5 Changes between 1.0.2e and 1.0.2f [xx XXX xxxx]
9 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
11 *) BN_mod_exp may produce incorrect results on x86_64
13 There is a carry propagating bug in the x86_64 Montgomery squaring
14 procedure. No EC algorithms are affected. Analysis suggests that attacks
15 against RSA and DSA as a result of this defect would be very difficult to
16 perform and are not believed likely. Attacks against DH are considered just
17 feasible (although very difficult) because most of the work necessary to
18 deduce information about a private key may be performed offline. The amount
19 of resources required for such an attack would be very significant and
20 likely only accessible to a limited number of attackers. An attacker would
21 additionally need online access to an unpatched system using the target
22 private key in a scenario with persistent DH parameters and a private
23 key that is shared between multiple clients. For example this can occur by
24 default in OpenSSL DHE based SSL/TLS ciphersuites.
26 This issue was reported to OpenSSL by Hanno Böck.
30 *) Certificate verify crash with missing PSS parameter
32 The signature verification routines will crash with a NULL pointer
33 dereference if presented with an ASN.1 signature using the RSA PSS
34 algorithm and absent mask generation function parameter. Since these
35 routines are used to verify certificate signature algorithms this can be
36 used to crash any certificate verification operation and exploited in a
37 DoS attack. Any application which performs certificate verification is
38 vulnerable including OpenSSL clients and servers which enable client
41 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
45 *) X509_ATTRIBUTE memory leak
47 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
48 memory. This structure is used by the PKCS#7 and CMS routines so any
49 application which reads PKCS#7 or CMS data from untrusted sources is
50 affected. SSL/TLS is not affected.
52 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
57 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
58 This changes the decoding behaviour for some invalid messages,
59 though the change is mostly in the more lenient direction, and
60 legacy behaviour is preserved as much as possible.
63 *) In DSA_generate_parameters_ex, if the provided seed is too short,
65 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
67 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
69 *) Alternate chains certificate forgery
71 During certificate verfification, OpenSSL will attempt to find an
72 alternative certificate chain if the first attempt to build such a chain
73 fails. An error in the implementation of this logic can mean that an
74 attacker could cause certain checks on untrusted certificates to be
75 bypassed, such as the CA flag, enabling them to use a valid leaf
76 certificate to act as a CA and "issue" an invalid certificate.
78 This issue was reported to OpenSSL by Adam Langley/David Benjamin
83 *) Race condition handling PSK identify hint
85 If PSK identity hints are received by a multi-threaded client then
86 the values are wrongly updated in the parent SSL_CTX structure. This can
87 result in a race condition potentially leading to a double free of the
92 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
94 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
95 incompatibility in the handling of HMAC. The previous ABI has now been
98 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
100 *) Malformed ECParameters causes infinite loop
102 When processing an ECParameters structure OpenSSL enters an infinite loop
103 if the curve specified is over a specially malformed binary polynomial
106 This can be used to perform denial of service against any
107 system which processes public keys, certificate requests or
108 certificates. This includes TLS clients and TLS servers with
109 client authentication enabled.
111 This issue was reported to OpenSSL by Joseph Barr-Pixton.
115 *) Exploitable out-of-bounds read in X509_cmp_time
117 X509_cmp_time does not properly check the length of the ASN1_TIME
118 string and can read a few bytes out of bounds. In addition,
119 X509_cmp_time accepts an arbitrary number of fractional seconds in the
122 An attacker can use this to craft malformed certificates and CRLs of
123 various sizes and potentially cause a segmentation fault, resulting in
124 a DoS on applications that verify certificates or CRLs. TLS clients
125 that verify CRLs are affected. TLS clients and servers with client
126 authentication enabled may be affected if they use custom verification
129 This issue was reported to OpenSSL by Robert Swiecki (Google), and
130 independently by Hanno Böck.
134 *) PKCS7 crash with missing EnvelopedContent
136 The PKCS#7 parsing code does not handle missing inner EncryptedContent
137 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
138 with missing content and trigger a NULL pointer dereference on parsing.
140 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
141 structures from untrusted sources are affected. OpenSSL clients and
142 servers are not affected.
144 This issue was reported to OpenSSL by Michal Zalewski (Google).
148 *) CMS verify infinite loop with unknown hash function
150 When verifying a signedData message the CMS code can enter an infinite loop
151 if presented with an unknown hash function OID. This can be used to perