee backward binary compatibility.
6063 *) Fix typo in get_cert_by_subject() in by_dir.c
6064 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6066 *) Rework the system to generate shared libraries:
6068 - Make note of the expected extension for the shared libraries and
6069 if there is a need for symbolic links from for example libcrypto.so.0
6070 to libcrypto.so.0.9.7. There is extended info in Configure for
6073 - Make as few rebuilds of the shared libraries as possible.
6075 - Still avoid linking the OpenSSL programs with the shared libraries.
6077 - When installing, install the shared libraries separately from the
6081 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6083 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6084 and not in SSL_clear because the latter is also used by the
6085 accept/connect functions; previously, the settings made by
6086 SSL_set_read_ahead would be lost during the handshake.
6087 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6089 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6090 Previously, it would create entries for disableed algorithms no
6094 *) Added several new manual pages for SSL_* function.
6097 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6099 *) In ssl23_get_client_hello, generate an error message when faced
6100 with an initial SSL 3.0/TLS record that is too small to contain the
6101 first two bytes of the ClientHello message, i.e. client_version.
6102 (Note that this is a pathologic case that probably has never happened
6103 in real life.) The previous approach was to use the version number
6104 from the record header as a substitute; but our protocol choice
6105 should not depend on that one because it is not authenticated
6106 by the Finished messages.
6109 *) More robust randomness gathering functions for Windows.
6110 [Jeffrey Altman <jaltman@columbia.edu>]
6112 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
6113 not set then we don't setup the error code for issuer check errors
6114 to avoid possibly overwriting other errors which the callback does
6115 handle. If an application does set the flag then we assume it knows
6116 what it is doing and can handle the new informational codes
6120 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
6121 a general "ANY" type, as such it should be able to decode anything
6122 including tagged types. However it didn't check the class so it would
6123 wrongly interpret tagged types in the same way as their universal
6124 counterpart and unknown types were just rejected. Changed so that the
6125 tagged and unknown types are handled in the same way as a SEQUENCE:
6126 that is the encoding is stored intact. There is also a new type
6127 "V_ASN1_OTHER" which is used when the class is not universal, in this
6128 case we have no idea what the actual type is so we just lump them all
6132 *) On VMS, stdout may very well lead to a file that is written to
6133 in a record-oriented fashion. That means that every write() will
6134 write a separate record, which will be read separately by the
6135 programs trying to read from it. This can be very confusing.
6137 The solution is to put a BIO filter in the way that will buffer
6138 text until a linefeed is reached, and then write everything a
6139 line at a time, so every record written will be an actual line,
6140 not chunks of lines and not (usually doesn't happen, but I've
6141 seen it once) several lines in one record. BIO_f_linebuffer() is
6144 Currently, it's a VMS-only method, because that's where it has
6145 been tested well enough.
6148 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
6149 it can return incorrect results.
6150 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
6151 but it was in 0.9.6-beta[12].)
6154 *) Disable the check for content being present when verifying detached
6155 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
6156 include zero length content when signing messages.
6159 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
6160 BIO_ctrl (for BIO pairs).
6163 *) Add DSO method for VMS.
6166 *) Bug fix: Montgomery multiplication could produce results with the
6170 *) Add RPM specification openssl.spec and modify it to build three
6171 packages. The default package contains applications, application
6172 documentation and run-time libraries. The devel package contains
6173 include files, static libraries and function documentation. The
6174 doc package contains the contents of the doc directory. The original
6175 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
6178 *) Add a large number of documentation files for many SSL routines.
6179 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6181 *) Add a configuration entry for Sony News 4.
6182 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
6184 *) Don't set the two most significant bits to one when generating a
6185 random number < q in the DSA library.
6188 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
6189 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
6190 the underlying transport is blocking) if a handshake took place.
6191 (The default behaviour is needed by applications such as s_client
6192 and s_server that use select() to determine when to use SSL_read;
6193 but for applications that know in advance when to expect data, it
6194 just makes things more complicated.)
6197 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
6201 *) Add a few more EBCDIC conditionals that make `req' and `x509'
6202 work better on such systems.
6203 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6205 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
6206 Update PKCS12_parse() so it copies the friendlyName and the
6207 keyid to the certificates aux info.
6210 *) Fix bug in PKCS7_verify() which caused an infinite loop
6211 if there was more than one signature.
6212 [Sven Uszpelkat <su@celocom.de>]
6214 *) Major change in util/mkdef.pl to include extra information
6215 about each symbol, as well as presentig variables as well
6216 as functions. This change means that there's n more need
6217 to rebuild the .num files when some algorithms are excluded.
6220 *) Allow the verify time to be set by an application,
6221 rather than always using the current time.
6224 *) Phase 2 verify code reorganisation. The certificate
6225 verify code now looks up an issuer certificate by a
6226 number of criteria: subject name, authority key id
6227 and key usage. It also verifies self signed certificates
6228 by the same criteria. The main comparison function is
6229 X509_check_issued() which performs these checks.
6231 Lot of changes were necessary in order to support this
6232 without completely rewriting the lookup code.
6234 Authority and subject key identifier are now cached.
6236 The LHASH 'certs' is X509_STORE has now been replaced
6237 by a STACK_OF(X509_OBJECT). This is mainly because an
6238 LHASH can't store or retrieve multiple objects with
6239 the same hash value.
6241 As a result various functions (which were all internal
6242 use only) have changed to handle the new X509_STORE
6243 structure. This will break anything that messed round
6244 with X509_STORE internally.
6246 The functions X509_STORE_add_cert() now checks for an
6247 exact match, rather than just subject name.
6249 The X509_STORE API doesn't directly support the retrieval
6250 of multiple certificates matching a given criteria, however
6251 this can be worked round by performing a lookup first
6252 (which will fill the cache with candidate certificates)
6253 and then examining the cache for matches. This is probably
6254 the best we can do without throwing out X509_LOOKUP
6255 entirely (maybe later...).
6257 The X509_VERIFY_CTX structure has been enhanced considerably.
6259 All certificate lookup operations now go via a get_issuer()
6260 callback. Although this currently uses an X509_STORE it
6261 can be replaced by custom lookups. This is a simple way
6262 to bypass the X509_STORE hackery necessary to make this
6263 work and makes it possible to use more efficient techniques
6264 in future. A very simple version which uses a simple
6265 STACK for its trusted certificate store is also provided
6266 using X509_STORE_CTX_trusted_stack().
6268 The verify_cb() and verify() callbacks now have equivalents
6269 in the X509_STORE_CTX structure.
6271 X509_STORE_CTX also has a 'flags' field which can be used
6272 to customise the verify behaviour.
6275 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
6276 excludes S/MIME capabilities.
6279 *) When a certificate request is read in keep a copy of the
6280 original encoding of the signed data and use it when outputing
6281 again. Signatures then use the original encoding rather than
6282 a decoded, encoded version which may cause problems if the
6283 request is improperly encoded.
6286 *) For consistency with other BIO_puts implementations, call
6287 buffer_write(b, ...) directly in buffer_puts instead of calling
6290 In BIO_puts, increment b->num_write as in BIO_write.
6291 [Peter.Sylvester@EdelWeb.fr]
6293 *) Fix BN_mul_word for the case where the word is 0. (We have to use
6294 BN_zero, we may not return a BIGNUM with an array consisting of
6298 *) Avoid calling abort() from within the library when problems are
6299 detected, except if preprocessor symbols have been defined
6300 (such as REF_CHECK, BN_DEBUG etc.).
6303 *) New openssl application 'rsautl'. This utility can be
6304 used for low level RSA operations. DER public key
6305 BIO/fp routines also added.
6308 *) New Configure entry and patches for compiling on QNX 4.
6309 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
6311 *) A demo state-machine implementation was sponsored by
6312 Nuron (http://www.nuron.com/) and is now available in
6313 demos/state_machine.
6316 *) New options added to the 'dgst' utility for signature
6317 generation and verification.
6320 *) Unrecognized PKCS#7 content types are now handled via a
6321 catch all ASN1_TYPE structure. This allows unsupported
6322 types to be stored as a "blob" and an application can
6323 encode and decode it manually.
6326 *) Fix various signed/unsigned issues to make a_strex.c
6328 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
6330 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
6331 length if passed a buffer. ASN1_INTEGER_to_BN failed
6332 if passed a NULL BN and its argument was negative.
6333 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
6335 *) Modification to PKCS#7 encoding routines to output definite
6336 length encoding. Since currently the whole structures are in
6337 memory there's not real point in using indefinite length
6338 constructed encoding. However if OpenSSL is compiled with
6339 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
6342 *) Added BIO_vprintf() and BIO_vsnprintf().
6345 *) Added more prefixes to parse for in the the strings written
6346 through a logging bio, to cover all the levels that are available
6347 through syslog. The prefixes are now:
6349 PANIC, EMERG, EMR => LOG_EMERG
6350 ALERT, ALR => LOG_ALERT
6351 CRIT, CRI => LOG_CRIT
6352 ERROR, ERR => LOG_ERR
6353 WARNING, WARN, WAR => LOG_WARNING
6354 NOTICE, NOTE, NOT => LOG_NOTICE
6355 INFO, INF => LOG_INFO
6356 DEBUG, DBG => LOG_DEBUG
6358 and as before, if none of those prefixes are present at the
6359 beginning of the string, LOG_ERR is chosen.
6361 On Win32, the LOG_* levels are mapped according to this:
6363 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
6364 LOG_WARNING => EVENTLOG_WARNING_TYPE
6365 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
6369 *) Made it possible to reconfigure with just the configuration
6370 argument "reconf" or "reconfigure". The command line arguments
6371 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
6372 and are retrieved from there when reconfiguring.
6376 [Assar Westerlund <assar@sics.se>, Richard Levitte]
6378 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
6381 *) The obj_dat.pl script was messing up the sorting of object
6382 names. The reason was that it compared the quoted version
6383 of strings as a result "OCSP" > "OCSP Signing" because
6384 " > SPACE. Changed script to store unquoted versions of
6385 names and add quotes on output. It was also omitting some
6386 names from the lookup table if they were given a default
6387 value (that is if SN is missing it is given the same
6388 value as LN and vice versa), these are now added on the
6389 grounds that if an object has a name we should be able to
6390 look it up. Finally added warning output when duplicate
6391 short or long names are found.
6394 *) Changes needed for Tandem NSK.
6395 [Scott Uroff <scott@xypro.com>]
6397 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
6398 RSA_padding_check_SSLv23(), special padding was never detected
6399 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
6400 version rollback attacks was not effective.
6402 In s23_clnt.c, don't use special rollback-attack detection padding
6403 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
6404 client; similarly, in s23_srvr.c, don't do the rollback check if
6405 SSL 2.0 is the only protocol enabled in the server.
6408 *) Make it possible to get hexdumps of unprintable data with 'openssl
6409 asn1parse'. By implication, the functions ASN1_parse_dump() and
6410 BIO_dump_indent() are added.
6413 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
6414 these print out strings and name structures based on various
6415 flags including RFC2253 support and proper handling of
6416 multibyte characters. Added options to the 'x509' utility
6417 to allow the various flags to be set.
6420 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
6421 Also change the functions X509_cmp_current_time() and
6422 X509_gmtime_adj() work with an ASN1_TIME structure,
6423 this will enable certificates using GeneralizedTime in validity
6424 dates to be checked.
6427 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
6428 negative public key encodings) on by default,
6429 NO_NEG_PUBKEY_BUG can be set to disable it.
6432 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
6433 content octets. An i2c_ASN1_OBJECT is unnecessary because
6434 the encoding can be trivially obtained from the structure.
6437 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
6438 not read locks (CRYPTO_r_[un]lock).
6441 *) A first attempt at creating official support for shared
6442 libraries through configuration. I've kept it so the
6443 default is static libraries only, and the OpenSSL programs
6444 are always statically linked for now, but there are
6445 preparations for dynamic linking in place.
6446 This has been tested on Linux and Tru64.
6449 *) Randomness polling function for Win9x, as described in:
6450 Peter Gutmann, Software Generation of Practically Strong
6454 *) Fix so PRNG is seeded in req if using an already existing
6458 *) New options to smime application. -inform and -outform
6459 allow alternative formats for the S/MIME message including
6460 PEM and DER. The -content option allows the content to be
6461 specified separately. This should allow things like Netscape
6462 form signing output easier to verify.
6465 *) Fix the ASN1 encoding of tags using the 'long form'.
6468 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
6469 STRING types. These convert content octets to and from the
6470 underlying type. The actual tag and length octets are
6471 already assumed to have been read in and checked. These
6472 are needed because all other string types have virtually
6473 identical handling apart from the tag. By having versions
6474 of the ASN1 functions that just operate on content octets
6475 IMPLICIT tagging can be handled properly. It also allows
6476 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
6477 and ASN1_INTEGER are identical apart from the tag.
6480 *) Change the handling of OID objects as follows:
6482 - New object identifiers are inserted in objects.txt, following
6483 the syntax given in objects.README.
6484 - objects.pl is used to process obj_mac.num and create a new
6486 - obj_dat.pl is used to create a new obj_dat.h, using the data in
6489 This is currently kind of a hack, and the perl code in objects.pl
6490 isn't very elegant, but it works as I intended. The simplest way
6491 to check that it worked correctly is to look in obj_dat.h and
6492 check the array nid_objs and make sure the objects haven't moved
6493 around (this is important!). Additions are OK, as well as
6494 consistent name changes.
6497 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
6500 *) Addition of the command line parameter '-rand file' to 'openssl req'.
6501 The given file adds to whatever has already been seeded into the
6502 random pool through the RANDFILE configuration file option or
6503 environment variable, or the default random state file.
6506 *) mkstack.pl now sorts each macro group into lexical order.
6507 Previously the output order depended on the order the files
6508 appeared in the directory, resulting in needless rewriting
6512 *) Patches to make OpenSSL compile under Win32 again. Mostly
6513 work arounds for the VC++ problem that it treats func() as
6514 func(void). Also stripped out the parts of mkdef.pl that
6515 added extra typesafe functions: these no longer exist.
6518 *) Reorganisation of the stack code. The macros are now all
6519 collected in safestack.h . Each macro is defined in terms of
6520 a "stack macro" of the form SKM_<name>(type, a, b). The
6521 DEBUG_SAFESTACK is now handled in terms of function casts,
6522 this has the advantage of retaining type safety without the
6523 use of additional functions. If DEBUG_SAFESTACK is not defined
6524 then the non typesafe macros are used instead. Also modified the
6525 mkstack.pl script to handle the new form. Needs testing to see
6526 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
6527 the default if no major problems. Similar behaviour for ASN1_SET_OF
6528 and PKCS12_STACK_OF.
6531 *) When some versions of IIS use the 'NET' form of private key the
6532 key derivation algorithm is different. Normally MD5(password) is
6533 used as a 128 bit RC4 key. In the modified case
6534 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
6535 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
6536 as the old Netscape_RSA functions except they have an additional
6537 'sgckey' parameter which uses the modified algorithm. Also added
6538 an -sgckey command line option to the rsa utility. Thanks to
6539 Adrian Peck <bertie@ncipher.com> for posting details of the modified
6540 algorithm to openssl-dev.
6543 *) The evp_local.h macros were using 'c.##kname' which resulted in
6544 invalid expansion on some systems (SCO 5.0.5 for example).
6545 Corrected to 'c.kname'.
6546 [Phillip Porch <root@theporch.com>]
6548 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
6549 a STACK of email addresses from a certificate or request, these look
6550 in the subject name and the subject alternative name extensions and
6551 omit any duplicate addresses.
6554 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
6555 This makes DSA verification about 2 % faster.
6558 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
6559 (meaning that now 2^5 values will be precomputed, which is only 4 KB
6560 plus overhead for 1024 bit moduli).
6561 This makes exponentiations about 0.5 % faster for 1024 bit
6562 exponents (as measured by "openssl speed rsa2048").
6565 *) Rename memory handling macros to avoid conflicts with other
6567 Malloc => OPENSSL_malloc
6568 Malloc_locked => OPENSSL_malloc_locked
6569 Realloc => OPENSSL_realloc
6570 Free => OPENSSL_free
6573 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
6574 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
6577 *) CygWin32 support.
6578 [John Jarvie <jjarvie@newsguy.com>]
6580 *) The type-safe stack code has been rejigged. It is now only compiled
6581 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
6582 by default all type-specific stack functions are "#define"d back to
6583 standard stack functions. This results in more streamlined output
6584 but retains the type-safety checking possibilities of the original
6588 *) The STACK code has been cleaned up, and certain type declarations
6589 that didn't make a lot of sense have been brought in line. This has
6590 also involved a cleanup of sorts in safestack.h to more correctly
6591 map type-safe stack functions onto their plain stack counterparts.
6592 This work has also resulted in a variety of "const"ifications of
6593 lots of the code, especially "_cmp" operations which should normally
6594 be prototyped with "const" parameters anyway.
6597 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
6598 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
6599 (The PRNG state consists of two parts, the large pool 'state' and 'md',
6600 where all of 'md' is used each time the PRNG is used, but 'state'
6601 is used only indexed by a cyclic counter. As entropy may not be
6602 well distributed from the beginning, 'md' is important as a
6603 chaining variable. However, the output function chains only half
6604 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
6605 all of 'md', and seeding with STATE_SIZE dummy bytes will result
6606 in all of 'state' being rewritten, with the new values depending
6607 on virtually all of 'md'. This overcomes the 80 bit limitation.)
6610 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
6611 the handshake is continued after ssl_verify_cert_chain();
6612 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
6613 can lead to 'unexplainable' connection aborts later.
6614 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
6616 *) Major EVP API cipher revision.
6617 Add hooks for extra EVP features. This allows various cipher
6618 parameters to be set in the EVP interface. Support added for variable
6619 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
6620 setting of RC2 and RC5 parameters.
6622 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
6625 Remove lots of duplicated code from the EVP library. For example *every*
6626 cipher init() function handles the 'iv' in the same way according to the
6627 cipher mode. They also all do nothing if the 'key' parameter is NULL and
6628 for CFB and OFB modes they zero ctx->num.
6630 New functionality allows removal of S/MIME code RC2 hack.
6632 Most of the routines have the same form and so can be declared in terms
6635 By shifting this to the top level EVP_CipherInit() it can be removed from
6636 all individual ciphers. If the cipher wants to handle IVs or keys
6637 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
6640 Change lots of functions like EVP_EncryptUpdate() to now return a
6641 value: although software versions of the algorithms cannot fail
6642 any installed hardware versions can.
6645 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
6646 this option is set, tolerate broken clients that send the negotiated
6647 protocol version number instead of the requested protocol version
6651 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
6652 i.e. non-zero for export ciphersuites, zero otherwise.
6653 Previous versions had this flag inverted, inconsistent with
6654 rsa_tmp_cb (..._TMP_RSA_CB).
6655 [Bodo Moeller; problem reported by Amit Chopra]
6657 *) Add missing DSA library text string. Work around for some IIS
6658 key files with invalid SEQUENCE encoding.
6661 *) Add a document (doc/standards.txt) that list all kinds of standards
6662 and so on that are implemented in OpenSSL.
6665 *) Enhance c_rehash script. Old version would mishandle certificates
6666 with the same subject name hash and wouldn't handle CRLs at all.
6667 Added -fingerprint option to crl utility, to support new c_rehash
6671 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
6674 *) Fix for SSL server purpose checking. Server checking was
6675 rejecting certificates which had extended key usage present
6676 but no ssl client purpose.
6677 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
6679 *) Make PKCS#12 code work with no password. The PKCS#12 spec
6680 is a little unclear about how a blank password is handled.
6681 Since the password in encoded as a BMPString with terminating
6682 double NULL a zero length password would end up as just the
6683 double NULL. However no password at all is different and is
6684 handled differently in the PKCS#12 key generation code. NS
6685 treats a blank password as zero length. MSIE treats it as no
6686 password on export: but it will try both on import. We now do
6687 the same: PKCS12_parse() tries zero length and no password if
6688 the password is set to "" or NULL (NULL is now a valid password:
6689 it wasn't before) as does the pkcs12 application.
6692 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
6693 perror when PEM_read_bio_X509_REQ fails, the error message must
6694 be obtained from the error queue.
6697 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
6698 it in ERR_remove_state if appropriate, and change ERR_get_state
6699 accordingly to avoid race conditions (this is necessary because
6700 thread_hash is no longer constant once set).
6703 *) Bugfix for linux-elf makefile.one.
6706 *) RSA_get_default_method() will now cause a default
6707 RSA_METHOD to be chosen if one doesn't exist already.
6708 Previously this was only set during a call to RSA_new()
6709 or RSA_new_method(NULL) meaning it was possible for
6710 RSA_get_default_method() to return NULL.
6713 *) Added native name translation to the existing DSO code
6714 that will convert (if the flag to do so is set) filenames
6715 that are sufficiently small and have no path information
6716 into a canonical native form. Eg. "blah" converted to
6717 "libblah.so" or "blah.dll" etc.
6720 *) New function ERR_error_string_n(e, buf, len) which is like
6721 ERR_error_string(e, buf), but writes at most 'len' bytes
6722 including the 0 terminator. For ERR_error_string_n, 'buf'
6724 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
6726 *) CONF library reworked to become more general. A new CONF
6727 configuration file reader "class" is implemented as well as a
6728 new functions (NCONF_*, for "New CONF") to handle it. The now
6729 old CONF_* functions are still there, but are reimplemented to
6730 work in terms of the new functions. Also, a set of functions
6731 to handle the internal storage of the configuration data is
6732 provided to make it easier to write new configuration file
6733 reader "classes" (I can definitely see something reading a
6734 configuration file in XML format, for example), called _CONF_*,
6735 or "the configuration storage API"...
6737 The new configuration file reading functions are:
6739 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
6740 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
6742 NCONF_default, NCONF_WIN32
6744 NCONF_dump_fp, NCONF_dump_bio
6746 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
6747 NCONF_new creates a new CONF object. This works in the same way
6748 as other interfaces in OpenSSL, like the BIO interface.
6749 NCONF_dump_* dump the internal storage of the configuration file,
6750 which is useful for debugging. All other functions take the same
6751 arguments as the old CONF_* functions wth the exception of the
6752 first that must be a `CONF *' instead of a `LHASH *'.
6754 To make it easer to use the new classes with the old CONF_* functions,
6755 the function CONF_set_default_method is provided.
6758 *) Add '-tls1' option to 'openssl ciphers', which was already
6759 mentioned in the documentation but had not been implemented.
6760 (This option is not yet really useful because even the additional
6761 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
6764 *) Initial DSO code added into libcrypto for letting OpenSSL (and
6765 OpenSSL-based applications) load shared libraries and bind to
6766 them in a portable way.
6767 [Geoff Thorpe, with contributions from Richard Levitte]
6769 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
6771 *) Make sure _lrotl and _lrotr are only used with MSVC.
6773 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
6774 (the default implementation of RAND_status).
6776 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
6777 to '-clrext' (= clear extensions), as intended and documented.
6778 [Bodo Moeller; inconsistency pointed out by Michael Attili
6781 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
6782 was larger than the MD block size.
6783 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
6785 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
6786 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
6787 using the passed key: if the passed key was a private key the result
6788 of X509_print(), for example, would be to print out all the private key
6792 *) des_quad_cksum() byte order bug fix.
6793 [Ulf Möller, using the problem description in krb4-0.9.7, where
6794 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
6796 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
6798 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
6800 *) For easily testing in shell scripts whether some command
6801 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
6802 returns with exit code 0 iff no command of the given name is available.
6803 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
6804 the output goes to stdout and nothing is printed to stderr.
6805 Additional arguments are always ignored.
6807 Since for each cipher there is a command of the same name,
6808 the 'no-cipher' compilation switches can be tested this way.
6810 ('openssl no-XXX' is not able to detect pseudo-commands such
6811 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
6814 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
6817 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
6818 is set; it will be thrown away anyway because each handshake creates
6820 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
6821 to parameters -- in previous versions (since OpenSSL 0.9.3) the
6822 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
6823 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
6826 *) New s_client option -ign_eof: EOF at stdin is ignored, and
6827 'Q' and 'R' lose their special meanings (quit/renegotiate).
6828 This is part of what -quiet does; unlike -quiet, -ign_eof
6829 does not suppress any output.
6832 *) Add compatibility options to the purpose and trust code. The
6833 purpose X509_PURPOSE_ANY is "any purpose" which automatically
6834 accepts a certificate or CA, this was the previous behaviour,
6835 with all the associated security issues.
6837 X509_TRUST_COMPAT is the old trust behaviour: only and
6838 automatically trust self signed roots in certificate store. A
6839 new trust setting X509_TRUST_DEFAULT is used to specify that
6840 a purpose has no associated trust setting and it should instead
6841 use the value in the default purpose.
6844 *) Fix the PKCS#8 DSA private key code so it decodes keys again
6845 and fix a memory leak.
6848 *) In util/mkerr.pl (which implements 'make errors'), preserve
6849 reason strings from the previous version of the .c file, as
6850 the default to have only downcase letters (and digits) in
6851 automatically generated reasons codes is not always appropriate.
6854 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
6855 using strerror. Previously, ERR_reason_error_string() returned
6856 library names as reason strings for SYSerr; but SYSerr is a special
6857 case where small numbers are errno values, not library numbers.
6860 *) Add '-dsaparam' option to 'openssl dhparam' application. This
6861 converts DSA parameters into DH parameters. (When creating parameters,
6862 DSA_generate_parameters is used.)
6865 *) Include 'length' (recommended exponent length) in C code generated
6866 by 'openssl dhparam -C'.
6869 *) The second argument to set_label in perlasm was already being used
6870 so couldn't be used as a "file scope" flag. Moved to third argument
6874 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
6875 instead of RAND_bytes for encryption IVs and salts.
6878 *) Include RAND_status() into RAND_METHOD instead of implementing
6879 it only for md_rand.c Otherwise replacing the PRNG by calling
6880 RAND_set_rand_method would be impossible.
6883 *) Don't let DSA_generate_key() enter an infinite loop if the random
6884 number generation fails.
6887 *) New 'rand' application for creating pseudo-random output.
6890 *) Added configuration support for Linux/IA64
6891 [Rolf Haberrecker <rolf@suse.de>]
6893 *) Assembler module support for Mingw32.
6896 *) Shared library support for HPUX (in shlib/).
6897 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
6899 *) Shared library support for Solaris gcc.
6900 [Lutz Behnke <behnke@trustcenter.de>]
6902 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
6904 *) PKCS7_encrypt() was adding text MIME headers twice because they
6905 were added manually and by SMIME_crlf_copy().
6908 *) In bntest.c don't call BN_rand with zero bits argument.
6909 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
6911 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
6912 case was implemented. This caused BN_div_recp() to fail occasionally.
6915 *) Add an optional second argument to the set_label() in the perl
6916 assembly language builder. If this argument exists and is set
6917 to 1 it signals that the assembler should use a symbol whose
6918 scope is the entire file, not just the current function. This
6919 is needed with MASM which uses the format label:: for this scope.
6920 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
6922 *) Change the ASN1 types so they are typedefs by default. Before
6923 almost all types were #define'd to ASN1_STRING which was causing
6924 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
6928 *) Change names of new functions to the new get1/get0 naming
6929 convention: After 'get1', the caller owns a reference count
6930 and has to call ..._free; 'get0' returns a pointer to some
6931 data structure without incrementing reference counters.
6932 (Some of the existing 'get' functions increment a reference
6933 counter, some don't.)
6934 Similarly, 'set1' and 'add1' functions increase reference
6935 counters or duplicate objects.
6938 *) Allow for the possibility of temp RSA key generation failure:
6939 the code used to assume it always worked and crashed on failure.
6942 *) Fix potential buffer overrun problem in BIO_printf().
6943 [Ulf Möller, using public domain code by Patrick Powell; problem
6944 pointed out by David Sacerdote <das33@cornell.edu>]
6946 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
6947 RAND_egd() and RAND_status(). In the command line application,
6948 the EGD socket can be specified like a seed file using RANDFILE
6952 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
6953 Some CAs (e.g. Verisign) distribute certificates in this form.
6956 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
6957 list to exclude them. This means that no special compilation option
6958 is needed to use anonymous DH: it just needs to be included in the
6962 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
6963 EVP_MD_type. The old functionality is available in a new macro called
6964 EVP_MD_md(). Change code that uses it and update docs.
6967 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
6968 where the 'void *' argument is replaced by a function pointer argument.
6969 Previously 'void *' was abused to point to functions, which works on
6970 many platforms, but is not correct. As these functions are usually
6971 called by macros defined in OpenSSL header files, most source code
6972 should work without changes.
6975 *) <openssl/opensslconf.h> (which is created by Configure) now contains
6976 sections with information on -D... compiler switches used for
6977 compiling the library so that applications can see them. To enable
6978 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
6979 must be defined. E.g.,
6980 #define OPENSSL_ALGORITHM_DEFINES
6981 #include <openssl/opensslconf.h>
6982 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
6983 [Richard Levitte, Ulf and Bodo Möller]
6985 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
6989 *) Change the 'other' type in certificate aux info to a STACK_OF
6990 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
6991 the required ASN1 format: arbitrary types determined by an OID.
6994 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
6995 argument to 'req'. This is not because the function is newer or
6996 better than others it just uses the work 'NEW' in the certificate
6997 request header lines. Some software needs this.
7000 *) Reorganise password command line arguments: now passwords can be
7001 obtained from various sources. Delete the PEM_cb function and make
7002 it the default behaviour: i.e. if the callback is NULL and the
7003 usrdata argument is not NULL interpret it as a null terminated pass
7004 phrase. If usrdata and the callback are NULL then the pass phrase
7005 is prompted for as usual.
7008 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7009 the support is automatically enabled. The resulting binaries will
7010 autodetect the card and use it if present.
7011 [Ben Laurie and Compaq Inc.]
7013 *) Work around for Netscape hang bug. This sends certificate request
7014 and server done in one record. Since this is perfectly legal in the
7015 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7016 the bugs/SSLv3 entry for more info.
7019 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7022 *) Add -rand argument to smime and pkcs12 applications and read/write
7026 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7029 *) Add command line password options to the remaining applications.
7032 *) Bug fix for BN_div_recp() for numerators with an even number of
7036 *) More tests in bntest.c, and changed test_bn output.
7039 *) ./config recognizes MacOS X now.
7042 *) Bug fix for BN_div() when the first words of num and divsor are
7043 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7046 *) Add support for various broken PKCS#8 formats, and command line
7047 options to produce them.
7050 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7051 get temporary BIGNUMs from a BN_CTX.
7054 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7058 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7059 include a #define from the old name to the new. The original intent
7060 was that statically linked binaries could for example just call
7061 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7062 link with digests. This never worked becayse SSLeay_add_all_digests()
7063 and SSLeay_add_all_ciphers() were in the same source file so calling
7064 one would link with the other. They are now in separate source files.
7067 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7070 *) Use a less unusual form of the Miller-Rabin primality test (it used
7071 a binary algorithm for exponentiation integrated into the Miller-Rabin
7072 loop, our standard modexp algorithms are faster).
7075 *) Support for the EBCDIC character set completed.
7076 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7078 *) Source code cleanups: use const where appropriate, eliminate casts,
7079 use void * instead of char * in lhash.
7082 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7083 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7084 this the server could overwrite ephemeral keys that the client
7088 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7089 using 50 iterations of the Rabin-Miller test.
7091 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7092 iterations of the Rabin-Miller test as required by the appendix
7093 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7094 As BN_is_prime_fasttest includes trial division, DSA parameter
7095 generation becomes much faster.
7097 This implies a change for the callback functions in DSA_is_prime
7098 and DSA_generate_parameters: The callback function is called once
7099 for each positive witness in the Rabin-Miller test, not just
7100 occasionally in the inner loop; and the parameters to the
7101 callback function now provide an iteration count for the outer
7102 loop rather than for the current invocation of the inner loop.
7103 DSA_generate_parameters additionally can call the callback
7104 function with an 'iteration count' of -1, meaning that a
7105 candidate has passed the trial division test (when q is generated
7106 from an application-provided seed, trial division is skipped).
7109 *) New function BN_is_prime_fasttest that optionally does trial
7110 division before starting the Rabin-Miller test and has
7111 an additional BN_CTX * argument (whereas BN_is_prime always
7112 has to allocate at least one BN_CTX).
7113 'callback(1, -1, cb_arg)' is called when a number has passed the
7114 trial division stage.
7117 *) Fix for bug in CRL encoding. The validity dates weren't being handled
7121 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
7124 *) New function BN_pseudo_rand().
7127 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
7128 bignum version of BN_from_montgomery() with the working code from
7129 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
7133 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
7134 made it impossible to use the same SSL_SESSION data structure in
7135 SSL2 clients in multiple threads.
7138 *) The return value of RAND_load_file() no longer counts bytes obtained
7139 by stat(). RAND_load_file(..., -1) is new and uses the complete file
7140 to seed the PRNG (previously an explicit byte count was required).
7141 [Ulf Möller, Bodo Möller]
7143 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
7144 used (char *) instead of (void *) and had casts all over the place.
7147 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
7150 *) Retain source code compatibility for BN_prime_checks macro:
7151 BN_is_prime(..., BN_prime_checks, ...) now uses
7152 BN_prime_checks_for_size to determine the appropriate number of
7153 Rabin-Miller iterations.
7156 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
7157 DH_CHECK_P_NOT_SAFE_PRIME.
7158 (Check if this is true? OpenPGP calls them "strong".)
7161 *) Merge the functionality of "dh" and "gendh" programs into a new program
7162 "dhparam". The old programs are retained for now but will handle DH keys
7163 (instead of parameters) in future.
7166 *) Make the ciphers, s_server and s_client programs check the return values
7167 when a new cipher list is set.
7170 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
7171 ciphers. Before when the 56bit ciphers were enabled the sorting was
7174 The syntax for the cipher sorting has been extended to support sorting by
7175 cipher-strength (using the strength_bits hard coded in the tables).
7176 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
7178 Fix a bug in the cipher-command parser: when supplying a cipher command
7179 string with an "undefined" symbol (neither command nor alphanumeric
7180 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
7181 an error is flagged.
7183 Due to the strength-sorting extension, the code of the
7184 ssl_create_cipher_list() function was completely rearranged. I hope that
7185 the readability was also increased :-)
7186 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7188 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
7189 for the first serial number and places 2 in the serial number file. This
7190 avoids problems when the root CA is created with serial number zero and
7191 the first user certificate has the same issuer name and serial number
7195 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
7196 the new code. Add documentation for this stuff.
7199 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
7200 X509_*() to X509at_*() on the grounds that they don't handle X509
7201 structures and behave in an analagous way to the X509v3 functions:
7202 they shouldn't be called directly but wrapper functions should be used
7205 So we also now have some wrapper functions that call the X509at functions
7206 when passed certificate requests. (TO DO: similar things can be done with
7207 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
7208 things. Some of these need some d2i or i2d and print functionality
7209 because they handle more complex structures.)
7212 *) Add missing #ifndefs that caused missing symbols when building libssl
7213 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
7214 NO_RSA in ssl/s2*.c.
7215 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
7217 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
7218 has a return value which indicates the quality of the random data
7219 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
7220 error queue. New function RAND_pseudo_bytes() generates output that is
7221 guaranteed to be unique but not unpredictable. RAND_add is like
7222 RAND_seed, but takes an extra argument for an entropy estimate
7223 (RAND_seed always assumes full entropy).
7226 *) Do more iterations of Rabin-Miller probable prime test (specifically,
7227 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
7228 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
7229 in crypto/bn/bn_prime.c for the complete table). This guarantees a
7230 false-positive rate of at most 2^-80 for random input.
7233 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
7236 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
7237 in the 0.9.5 release), this returns the chain
7238 from an X509_CTX structure with a dup of the stack and all
7239 the X509 reference counts upped: so the stack will exist
7240 after X509_CTX_cleanup() has been called. Modify pkcs12.c
7243 Also make SSL_SESSION_print() print out the verify return
7247 *) Add manpage for the pkcs12 command. Also change the default
7248 behaviour so MAC iteration counts are used unless the new
7249 -nomaciter option is used. This improves file security and
7250 only older versions of MSIE (4.0 for example) need it.
7253 *) Honor the no-xxx Configure options when creating .DEF files.
7256 *) Add PKCS#10 attributes to field table: challengePassword,
7257 unstructuredName and unstructuredAddress. These are taken from
7258 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
7259 international characters are used.
7261 More changes to X509_ATTRIBUTE code: allow the setting of types
7262 based on strings. Remove the 'loc' parameter when adding
7263 attributes because these will be a SET OF encoding which is sorted
7267 *) Initial changes to the 'req' utility to allow request generation
7268 automation. This will allow an application to just generate a template
7269 file containing all the field values and have req construct the
7272 Initial support for X509_ATTRIBUTE handling. Stacks of these are
7273 used all over the place including certificate requests and PKCS#7
7274 structures. They are currently handled manually where necessary with
7275 some primitive wrappers for PKCS#7. The new functions behave in a
7276 manner analogous to the X509 extension functions: they allow
7277 attributes to be looked up by NID and added.
7279 Later something similar to the X509V3 code would be desirable to
7280 automatically handle the encoding, decoding and printing of the
7281 more complex types. The string types like challengePassword can
7282 be handled by the string table functions.
7284 Also modified the multi byte string table handling. Now there is
7285 a 'global mask' which masks out certain types. The table itself
7286 can use the flag STABLE_NO_MASK to ignore the mask setting: this
7287 is useful when for example there is only one permissible type
7288 (as in countryName) and using the mask might result in no valid
7292 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
7293 SSL_get_peer_finished to allow applications to obtain the latest
7294 Finished messages sent to the peer or expected from the peer,
7295 respectively. (SSL_get_peer_finished is usually the Finished message
7296 actually received from the peer, otherwise the protocol will be aborted.)
7298 As the Finished message are message digests of the complete handshake
7299 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
7300 be used for external authentication procedures when the authentication
7301 provided by SSL/TLS is not desired or is not enough.
7304 *) Enhanced support for Alpha Linux is added. Now ./config checks if
7305 the host supports BWX extension and if Compaq C is present on the
7306 $PATH. Just exploiting of the BWX extension results in 20-30%
7307 performance kick for some algorithms, e.g. DES and RC4 to mention
7308 a couple. Compaq C in turn generates ~20% faster code for MD5 and
7312 *) Add support for MS "fast SGC". This is arguably a violation of the
7313 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
7314 weak crypto and after checking the certificate is SGC a second one
7315 with strong crypto. MS SGC stops the first handshake after receiving
7316 the server certificate message and sends a second client hello. Since
7317 a server will typically do all the time consuming operations before
7318 expecting any further messages from the client (server key exchange
7319 is the most expensive) there is little difference between the two.
7321 To get OpenSSL to support MS SGC we have to permit a second client
7322 hello message after we have sent server done. In addition we have to
7323 reset the MAC if we do get this second client hello.
7326 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
7327 if a DER encoded private key is RSA or DSA traditional format. Changed
7328 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
7329 format DER encoded private key. Newer code should use PKCS#8 format which
7330 has the key type encoded in the ASN1 structure. Added DER private key
7331 support to pkcs8 application.
7334 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
7335 ciphersuites has been selected (as required by the SSL 3/TLS 1
7336 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
7337 is set, we interpret this as a request to violate the specification
7338 (the worst that can happen is a handshake failure, and 'correct'
7339 behaviour would result in a handshake failure anyway).
7342 *) In SSL_CTX_add_session, take into account that there might be multiple
7343 SSL_SESSION structures with the same session ID (e.g. when two threads
7344 concurrently obtain them from an external cache).
7345 The internal cache can handle only one SSL_SESSION with a given ID,
7346 so if there's a conflict, we now throw out the old one to achieve
7350 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
7351 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
7352 some routines that use cipher OIDs: some ciphers do not have OIDs
7353 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
7357 *) Simplify the trust setting structure and code. Now we just have
7358 two sequences of OIDs for trusted and rejected settings. These will
7359 typically have values the same as the extended key usage extension
7360 and any application specific purposes.
7362 The trust checking code now has a default behaviour: it will just
7363 check for an object with the same NID as the passed id. Functions can
7364 be provided to override either the default behaviour or the behaviour
7365 for a given id. SSL client, server and email already have functions
7366 in place for compatibility: they check the NID and also return "trusted"
7367 if the certificate is self signed.
7370 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
7371 traditional format into an EVP_PKEY structure.
7374 *) Add a password callback function PEM_cb() which either prompts for
7375 a password if usr_data is NULL or otherwise assumes it is a null
7376 terminated password. Allow passwords to be passed on command line
7377 environment or config files in a few more utilities.
7380 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
7381 keys. Add some short names for PKCS#8 PBE algorithms and allow them
7382 to be specified on the command line for the pkcs8 and pkcs12 utilities.
7383 Update documentation.
7386 *) Support for ASN1 "NULL" type. This could be handled before by using
7387 ASN1_TYPE but there wasn't any function that would try to read a NULL
7388 and produce an error if it couldn't. For compatibility we also have
7389 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
7390 don't allocate anything because they don't need to.
7393 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
7395 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
7397 *) Rebuild of the memory allocation routines used by OpenSSL code and
7398 possibly others as well. The purpose is to make an interface that
7399 provide hooks so anyone can build a separate set of allocation and
7400 deallocation routines to be used by OpenSSL, for example memory
7401 pool implementations, or something else, which was previously hard
7402 since Malloc(), Realloc() and Free() were defined as macros having
7403 the values malloc, realloc and free, respectively (except for Win32
7404 compilations). The same is provided for memory debugging code.
7405 OpenSSL already comes with functionality to find memory leaks, but
7406 this gives people a chance to debug other memory problems.
7408 With these changes, a new set of functions and macros have appeared:
7410 CRYPTO_set_mem_debug_functions() [F]
7411 CRYPTO_get_mem_debug_functions() [F]
7412 CRYPTO_dbg_set_options() [F]
7413 CRYPTO_dbg_get_options() [F]
7414 CRYPTO_malloc_debug_init() [M]
7416 The memory debug functions are NULL by default, unless the library
7417 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
7418 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
7419 gives the standard debugging functions that come with OpenSSL) or
7420 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
7421 provided by the library user) must be used. When the standard
7422 debugging functions are used, CRYPTO_dbg_set_options can be used to
7423 request additional information:
7424 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
7425 the CRYPTO_MDEBUG_xxx macro when compiling the library.
7427 Also, things like CRYPTO_set_mem_functions will always give the
7428 expected result (the new set of functions is used for allocation
7429 and deallocation) at all times, regardless of platform and compiler
7432 To finish it up, some functions that were never use in any other
7433 way than through macros have a new API and new semantic:
7436 CRYPTO_dbg_realloc()
7439 All macros of value have retained their old syntax.
7440 [Richard Levitte and Bodo Moeller]
7442 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
7443 ordering of SMIMECapabilities wasn't in "strength order" and there
7444 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
7448 *) Some ASN1 types with illegal zero length encoding (INTEGER,
7449 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
7450 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
7452 *) Merge in my S/MIME library for OpenSSL. This provides a simple
7453 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
7454 functionality to handle multipart/signed properly) and a utility
7455 called 'smime' to call all this stuff. This is based on code I
7456 originally wrote for Celo who have kindly allowed it to be
7457 included in OpenSSL.
7460 *) Add variants des_set_key_checked and des_set_key_unchecked of
7461 des_set_key (aka des_key_sched). Global variable des_check_key
7462 decides which of these is called by des_set_key; this way
7463 des_check_key behaves as it always did, but applications and
7464 the library itself, which was buggy for des_check_key == 1,
7465 have a cleaner way to pick the version they need.
7468 *) New function PKCS12_newpass() which changes the password of a
7472 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
7473 dynamic mix. In both cases the ids can be used as an index into the
7474 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
7475 functions so they accept a list of the field values and the
7476 application doesn't need to directly manipulate the X509_TRUST
7480 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
7484 *) Modify the way the V3 extension code looks up extensions. This now
7485 works in a similar way to the object code: we have some "standard"
7486 extensions in a static table which is searched with OBJ_bsearch()
7487 and the application can add dynamic ones if needed. The file
7488 crypto/x509v3/ext_dat.h now has the info: this file needs to be
7489 updated whenever a new extension is added to the core code and kept
7490 in ext_nid order. There is a simple program 'tabtest.c' which checks
7491 this. New extensions are not added too often so this file can readily
7492 be maintained manually.
7494 There are two big advantages in doing things this way. The extensions
7495 can be looked up immediately and no longer need to be "added" using
7496 X509V3_add_standard_extensions(): this function now does nothing.
7497 [Side note: I get *lots* of email saying the extension code doesn't
7498 work because people forget to call this function]
7499 Also no dynamic allocation is done unless new extensions are added:
7500 so if we don't add custom extensions there is no need to call
7501 X509V3_EXT_cleanup().
7504 *) Modify enc utility's salting as follows: make salting the default. Add a
7505 magic header, so unsalted files fail gracefully instead of just decrypting
7506 to garbage. This is because not salting is a big security hole, so people
7507 should be discouraged from doing it.
7510 *) Fixes and enhancements to the 'x509' utility. It allowed a message
7511 digest to be passed on the command line but it only used this
7512 parameter when signing a certificate. Modified so all relevant
7513 operations are affected by the digest parameter including the
7514 -fingerprint and -x509toreq options. Also -x509toreq choked if a
7515 DSA key was used because it didn't fix the digest.
7518 *) Initial certificate chain verify code. Currently tests the untrusted
7519 certificates for consistency with the verify purpose (which is set
7520 when the X509_STORE_CTX structure is set up) and checks the pathlength.
7522 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
7523 this is because it will reject chains with invalid extensions whereas
7524 every previous version of OpenSSL and SSLeay made no checks at all.
7526 Trust code: checks the root CA for the relevant trust settings. Trust
7527 settings have an initial value consistent with the verify purpose: e.g.
7528 if the verify purpose is for SSL client use it expects the CA to be
7529 trusted for SSL client use. However the default value can be changed to
7530 permit custom trust settings: one example of this would be to only trust
7531 certificates from a specific "secure" set of CAs.
7533 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
7534 which should be used for version portability: especially since the
7535 verify structure is likely to change more often now.
7537 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
7538 to set them. If not set then assume SSL clients will verify SSL servers
7541 Two new options to the verify program: -untrusted allows a set of
7542 untrusted certificates to be passed in and -purpose which sets the
7543 intended purpose of the certificate. If a purpose is set then the
7544 new chain verify code is used to check extension consistency.
7547 *) Support for the authority information access extension.
7550 *) Modify RSA and DSA PEM read routines to transparently handle
7551 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
7552 public keys in a format compatible with certificate
7553 SubjectPublicKeyInfo structures. Unfortunately there were already
7554 functions called *_PublicKey_* which used various odd formats so
7555 these are retained for compatibility: however the DSA variants were
7556 never in a public release so they have been deleted. Changed dsa/rsa
7557 utilities to handle the new format: note no releases ever handled public
7558 keys so we should be OK.
7560 The primary motivation for this change is to avoid the same fiasco
7561 that dogs private keys: there are several incompatible private key
7562 formats some of which are standard and some OpenSSL specific and
7563 require various evil hacks to allow partial transparent handling and
7564 even then it doesn't work with DER formats. Given the option anything
7565 other than PKCS#8 should be dumped: but the other formats have to
7566 stay in the name of compatibility.
7568 With public keys and the benefit of hindsight one standard format
7569 is used which works with EVP_PKEY, RSA or DSA structures: though
7570 it clearly returns an error if you try to read the wrong kind of key.
7572 Added a -pubkey option to the 'x509' utility to output the public key.
7573 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
7574 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
7575 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
7576 that do the same as the EVP_PKEY_assign_*() except they up the
7577 reference count of the added key (they don't "swallow" the
7581 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
7582 CRLs would fail if the file contained no certificates or no CRLs:
7583 added a new function to read in both types and return the number
7584 read: this means that if none are read it will be an error. The
7585 DER versions of the certificate and CRL reader would always fail
7586 because it isn't possible to mix certificates and CRLs in DER format
7587 without choking one or the other routine. Changed this to just read
7588 a certificate: this is the best we can do. Also modified the code
7589 in apps/verify.c to take notice of return codes: it was previously
7590 attempting to read in certificates from NULL pointers and ignoring
7591 any errors: this is one reason why the cert and CRL reader seemed
7592 to work. It doesn't check return codes from the default certificate
7593 routines: these may well fail if the certificates aren't installed.
7596 *) Code to support otherName option in GeneralName.
7599 *) First update to verify code. Change the verify utility
7600 so it warns if it is passed a self signed certificate:
7601 for consistency with the normal behaviour. X509_verify
7602 has been modified to it will now verify a self signed
7603 certificate if *exactly* the same certificate appears
7604 in the store: it was previously impossible to trust a
7605 single self signed certificate. This means that:
7606 openssl verify ss.pem
7607 now gives a warning about a self signed certificate but
7608 openssl verify -CAfile ss.pem ss.pem
7612 *) For servers, store verify_result in SSL_SESSION data structure
7613 (and add it to external session representation).
7614 This is needed when client certificate verifications fails,
7615 but an application-provided verification callback (set by
7616 SSL_CTX_set_cert_verify_callback) allows accepting the session
7617 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
7618 but returns 1): When the session is reused, we have to set
7619 ssl->verify_result to the appropriate error code to avoid
7621 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
7623 *) Fix a bug in the new PKCS#7 code: it didn't consider the
7624 case in PKCS7_dataInit() where the signed PKCS7 structure
7625 didn't contain any existing data because it was being created.
7626 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
7628 *) Add a salt to the key derivation routines in enc.c. This
7629 forms the first 8 bytes of the encrypted file. Also add a
7630 -S option to allow a salt to be input on the command line.
7633 *) New function X509_cmp(). Oddly enough there wasn't a function
7634 to compare two certificates. We do this by working out the SHA1
7635 hash and comparing that. X509_cmp() will be needed by the trust
7639 *) SSL_get1_session() is like SSL_get_session(), but increments
7640 the reference count in the SSL_SESSION returned.
7641 [Geoff Thorpe <geoff@eu.c2.net>]
7643 *) Fix for 'req': it was adding a null to request attributes.
7644 Also change the X509_LOOKUP and X509_INFO code to handle
7645 certificate auxiliary information.
7648 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
7652 *) Add the possibility to add extra information to the memory leak
7653 detecting output, to form tracebacks, showing from where each
7654 allocation was originated: CRYPTO_push_info("constant string") adds
7655 the string plus current file name and line number to a per-thread
7656 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
7657 is like calling CYRPTO_pop_info() until the stack is empty.
7658 Also updated memory leak detection code to be multi-thread-safe.
7661 *) Add options -text and -noout to pkcs7 utility and delete the
7662 encryption options which never did anything. Update docs.
7665 *) Add options to some of the utilities to allow the pass phrase
7666 to be included on either the command line (not recommended on
7667 OSes like Unix) or read from the environment. Update the
7668 manpages and fix a few bugs.
7671 *) Add a few manpages for some of the openssl commands.
7674 *) Fix the -revoke option in ca. It was freeing up memory twice,
7675 leaking and not finding already revoked certificates.
7678 *) Extensive changes to support certificate auxiliary information.
7679 This involves the use of X509_CERT_AUX structure and X509_AUX
7680 functions. An X509_AUX function such as PEM_read_X509_AUX()
7681 can still read in a certificate file in the usual way but it
7682 will also read in any additional "auxiliary information". By
7683 doing things this way a fair degree of compatibility can be
7684 retained: existing certificates can have this information added
7685 using the new 'x509' options.
7687 Current auxiliary information includes an "alias" and some trust
7688 settings. The trust settings will ultimately be used in enhanced
7689 certificate chain verification routines: currently a certificate
7690 can only be trusted if it is self signed and then it is trusted
7694 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
7695 The problem was that one of the replacement routines had not been working
7696 since SSLeay releases. For now the offending routine has been replaced
7697 with non-optimised assembler. Even so, this now gives around 95%
7698 performance improvement for 1024 bit RSA signs.
7701 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
7702 handling. Most clients have the effective key size in bits equal to
7703 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
7704 A few however don't do this and instead use the size of the decrypted key
7705 to determine the RC2 key length and the AlgorithmIdentifier to determine
7706 the effective key length. In this case the effective key length can still
7707 be 40 bits but the key length can be 168 bits for example. This is fixed
7708 by manually forcing an RC2 key into the EVP_PKEY structure because the
7709 EVP code can't currently handle unusual RC2 key sizes: it always assumes
7710 the key length and effective key length are equal.
7713 *) Add a bunch of functions that should simplify the creation of
7714 X509_NAME structures. Now you should be able to do:
7715 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
7716 and have it automatically work out the correct field type and fill in
7717 the structures. The more adventurous can try:
7718 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
7719 and it will (hopefully) work out the correct multibyte encoding.
7722 *) Change the 'req' utility to use the new field handling and multibyte
7723 copy routines. Before the DN field creation was handled in an ad hoc
7724 way in req, ca, and x509 which was rather broken and didn't support
7725 BMPStrings or UTF8Strings. Since some software doesn't implement
7726 BMPStrings or UTF8Strings yet, they can be enabled using the config file
7727 using the dirstring_type option. See the new comment in the default
7728 openssl.cnf for more info.
7731 *) Make crypto/rand/md_rand.c more robust:
7732 - Assure unique random numbers after fork().
7733 - Make sure that concurrent threads access the global counter and
7734 md serializably so that we never lose entropy in them
7735 or use exactly the same state in multiple threads.
7736 Access to the large state is not always serializable because
7737 the additional locking could be a performance killer, and
7738 md should be large enough anyway.
7741 *) New file apps/app_rand.c with commonly needed functionality
7742 for handling the random seed file.
7744 Use the random seed file in some applications that previously did not:
7746 dsaparam -genkey (which also ignored its '-rand' option),
7749 x509 (when signing).
7750 Except on systems with /dev/urandom, it is crucial to have a random
7751 seed file at least for key creation, DSA signing, and for DH exchanges;
7752 for RSA signatures we could do without one.
7754 gendh and gendsa (unlike genrsa) used to read only the first byte
7755 of each file listed in the '-rand' option. The function as previously
7756 found in genrsa is now in app_rand.c and is used by all programs
7757 that support '-rand'.
7760 *) In RAND_write_file, use mode 0600 for creating files;
7761 don't just chmod when it may be too late.
7764 *) Report an error from X509_STORE_load_locations
7765 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
7768 *) New function ASN1_mbstring_copy() this copies a string in either
7769 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
7770 into an ASN1_STRING type. A mask of permissible types is passed
7771 and it chooses the "minimal" type to use or an error if not type
7775 *) Add function equivalents to the various macros in asn1.h. The old
7776 macros are retained with an M_ prefix. Code inside the library can
7777 use the M_ macros. External code (including the openssl utility)
7778 should *NOT* in order to be "shared library friendly".
7781 *) Add various functions that can check a certificate's extensions
7782 to see if it usable for various purposes such as SSL client,
7783 server or S/MIME and CAs of these types. This is currently
7784 VERY EXPERIMENTAL but will ultimately be used for certificate chain
7785 verification. Also added a -purpose flag to x509 utility to
7786 print out all the purposes.
7789 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
7793 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
7794 for, obtain and decode and extension and obtain its critical flag.
7795 This allows all the necessary extension code to be handled in a
7796 single function call.
7799 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
7800 platforms. See crypto/rc4/rc4_enc.c for further details.
7803 *) New -noout option to asn1parse. This causes no output to be produced
7804 its main use is when combined with -strparse and -out to extract data
7805 from a file (which may not be in ASN.1 format).
7808 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
7809 when producing the local key id.
7810 [Richard Levitte <levitte@stacken.kth.se>]
7812 *) New option -dhparam in s_server. This allows a DH parameter file to be
7813 stated explicitly. If it is not stated then it tries the first server
7814 certificate file. The previous behaviour hard coded the filename
7818 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
7819 a public key to be input or output. For example:
7820 openssl rsa -in key.pem -pubout -out pubkey.pem
7821 Also added necessary DSA public key functions to handle this.
7824 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
7825 in the message. This was handled by allowing
7826 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
7827 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
7829 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
7830 to the end of the strings whereas this didn't. This would cause problems
7831 if strings read with d2i_ASN1_bytes() were later modified.
7832 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
7834 *) Fix for base64 decode bug. When a base64 bio reads only one line of
7835 data and it contains EOF it will end up returning an error. This is
7836 caused by input 46 bytes long. The cause is due to the way base64
7837 BIOs find the start of base64 encoded data. They do this by trying a
7838 trial decode on each line until they find one that works. When they
7839 do a flag is set and it starts again knowing it can pass all the
7840 data directly through the decoder. Unfortunately it doesn't reset
7841 the context it uses. This means that if EOF is reached an attempt
7842 is made to pass two EOFs through the context and this causes the
7843 resulting error. This can also cause other problems as well. As is
7844 usual with these problems it takes *ages* to find and the fix is
7845 trivial: move one line.
7846 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
7848 *) Ugly workaround to get s_client and s_server working under Windows. The
7849 old code wouldn't work because it needed to select() on sockets and the
7850 tty (for keypresses and to see if data could be written). Win32 only
7851 supports select() on sockets so we select() with a 1s timeout on the
7852 sockets and then see if any characters are waiting to be read, if none
7853 are present then we retry, we also assume we can always write data to
7854 the tty. This isn't nice because the code then blocks until we've
7855 received a complete line of data and it is effectively polling the
7856 keyboard at 1s intervals: however it's quite a bit better than not
7857 working at all :-) A dedicated Windows application might handle this
7858 with an event loop for example.
7861 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
7862 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
7863 will be called when RSA_sign() and RSA_verify() are used. This is useful
7864 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
7865 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
7866 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
7867 This necessitated the support of an extra signature type NID_md5_sha1
7868 for SSL signatures and modifications to the SSL library to use it instead
7869 of calling RSA_public_decrypt() and RSA_private_encrypt().
7872 *) Add new -verify -CAfile and -CApath options to the crl program, these
7873 will lookup a CRL issuers certificate and verify the signature in a
7874 similar way to the verify program. Tidy up the crl program so it
7875 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
7876 less strict. It will now permit CRL extensions even if it is not
7877 a V2 CRL: this will allow it to tolerate some broken CRLs.
7880 *) Initialize all non-automatic variables each time one of the openssl
7881 sub-programs is started (this is necessary as they may be started
7882 multiple times from the "OpenSSL>" prompt).
7883 [Lennart Bang, Bodo Moeller]
7885 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
7886 removing all other RSA functionality (this is what NO_RSA does). This
7887 is so (for example) those in the US can disable those operations covered
7888 by the RSA patent while allowing storage and parsing of RSA keys and RSA
7892 *) Non-copying interface to BIO pairs.
7893 (still largely untested)
7896 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
7897 ASCII string. This was handled independently in various places before.
7900 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
7901 UTF8 strings a character at a time.
7904 *) Use client_version from client hello to select the protocol
7905 (s23_srvr.c) and for RSA client key exchange verification
7906 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
7909 *) Add various utility functions to handle SPKACs, these were previously
7910 handled by poking round in the structure internals. Added new function
7911 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
7912 print, verify and generate SPKACs. Based on an original idea from
7913 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
7916 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
7919 *) Allow the config file extension section to be overwritten on the
7920 command line. Based on an original idea from Massimiliano Pala
7921 <madwolf@comune.modena.it>. The new option is called -extensions
7922 and can be applied to ca, req and x509. Also -reqexts to override
7923 the request extensions in req and -crlexts to override the crl extensions
7927 *) Add new feature to the SPKAC handling in ca. Now you can include
7928 the same field multiple times by preceding it by "XXXX." for example:
7931 this is the same syntax as used in the req config file.
7934 *) Allow certificate extensions to be added to certificate requests. These
7935 are specified in a 'req_extensions' option of the req section of the
7936 config file. They can be printed out with the -text option to req but
7937 are otherwise ignored at present.
7940 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
7941 data read consists of only the final block it would not decrypted because
7942 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
7943 A misplaced 'break' also meant the decrypted final block might not be
7944 copied until the next read.
7947 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
7948 a few extra parameters to the DH structure: these will be useful if
7949 for example we want the value of 'q' or implement X9.42 DH.
7952 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
7953 provides hooks that allow the default DSA functions or functions on a
7954 "per key" basis to be replaced. This allows hardware acceleration and
7955 hardware key storage to be handled without major modification to the
7956 library. Also added low level modexp hooks and CRYPTO_EX structure and
7957 associated functions.
7960 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
7961 as "read only": it can't be written to and the buffer it points to will
7962 not be freed. Reading from a read only BIO is much more efficient than
7963 a normal memory BIO. This was added because there are several times when
7964 an area of memory needs to be read from a BIO. The previous method was
7965 to create a memory BIO and write the data to it, this results in two
7966 copies of the data and an O(n^2) reading algorithm. There is a new
7967 function BIO_new_mem_buf() which creates a read only memory BIO from
7968 an area of memory. Also modified the PKCS#7 routines to use read only
7972 *) Bugfix: ssl23_get_client_hello did not work properly when called in
7973 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
7974 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
7975 but a retry condition occured while trying to read the rest.
7978 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
7979 NID_pkcs7_encrypted by default: this was wrong since this should almost
7980 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
7981 the encrypted data type: this is a more sensible place to put it and it
7982 allows the PKCS#12 code to be tidied up that duplicated this
7986 *) Changed obj_dat.pl script so it takes its input and output files on
7987 the command line. This should avoid shell escape redirection problems
7991 *) Initial support for certificate extension requests, these are included
7992 in things like Xenroll certificate requests. Included functions to allow
7993 extensions to be obtained and added.
7996 *) -crlf option to s_client and s_server for sending newlines as
7997 CRLF (as required by many protocols).
8000 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8002 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8003 [Ralf S. Engelschall]
8005 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8006 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8008 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8012 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8013 DH parameters/keys (q is lost during that conversion, but the resulting
8014 DH parameters contain its length).
8016 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8017 much faster than DH_generate_parameters (which creates parameters
8018 where p = 2*q + 1), and also the smaller q makes DH computations
8019 much more efficient (160-bit exponentiation instead of 1024-bit
8020 exponentiation); so this provides a convenient way to support DHE
8021 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8022 utter importance to use
8023 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8025 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8026 when such DH parameters are used, because otherwise small subgroup
8027 attacks may become possible!
8030 *) Avoid memory leak in i2d_DHparams.
8033 *) Allow the -k option to be used more than once in the enc program:
8034 this allows the same encrypted message to be read by multiple recipients.
8037 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8038 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8039 it will always use the numerical form of the OID, even if it has a short
8043 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8044 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8045 otherwise bn_mod_exp was called. In the case of hardware keys for example
8046 no private key components need be present and it might store extra data
8047 in the RSA structure, which cannot be accessed from bn_mod_exp.
8048 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8049 private key operations.
8052 *) Added support for SPARC Linux.
8055 *) pem_password_cb function type incompatibly changed from
8056 typedef int pem_password_cb(char *buf, int size, int rwflag);
8058 ....(char *buf, int size, int rwflag, void *userdata);
8059 so that applications can pass data to their callbacks:
8060 The PEM[_ASN1]_{read,write}... functions and macros now take an
8061 additional void * argument, which is just handed through whenever
8062 the password callback is called.
8063 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8065 New function SSL_CTX_set_default_passwd_cb_userdata.
8067 Compatibility note: As many C implementations push function arguments
8068 onto the stack in reverse order, the new library version is likely to
8069 interoperate with programs that have been compiled with the old
8070 pem_password_cb definition (PEM_whatever takes some data that
8071 happens to be on the stack as its last argument, and the callback
8072 just ignores this garbage); but there is no guarantee whatsoever that
8075 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8076 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8077 problems not only on Windows, but also on some Unix platforms.
8078 To avoid problematic command lines, these definitions are now in an
8079 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8080 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8083 *) MIPS III/IV assembler module is reimplemented.
8086 *) More DES library cleanups: remove references to srand/rand and
8087 delete an unused file.
8090 *) Add support for the the free Netwide assembler (NASM) under Win32,
8091 since not many people have MASM (ml) and it can be hard to obtain.
8092 This is currently experimental but it seems to work OK and pass all
8093 the tests. Check out INSTALL.W32 for info.
8096 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8097 without temporary keys kept an extra copy of the server key,
8098 and connections with temporary keys did not free everything in case
8102 *) New function RSA_check_key and new openssl rsa option -check
8103 for verifying the consistency of RSA keys.
8104 [Ulf Moeller, Bodo Moeller]
8106 *) Various changes to make Win32 compile work:
8107 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
8108 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
8109 comparison" warnings.
8110 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
8113 *) Add a debugging option to PKCS#5 v2 key generation function: when
8114 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
8115 derived keys are printed to stderr.
8118 *) Copy the flags in ASN1_STRING_dup().
8119 [Roman E. Pavlov <pre@mo.msk.ru>]
8121 *) The x509 application mishandled signing requests containing DSA
8122 keys when the signing key was also DSA and the parameters didn't match.
8124 It was supposed to omit the parameters when they matched the signing key:
8125 the verifying software was then supposed to automatically use the CA's
8126 parameters if they were absent from the end user certificate.
8128 Omitting parameters is no longer recommended. The test was also
8129 the wrong way round! This was probably due to unusual behaviour in
8130 EVP_cmp_parameters() which returns 1 if the parameters match.
8131 This meant that parameters were omitted when they *didn't* match and
8132 the certificate was useless. Certificates signed with 'ca' didn't have
8134 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
8136 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
8137 The interface is as follows:
8138 Applications can use
8139 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
8140 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
8141 "off" is now the default.
8142 The library internally uses
8143 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
8144 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
8145 to disable memory-checking temporarily.
8147 Some inconsistent states that previously were possible (and were
8148 even the default) are now avoided.
8150 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
8151 with each memory chunk allocated; this is occasionally more helpful
8152 than just having a counter.
8154 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
8156 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
8160 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
8161 which largely parallels "options", but is for changing API behaviour,
8162 whereas "options" are about protocol behaviour.
8163 Initial "mode" flags are:
8165 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
8166 a single record has been written.
8167 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
8168 retries use the same buffer location.
8169 (But all of the contents must be
8173 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
8176 *) Fix problems with no-hmac etc.
8177 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
8179 *) New functions RSA_get_default_method(), RSA_set_method() and
8180 RSA_get_method(). These allows replacement of RSA_METHODs without having
8181 to mess around with the internals of an RSA structure.
8184 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
8185 Also really enable memory leak checks in openssl.c and in some
8187 [Chad C. Mulligan, Bodo Moeller]
8189 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
8190 up the length of negative integers. This has now been simplified to just
8191 store the length when it is first determined and use it later, rather
8192 than trying to keep track of where data is copied and updating it to
8194 [Steve Henson, reported by Brien Wheeler
8195 <bwheeler@authentica-security.com>]
8197 *) Add a new function PKCS7_signatureVerify. This allows the verification
8198 of a PKCS#7 signature but with the signing certificate passed to the
8199 function itself. This contrasts with PKCS7_dataVerify which assumes the
8200 certificate is present in the PKCS#7 structure. This isn't always the
8201 case: certificates can be omitted from a PKCS#7 structure and be
8202 distributed by "out of band" means (such as a certificate database).
8205 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
8206 function prototypes in pem.h, also change util/mkdef.pl to add the
8207 necessary function names.
8210 *) mk1mf.pl (used by Windows builds) did not properly read the
8211 options set by Configure in the top level Makefile, and Configure
8212 was not even able to write more than one option correctly.
8213 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
8216 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
8217 file to be loaded from a BIO or FILE pointer. The BIO version will
8218 for example allow memory BIOs to contain config info.
8221 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
8222 Whoever hopes to achieve shared-library compatibility across versions
8223 must use this, not the compile-time macro.
8224 (Exercise 0.9.4: Which is the minimum library version required by
8226 Note: All this applies only to multi-threaded programs, others don't
8230 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
8231 through a BIO pair triggered the default case, i.e.
8232 SSLerr(...,SSL_R_UNKNOWN_STATE).
8235 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
8236 can use the SSL library even if none of the specific BIOs is
8240 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
8241 for the encoded length.
8242 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
8244 *) Add initial documentation of the X509V3 functions.
8247 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
8248 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
8249 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
8250 secure PKCS#8 private key format with a high iteration count.
8253 *) Fix determination of Perl interpreter: A perl or perl5
8254 _directory_ in $PATH was also accepted as the interpreter.
8255 [Ralf S. Engelschall]
8257 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
8258 wrong with it but it was very old and did things like calling
8259 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
8263 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
8264 to use the new extension code.
8267 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
8268 with macros. This should make it easier to change their form, add extra
8269 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
8273 *) Add to configuration table a new entry that can specify an alternative
8274 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
8275 according to Mark Crispin <MRC@Panda.COM>.
8279 *) DES CBC did not update the IV. Weird.
8282 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
8283 Changing the behaviour of the former might break existing programs --
8284 where IV updating is needed, des_ncbc_encrypt can be used.
8287 *) When bntest is run from "make test" it drives bc to check its
8288 calculations, as well as internally checking them. If an internal check
8289 fails, it needs to cause bc to give a non-zero result or make test carries
8290 on without noticing the failure. Fixed.
8293 *) DES library cleanups.
8296 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
8297 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
8298 ciphers. NOTE: although the key derivation function has been verified
8299 against some published test vectors it has not been extensively tested
8300 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
8304 *) Instead of "mkdir -p", which is not fully portable, use new
8305 Perl script "util/mkdir-p.pl".
8308 *) Rewrite the way password based encryption (PBE) is handled. It used to
8309 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
8310 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
8311 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
8312 the 'parameter' field of the AlgorithmIdentifier is passed to the
8313 underlying key generation function so it must do its own ASN1 parsing.
8314 This has also changed the EVP_PBE_CipherInit() function which now has a
8315 'parameter' argument instead of literal salt and iteration count values
8316 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
8319 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
8320 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
8321 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
8322 KEY" because this clashed with PKCS#8 unencrypted string. Since this
8323 value was just used as a "magic string" and not used directly its
8324 value doesn't matter.
8327 *) Introduce some semblance of const correctness to BN. Shame C doesn't
8331 *) "linux-sparc64" configuration (ultrapenguin).
8332 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
8333 "linux-sparc" configuration.
8334 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
8336 *) config now generates no-xxx options for missing ciphers.
8339 *) Support the EBCDIC character set (work in progress).
8340 File ebcdic.c not yet included because it has a different license.
8341 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8343 *) Support BS2000/OSD-POSIX.
8344 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8346 *) Make callbacks for key generation use void * instead of char *.
8349 *) Make S/MIME samples compile (not yet tested).
8352 *) Additional typesafe stacks.
8355 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
8359 Changes between 0.9.3 and 0.9.3a [29 May 1999]
8361 *) New configuration variant "sco5-gcc".
8363 *) Updated some demos.
8364 [Sean O Riordain, Wade Scholine]
8366 *) Add missing BIO_free at exit of pkcs12 application.
8369 *) Fix memory leak in conf.c.
8372 *) Updates for Win32 to assembler version of MD5.
8375 *) Set #! path to perl in apps/der_chop to where we found it
8376 instead of using a fixed path.
8379 *) SHA library changes for irix64-mips4-cc.
8382 *) Improvements for VMS support.
8386 Changes between 0.9.2b and 0.9.3 [24 May 1999]
8388 *) Bignum library bug fix. IRIX 6 passes "make test" now!
8389 This also avoids the problems with SC4.2 and unpatched SC5.
8390 [Andy Polyakov <appro@fy.chalmers.se>]
8392 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
8393 These are required because of the typesafe stack would otherwise break
8394 existing code. If old code used a structure member which used to be STACK
8395 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
8396 sk_num or sk_value it would produce an error because the num, data members
8397 are not present in STACK_OF. Now it just produces a warning. sk_set
8398 replaces the old method of assigning a value to sk_value
8399 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
8400 that does this will no longer work (and should use sk_set instead) but
8401 this could be regarded as a "questionable" behaviour anyway.
8404 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
8405 correctly handle encrypted S/MIME data.
8408 *) Change type of various DES function arguments from des_cblock
8409 (which means, in function argument declarations, pointer to char)
8410 to des_cblock * (meaning pointer to array with 8 char elements),
8411 which allows the compiler to do more typechecking; it was like
8412 that back in SSLeay, but with lots of ugly casts.
8414 Introduce new type const_des_cblock.
8417 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
8418 problems: find RecipientInfo structure that matches recipient certificate
8419 and initialise the ASN1 structures properly based on passed cipher.
8422 *) Belatedly make the BN tests actually check the results.
8425 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
8426 to and from BNs: it was completely broken. New compilation option
8427 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
8428 key elements as negative integers.
8431 *) Reorganize and speed up MD5.
8432 [Andy Polyakov <appro@fy.chalmers.se>]
8435 [Richard Levitte <richard@levitte.org>]
8437 *) New option -out to asn1parse to allow the parsed structure to be
8438 output to a file. This is most useful when combined with the -strparse
8439 option to examine the output of things like OCTET STRINGS.
8442 *) Make SSL library a little more fool-proof by not requiring any longer
8443 that SSL_set_{accept,connect}_state be called before
8444 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
8445 in many applications because usually everything *appeared* to work as
8446 intended anyway -- now it really works as intended).
8449 *) Move openssl.cnf out of lib/.
8452 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
8453 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
8454 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
8455 [Ralf S. Engelschall]
8457 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
8458 handle PKCS#7 enveloped data properly.
8459 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
8461 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
8462 copying pointers. The cert_st handling is changed by this in
8463 various ways (and thus what used to be known as ctx->default_cert
8464 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
8465 any longer when s->cert does not give us what we need).
8466 ssl_cert_instantiate becomes obsolete by this change.
8467 As soon as we've got the new code right (possibly it already is?),
8468 we have solved a couple of bugs of the earlier code where s->cert
8469 was used as if it could not have been shared with other SSL structures.
8471 Note that using the SSL API in certain dirty ways now will result
8472 in different behaviour than observed with earlier library versions:
8473 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
8474 does not influence s as it used to.
8476 In order to clean up things more thoroughly, inside SSL_SESSION
8477 we don't use CERT any longer, but a new structure SESS_CERT
8478 that holds per-session data (if available); currently, this is
8479 the peer's certificate chain and, for clients, the server's certificate
8480 and temporary key. CERT holds only those values that can have
8481 meaningful defaults in an SSL_CTX.
8484 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
8485 from the internal representation. Various PKCS#7 fixes: remove some
8486 evil casts and set the enc_dig_alg field properly based on the signing
8490 *) Allow PKCS#12 password to be set from the command line or the
8491 environment. Let 'ca' get its config file name from the environment
8492 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
8496 *) Allow certificate policies extension to use an IA5STRING for the
8497 organization field. This is contrary to the PKIX definition but
8498 VeriSign uses it and IE5 only recognises this form. Document 'x509'
8502 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
8503 without disallowing inline assembler and the like for non-pedantic builds.
8506 *) Support Borland C++ builder.
8507 [Janez Jere <jj@void.si>, modified by Ulf Möller]
8512 *) SHA-1 cleanups and performance enhancements.
8513 [Andy Polyakov <appro@fy.chalmers.se>]
8515 *) Sparc v8plus assembler for the bignum library.
8516 [Andy Polyakov <appro@fy.chalmers.se>]
8518 *) Accept any -xxx and +xxx compiler options in Configure.
8521 *) Update HPUX configuration.
8524 *) Add missing sk_<type>_unshift() function to safestack.h
8525 [Ralf S. Engelschall]
8527 *) New function SSL_CTX_use_certificate_chain_file that sets the
8528 "extra_cert"s in addition to the certificate. (This makes sense
8529 only for "PEM" format files, as chains as a whole are not
8533 *) Support verify_depth from the SSL API.
8534 x509_vfy.c had what can be considered an off-by-one-error:
8535 Its depth (which was not part of the external interface)
8536 was actually counting the number of certificates in a chain;
8537 now it really counts the depth.
8540 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
8541 instead of X509err, which often resulted in confusing error
8542 messages since the error codes are not globally unique
8543 (e.g. an alleged error in ssl3_accept when a certificate
8544 didn't match the private key).
8546 *) New function SSL_CTX_set_session_id_context that allows to set a default
8547 value (so that you don't need SSL_set_session_id_context for each
8548 connection using the SSL_CTX).
8551 *) OAEP decoding bug fix.
8554 *) Support INSTALL_PREFIX for package builders, as proposed by
8558 *) New Configure options "threads" and "no-threads". For systems
8559 where the proper compiler options are known (currently Solaris
8560 and Linux), "threads" is the default.
8563 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
8566 *) Install various scripts to $(OPENSSLDIR)/misc, not to
8567 $(INSTALLTOP)/bin -- they shouldn't clutter directories
8568 such as /usr/local/bin.
8571 *) "make linux-shared" to build shared libraries.
8572 [Niels Poppe <niels@netbox.org>]
8574 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
8577 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
8578 extension adding in x509 utility.
8581 *) Remove NOPROTO sections and error code comments.
8584 *) Partial rewrite of the DEF file generator to now parse the ANSI
8588 *) New Configure options --prefix=DIR and --openssldir=DIR.
8591 *) Complete rewrite of the error code script(s). It is all now handled
8592 by one script at the top level which handles error code gathering,
8593 header rewriting and C source file generation. It should be much better
8594 than the old method: it now uses a modified version of Ulf's parser to
8595 read the ANSI prototypes in all header files (thus the old K&R definitions
8596 aren't needed for error creation any more) and do a better job of
8597 translating function codes into names. The old 'ASN1 error code imbedded
8598 in a comment' is no longer necessary and it doesn't use .err files which
8599 have now been deleted. Also the error code call doesn't have to appear all
8600 on one line (which resulted in some large lines...).
8603 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
8606 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
8607 0 (which usually indicates a closed connection), but continue reading.
8610 *) Fix some race conditions.
8613 *) Add support for CRL distribution points extension. Add Certificate
8614 Policies and CRL distribution points documentation.
8617 *) Move the autogenerated header file parts to crypto/opensslconf.h.
8620 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
8621 8 of keying material. Merlin has also confirmed interop with this fix
8622 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
8623 [Merlin Hughes <merlin@baltimore.ie>]
8625 *) Fix lots of warnings.
8626 [Richard Levitte <levitte@stacken.kth.se>]
8628 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
8629 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
8630 [Richard Levitte <levitte@stacken.kth.se>]
8632 *) Fix problems with sizeof(long) == 8.
8633 [Andy Polyakov <appro@fy.chalmers.se>]
8635 *) Change functions to ANSI C.
8638 *) Fix typos in error codes.
8639 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
8641 *) Remove defunct assembler files from Configure.
8644 *) SPARC v8 assembler BIGNUM implementation.
8645 [Andy Polyakov <appro@fy.chalmers.se>]
8647 *) Support for Certificate Policies extension: both print and set.
8648 Various additions to support the r2i method this uses.
8651 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
8652 return a const string when you are expecting an allocated buffer.
8655 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
8656 types DirectoryString and DisplayText.
8659 *) Add code to allow r2i extensions to access the configuration database,
8660 add an LHASH database driver and add several ctx helper functions.
8663 *) Fix an evil bug in bn_expand2() which caused various BN functions to
8664 fail when they extended the size of a BIGNUM.
8667 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
8668 support typesafe stack.
8671 *) Fix typo in SSL_[gs]et_options().
8672 [Nils Frostberg <nils@medcom.se>]
8674 *) Delete various functions and files that belonged to the (now obsolete)
8675 old X509V3 handling code.
8678 *) New Configure option "rsaref".
8681 *) Don't auto-generate pem.h.
8684 *) Introduce type-safe ASN.1 SETs.
8687 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
8688 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
8690 *) Introduce type-safe STACKs. This will almost certainly break lots of code
8691 that links with OpenSSL (well at least cause lots of warnings), but fear
8692 not: the conversion is trivial, and it eliminates loads of evil casts. A
8693 few STACKed things have been converted already. Feel free to convert more.
8694 In the fullness of time, I'll do away with the STACK type altogether.
8697 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
8698 specified in <certfile> by updating the entry in the index.txt file.
8699 This way one no longer has to edit the index.txt file manually for
8700 revoking a certificate. The -revoke option does the gory details now.
8701 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
8703 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
8704 `-text' option at all and this way the `-noout -text' combination was
8705 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
8706 [Ralf S. Engelschall]
8708 *) Make sure a corresponding plain text error message exists for the
8709 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
8710 verify callback function determined that a certificate was revoked.
8711 [Ralf S. Engelschall]
8713 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
8714 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
8715 all available cipers including rc5, which was forgotten until now.
8716 In order to let the testing shell script know which algorithms
8717 are available, a new (up to now undocumented) command
8718 "openssl list-cipher-commands" is used.
8721 *) Bugfix: s_client occasionally would sleep in select() when
8722 it should have checked SSL_pending() first.
8725 *) New functions DSA_do_sign and DSA_do_verify to provide access to
8726 the raw DSA values prior to ASN.1 encoding.
8729 *) Tweaks to Configure
8730 [Niels Poppe <niels@netbox.org>]
8732 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
8736 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
8739 *) New config option to avoid instructions that are illegal on the 80386.
8740 The default code is faster, but requires at least a 486.
8743 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
8744 SSL2_SERVER_VERSION (not used at all) macros, which are now the
8745 same as SSL2_VERSION anyway.
8748 *) New "-showcerts" option for s_client.
8751 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
8752 application. Various cleanups and fixes.
8755 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
8756 modify error routines to work internally. Add error codes and PBE init
8757 to library startup routines.
8760 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
8761 packing functions to asn1 and evp. Changed function names and error
8762 codes along the way.
8765 *) PKCS12 integration: and so it begins... First of several patches to
8766 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
8767 objects to objects.h
8770 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
8771 and display support for Thawte strong extranet extension.
8774 *) Add LinuxPPC support.
8775 [Jeff Dubrule <igor@pobox.org>]
8777 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
8778 bn_div_words in alpha.s.
8779 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
8781 *) Make sure the RSA OAEP test is skipped under -DRSAref because
8782 OAEP isn't supported when OpenSSL is built with RSAref.
8783 [Ulf Moeller <ulf@fitug.de>]
8785 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
8786 so they no longer are missing under -DNOPROTO.
8787 [Soren S. Jorvang <soren@t.dk>]
8790 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
8792 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
8793 doesn't work when the session is reused. Coming soon!
8796 *) Fix a security hole, that allows sessions to be reused in the wrong
8797 context thus bypassing client cert protection! All software that uses
8798 client certs and session caches in multiple contexts NEEDS PATCHING to
8799 allow session reuse! A fuller solution is in the works.
8800 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
8802 *) Some more source tree cleanups (removed obsolete files
8803 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
8804 permission on "config" script to be executable) and a fix for the INSTALL
8806 [Ulf Moeller <ulf@fitug.de>]
8808 *) Remove some legacy and erroneous uses of malloc, free instead of
8810 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
8812 *) Make rsa_oaep_test return non-zero on error.
8813 [Ulf Moeller <ulf@fitug.de>]
8815 *) Add support for native Solaris shared libraries. Configure
8816 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
8817 if someone would make that last step automatic.
8818 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
8820 *) ctx_size was not built with the right compiler during "make links". Fixed.
8823 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
8824 except NULL ciphers". This means the default cipher list will no longer
8825 enable NULL ciphers. They need to be specifically enabled e.g. with
8826 the string "DEFAULT:eNULL".
8829 *) Fix to RSA private encryption routines: if p < q then it would
8830 occasionally produce an invalid result. This will only happen with
8831 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
8834 *) Be less restrictive and allow also `perl util/perlpath.pl
8835 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
8836 because this way one can also use an interpreter named `perl5' (which is
8837 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
8838 installed as `perl').
8839 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8841 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
8842 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8844 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
8845 advapi32.lib to Win32 build and change the pem test comparision
8846 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
8847 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
8848 and crypto/des/ede_cbcm_enc.c.
8851 *) DES quad checksum was broken on big-endian architectures. Fixed.
8854 *) Comment out two functions in bio.h that aren't implemented. Fix up the
8855 Win32 test batch file so it (might) work again. The Win32 test batch file
8856 is horrible: I feel ill....
8859 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
8860 in e_os.h. Audit of header files to check ANSI and non ANSI
8861 sections: 10 functions were absent from non ANSI section and not exported
8862 from Windows DLLs. Fixed up libeay.num for new functions.
8865 *) Make `openssl version' output lines consistent.
8866 [Ralf S. Engelschall]
8868 *) Fix Win32 symbol export lists for BIO functions: Added
8869 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
8870 to ms/libeay{16,32}.def.
8871 [Ralf S. Engelschall]
8873 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
8874 fine under Unix and passes some trivial tests I've now added. But the
8875 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
8876 added to make sure no one expects that this stuff really works in the
8877 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
8878 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
8880 [Ralf S. Engelschall]
8882 *) Fix the generation of two part addresses in perl.
8883 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
8885 *) Add config entry for Linux on MIPS.
8886 [John Tobey <jtobey@channel1.com>]
8888 *) Make links whenever Configure is run, unless we are on Windoze.
8891 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
8892 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
8896 *) Add a useful kludge to allow package maintainers to specify compiler and
8897 other platforms details on the command line without having to patch the
8898 Configure script everytime: One now can use ``perl Configure
8899 <id>:<details>'', i.e. platform ids are allowed to have details appended
8900 to them (seperated by colons). This is treated as there would be a static
8901 pre-configured entry in Configure's %table under key <id> with value
8902 <details> and ``perl Configure <id>'' is called. So, when you want to
8903 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
8904 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
8905 now, which overrides the FreeBSD-elf entry on-the-fly.
8906 [Ralf S. Engelschall]
8908 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
8911 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
8912 on the `perl Configure ...' command line. This way one can compile
8913 OpenSSL libraries with Position Independent Code (PIC) which is needed
8914 for linking it into DSOs.
8915 [Ralf S. Engelschall]
8917 *) Remarkably, export ciphers were totally broken and no-one had noticed!
8921 *) Cleaned up the LICENSE document: The official contact for any license
8922 questions now is the OpenSSL core team under openssl-core@openssl.org.
8923 And add a paragraph about the dual-license situation to make sure people
8924 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
8925 to the OpenSSL toolkit.
8926 [Ralf S. Engelschall]
8928 *) General source tree makefile cleanups: Made `making xxx in yyy...'
8929 display consistent in the source tree and replaced `/bin/rm' by `rm'.
8930 Additonally cleaned up the `make links' target: Remove unnecessary
8931 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
8932 to speed processing and no longer clutter the display with confusing
8933 stuff. Instead only the actually done links are displayed.
8934 [Ralf S. Engelschall]
8936 *) Permit null encryption ciphersuites, used for authentication only. It used
8937 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
8938 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
8942 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
8943 signed attributes when verifying signatures (this would break them),
8944 the detached data encoding was wrong and public keys obtained using
8945 X509_get_pubkey() weren't freed.
8948 *) Add text documentation for the BUFFER functions. Also added a work around
8949 to a Win95 console bug. This was triggered by the password read stuff: the
8950 last character typed gets carried over to the next fread(). If you were
8951 generating a new cert request using 'req' for example then the last
8952 character of the passphrase would be CR which would then enter the first
8956 *) Added the new `Includes OpenSSL Cryptography Software' button as
8957 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
8958 button and can be used by applications based on OpenSSL to show the
8959 relationship to the OpenSSL project.
8960 [Ralf S. Engelschall]
8962 *) Remove confusing variables in function signatures in files
8963 ssl/ssl_lib.c and ssl/ssl.h.
8964 [Lennart Bong <lob@kulthea.stacken.kth.se>]
8966 *) Don't install bss_file.c under PREFIX/include/
8967 [Lennart Bong <lob@kulthea.stacken.kth.se>]
8969 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
8970 functions that return function pointers and has support for NT specific
8971 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
8972 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
8973 unsigned to signed types: this was killing the Win32 compile.
8976 *) Add new certificate file to stack functions,
8977 SSL_add_dir_cert_subjects_to_stack() and
8978 SSL_add_file_cert_subjects_to_stack(). These largely supplant
8979 SSL_load_client_CA_file(), and can be used to add multiple certs easily
8980 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
8981 This means that Apache-SSL and similar packages don't have to mess around
8982 to add as many CAs as they want to the preferred list.
8985 *) Experiment with doxygen documentation. Currently only partially applied to
8987 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
8988 openssl.doxy as the configuration file.
8991 *) Get rid of remaining C++-style comments which strict C compilers hate.
8992 [Ralf S. Engelschall, pointed out by Carlos Amengual]
8994 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
8995 compiled in by default: it has problems with large keys.
8998 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
8999 DH private keys and/or callback functions which directly correspond to
9000 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9001 is needed for applications which have to configure certificates on a
9002 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9004 For the RSA certificate situation is makes no difference, but
9005 for the DSA certificate situation this fixes the "no shared cipher"
9006 problem where the OpenSSL cipher selection procedure failed because the
9007 temporary keys were not overtaken from the context and the API provided
9008 no way to reconfigure them.
9009 The new functions now let applications reconfigure the stuff and they
9010 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9011 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9012 non-public-API function ssl_cert_instantiate() is used as a helper
9013 function and also to reduce code redundancy inside ssl_rsa.c.
9014 [Ralf S. Engelschall]
9016 *) Move s_server -dcert and -dkey options out of the undocumented feature
9017 area because they are useful for the DSA situation and should be
9018 recognized by the users.
9019 [Ralf S. Engelschall]
9021 *) Fix the cipher decision scheme for export ciphers: the export bits are
9022 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9023 SSL_EXP_MASK. So, the original variable has to be used instead of the
9024 already masked variable.
9025 [Richard Levitte <levitte@stacken.kth.se>]
9027 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9028 [Richard Levitte <levitte@stacken.kth.se>]
9030 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9031 from `int' to `unsigned int' because it's a length and initialized by
9032 EVP_DigestFinal() which expects an `unsigned int *'.
9033 [Richard Levitte <levitte@stacken.kth.se>]
9035 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9036 script. Instead use the usual Shell->Perl transition trick.
9037 [Ralf S. Engelschall]
9039 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9040 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9041 -noout -modulus' as it's already the case for `openssl rsa -noout
9042 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9043 currently the public key is printed (a decision which was already done by
9044 `openssl dsa -modulus' in the past) which serves a similar purpose.
9045 Additionally the NO_RSA no longer completely removes the whole -modulus
9046 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9048 [Ralf S. Engelschall]
9050 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9051 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9052 [Arne Ansper <arne@ats.cyber.ee>]
9054 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9055 to be added. Now both 'req' and 'ca' can use new objects defined in the
9059 *) Add cool BIO that does syslog (or event log on NT).
9060 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9062 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9063 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9064 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9065 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9068 *) Add preliminary config info for new extension code.
9071 *) Make RSA_NO_PADDING really use no padding.
9072 [Ulf Moeller <ulf@fitug.de>]
9074 *) Generate errors when private/public key check is done.
9077 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9078 for some CRL extensions and new objects added.
9081 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9082 key usage extension and fuller support for authority key id.
9085 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9086 padding method for RSA, which is recommended for new applications in PKCS
9087 #1 v2.0 (RFC 2437, October 1998).
9088 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9089 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9090 against Bleichbacher's attack on RSA.
9091 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9094 *) Updates to the new SSL compression code
9095 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9097 *) Fix so that the version number in the master secret, when passed
9098 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9099 (because the server will not accept higher), that the version number
9100 is 0x03,0x01, not 0x03,0x00
9101 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9103 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9104 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9105 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
9108 *) Support for RAW extensions where an arbitrary extension can be
9109 created by including its DER encoding. See apps/openssl.cnf for
9113 *) Make sure latest Perl versions don't interpret some generated C array
9114 code as Perl array code in the crypto/err/err_genc.pl script.
9115 [Lars Weber <3weber@informatik.uni-hamburg.de>]
9117 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
9118 not many people have the assembler. Various Win32 compilation fixes and
9119 update to the INSTALL.W32 file with (hopefully) more accurate Win32
9123 *) Modify configure script 'Configure' to automatically create crypto/date.h
9124 file under Win32 and also build pem.h from pem.org. New script
9125 util/mkfiles.pl to create the MINFO file on environments that can't do a
9126 'make files': perl util/mkfiles.pl >MINFO should work.
9129 *) Major rework of DES function declarations, in the pursuit of correctness
9130 and purity. As a result, many evil casts evaporated, and some weirdness,
9131 too. You may find this causes warnings in your code. Zapping your evil
9132 casts will probably fix them. Mostly.
9135 *) Fix for a typo in asn1.h. Bug fix to object creation script
9136 obj_dat.pl. It considered a zero in an object definition to mean
9137 "end of object": none of the objects in objects.h have any zeros
9138 so it wasn't spotted.
9139 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
9141 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
9142 Masking (CBCM). In the absence of test vectors, the best I have been able
9143 to do is check that the decrypt undoes the encrypt, so far. Send me test
9144 vectors if you have them.
9147 *) Correct calculation of key length for export ciphers (too much space was
9148 allocated for null ciphers). This has not been tested!
9151 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
9152 message is now correct (it understands "crypto" and "ssl" on its
9153 command line). There is also now an "update" option. This will update
9154 the util/ssleay.num and util/libeay.num files with any new functions.
9156 perl util/mkdef.pl crypto ssl update
9157 it will update them.
9160 *) Overhauled the Perl interface (perl/*):
9161 - ported BN stuff to OpenSSL's different BN library
9162 - made the perl/ source tree CVS-aware
9163 - renamed the package from SSLeay to OpenSSL (the files still contain
9164 their history because I've copied them in the repository)
9165 - removed obsolete files (the test scripts will be replaced
9166 by better Test::Harness variants in the future)
9167 [Ralf S. Engelschall]
9169 *) First cut for a very conservative source tree cleanup:
9170 1. merge various obsolete readme texts into doc/ssleay.txt
9171 where we collect the old documents and readme texts.
9172 2. remove the first part of files where I'm already sure that we no
9173 longer need them because of three reasons: either they are just temporary
9174 files which were left by Eric or they are preserved original files where
9175 I've verified that the diff is also available in the CVS via "cvs diff
9176 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
9177 the crypto/md/ stuff).
9178 [Ralf S. Engelschall]
9180 *) More extension code. Incomplete support for subject and issuer alt
9181 name, issuer and authority key id. Change the i2v function parameters
9182 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
9183 what that's for :-) Fix to ASN1 macro which messed up
9184 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
9187 *) Preliminary support for ENUMERATED type. This is largely copied from the
9191 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
9192 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9194 *) Make sure `make rehash' target really finds the `openssl' program.
9195 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9197 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
9198 like to hear about it if this slows down other processors.
9201 *) Add CygWin32 platform information to Configure script.
9202 [Alan Batie <batie@aahz.jf.intel.com>]
9204 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
9205 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
9207 *) New program nseq to manipulate netscape certificate sequences
9210 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
9214 *) Fixes to BN code. Previously the default was to define BN_RECURSION
9215 but the BN code had some problems that would cause failures when
9216 doing certificate verification and some other functions.
9217 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9219 *) Add ASN1 and PEM code to support netscape certificate sequences.
9222 *) Add ASN1 and PEM code to support netscape certificate sequences.
9225 *) Add several PKIX and private extended key usage OIDs.
9228 *) Modify the 'ca' program to handle the new extension code. Modify
9229 openssl.cnf for new extension format, add comments.
9232 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
9233 and add a sample to openssl.cnf so req -x509 now adds appropriate
9237 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
9238 error code, add initial support to X509_print() and x509 application.
9241 *) Takes a deep breath and start addding X509 V3 extension support code. Add
9242 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
9243 stuff is currently isolated and isn't even compiled yet.
9246 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
9247 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
9248 Removed the versions check from X509 routines when loading extensions:
9249 this allows certain broken certificates that don't set the version
9250 properly to be processed.
9253 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
9254 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
9255 can still be regenerated with "make depend".
9258 *) Spelling mistake in C version of CAST-128.
9259 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
9261 *) Changes to the error generation code. The perl script err-code.pl
9262 now reads in the old error codes and retains the old numbers, only
9263 adding new ones if necessary. It also only changes the .err files if new
9264 codes are added. The makefiles have been modified to only insert errors
9265 when needed (to avoid needlessly modifying header files). This is done
9266 by only inserting errors if the .err file is newer than the auto generated
9267 C file. To rebuild all the error codes from scratch (the old behaviour)
9268 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
9269 or delete all the .err files.
9272 *) CAST-128 was incorrectly implemented for short keys. The C version has
9273 been fixed, but is untested. The assembler versions are also fixed, but
9274 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
9275 to regenerate it if needed.
9276 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
9277 Hagino <itojun@kame.net>]
9279 *) File was opened incorrectly in randfile.c.
9280 [Ulf Möller <ulf@fitug.de>]
9282 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
9283 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
9284 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
9285 al: it's just almost always a UTCTime. Note this patch adds new error
9286 codes so do a "make errors" if there are problems.
9289 *) Correct Linux 1 recognition in config.
9290 [Ulf Möller <ulf@fitug.de>]
9292 *) Remove pointless MD5 hash when using DSA keys in ca.
9293 [Anonymous <nobody@replay.com>]
9295 *) Generate an error if given an empty string as a cert directory. Also
9296 generate an error if handed NULL (previously returned 0 to indicate an
9297 error, but didn't set one).
9298 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
9300 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
9303 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
9304 parameters. This was causing a warning which killed off the Win32 compile.
9307 *) Remove C++ style comments from crypto/bn/bn_local.h.
9308 [Neil Costigan <neil.costigan@celocom.com>]
9310 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
9311 based on a text string, looking up short and long names and finally
9312 "dot" format. The "dot" format stuff didn't work. Added new function
9313 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
9314 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
9315 OID is not part of the table.
9318 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
9319 X509_LOOKUP_by_alias().
9322 *) Sort openssl functions by name.
9325 *) Get the gendsa program working (hopefully) and add it to app list. Remove
9326 encryption from sample DSA keys (in case anyone is interested the password
9330 *) Make _all_ *_free functions accept a NULL pointer.
9331 [Frans Heymans <fheymans@isaserver.be>]
9333 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
9335 [Anonymous <nobody@replay.com>]
9337 *) s_server should send the CAfile as acceptable CAs, not its own cert.
9338 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9340 *) Don't blow it for numeric -newkey arguments to apps/req.
9341 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9343 *) Temp key "for export" tests were wrong in s3_srvr.c.
9344 [Anonymous <nobody@replay.com>]
9346 *) Add prototype for temp key callback functions
9347 SSL_CTX_set_tmp_{rsa,dh}_callback().
9350 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
9351 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
9354 *) X509_name_add_entry() freed the wrong thing after an error.
9355 [Arne Ansper <arne@ats.cyber.ee>]
9357 *) rsa_eay.c would attempt to free a NULL context.
9358 [Arne Ansper <arne@ats.cyber.ee>]
9360 *) BIO_s_socket() had a broken should_retry() on Windoze.
9361 [Arne Ansper <arne@ats.cyber.ee>]
9363 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
9364 [Arne Ansper <arne@ats.cyber.ee>]
9366 *) Make sure the already existing X509_STORE->depth variable is initialized
9367 in X509_STORE_new(), but document the fact that this variable is still
9368 unused in the certificate verification process.
9369 [Ralf S. Engelschall]
9371 *) Fix the various library and apps files to free up pkeys obtained from
9372 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
9375 *) Fix reference counting in X509_PUBKEY_get(). This makes
9376 demos/maurice/example2.c work, amongst others, probably.
9377 [Steve Henson and Ben Laurie]
9379 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
9380 `openssl' and second, the shortcut symlinks for the `openssl <command>'
9381 are no longer created. This way we have a single and consistent command
9382 line interface `openssl <command>', similar to `cvs <command>'.
9383 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
9385 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
9386 BIT STRING wrapper always have zero unused bits.
9389 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
9392 *) Make the top-level INSTALL documentation easier to understand.
9395 *) Makefiles updated to exit if an error occurs in a sub-directory
9396 make (including if user presses ^C) [Paul Sutton]
9398 *) Make Montgomery context stuff explicit in RSA data structure.
9401 *) Fix build order of pem and err to allow for generated pem.h.
9404 *) Fix renumbering bug in X509_NAME_delete_entry().
9407 *) Enhanced the err-ins.pl script so it makes the error library number
9408 global and can add a library name. This is needed for external ASN1 and
9409 other error libraries.
9412 *) Fixed sk_insert which never worked properly.
9415 *) Fix ASN1 macros so they can handle indefinite length construted
9416 EXPLICIT tags. Some non standard certificates use these: they can now
9420 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
9421 into a single doc/ssleay.txt bundle. This way the information is still
9422 preserved but no longer messes up this directory. Now it's new room for
9423 the new set of documenation files.
9424 [Ralf S. Engelschall]
9426 *) SETs were incorrectly DER encoded. This was a major pain, because they
9427 shared code with SEQUENCEs, which aren't coded the same. This means that
9428 almost everything to do with SETs or SEQUENCEs has either changed name or
9429 number of arguments.
9430 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
9432 *) Fix test data to work with the above.
9435 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
9436 was already fixed by Eric for 0.9.1 it seems.
9437 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
9439 *) Autodetect FreeBSD3.
9442 *) Fix various bugs in Configure. This affects the following platforms:
9446 unixware-2.0-pentium
9450 *) Eliminate generated files from CVS. Reorder tests to regenerate files
9451 before they are needed.
9454 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
9458 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
9460 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
9461 changed SSLeay to OpenSSL in version strings.
9462 [Ralf S. Engelschall]
9464 *) Some fixups to the top-level documents.
9467 *) Fixed the nasty bug where rsaref.h was not found under compile-time
9468 because the symlink to include/ was missing.
9469 [Ralf S. Engelschall]
9471 *) Incorporated the popular no-RSA/DSA-only patches
9472 which allow to compile a RSA-free SSLeay.
9473 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
9475 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
9476 when "ssleay" is still not found.
9477 [Ralf S. Engelschall]
9479 *) Added more platforms to Configure: Cray T3E, HPUX 11,
9480 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
9482 *) Updated the README file.
9483 [Ralf S. Engelschall]
9485 *) Added various .cvsignore files in the CVS repository subdirs
9486 to make a "cvs update" really silent.
9487 [Ralf S. Engelschall]
9489 *) Recompiled the error-definition header files and added
9490 missing symbols to the Win32 linker tables.
9491 [Ralf S. Engelschall]
9493 *) Cleaned up the top-level documents;
9494 o new files: CHANGES and LICENSE
9495 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
9496 o merged COPYRIGHT into LICENSE
9497 o removed obsolete TODO file
9498 o renamed MICROSOFT to INSTALL.W32
9499 [Ralf S. Engelschall]
9501 *) Removed dummy files from the 0.9.1b source tree:
9502 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
9503 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
9504 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
9505 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
9506 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
9507 [Ralf S. Engelschall]
9509 *) Added various platform portability fixes.
9512 *) The Genesis of the OpenSSL rpject:
9513 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
9514 Young and Tim J. Hudson created while they were working for C2Net until
9516 [The OpenSSL Project]
9519 Changes between 0.9.0b and 0.9.1b [not released]
9521 *) Updated a few CA certificates under certs/
9524 *) Changed some BIGNUM api stuff.
9527 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
9528 DGUX x86, Linux Alpha, etc.
9531 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
9532 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
9536 *) Add -strparse option to asn1pars program which parses nested
9538 [Dr Stephen Henson <shenson@bigfoot.com>]
9540 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
9543 *) DSA fix for "ca" program.
9546 *) Added "-genkey" option to "dsaparam" program.
9549 *) Added RIPE MD160 (rmd160) message digest.
9552 *) Added -a (all) option to "ssleay version" command.
9555 *) Added PLATFORM define which is the id given to Configure.
9558 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
9561 *) Extended the ASN.1 parser routines.
9564 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
9567 *) Added a BN_CTX to the BN library.
9570 *) Fixed the weak key values in DES library
9573 *) Changed API in EVP library for cipher aliases.
9576 *) Added support for RC2/64bit cipher.
9579 *) Converted the lhash library to the crypto/mem.c functions.
9582 *) Added more recognized ASN.1 object ids.
9585 *) Added more RSA padding checks for SSL/TLS.
9588 *) Added BIO proxy/filter functionality.
9591 *) Added extra_certs to SSL_CTX which can be used
9592 send extra CA certificates to the client in the CA cert chain sending
9593 process. It can be configured with SSL_CTX_add_extra_chain_cert().
9596 *) Now Fortezza is denied in the authentication phase because
9597 this is key exchange mechanism is not supported by SSLeay at all.
9600 *) Additional PKCS1 checks.
9603 *) Support the string "TLSv1" for all TLS v1 ciphers.
9606 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
9607 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
9610 *) Fixed a few memory leaks.
9613 *) Fixed various code and comment typos.
9616 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
9617 bytes sent in the client random.
9618 [Edward Bishop <ebishop@spyglass.com>]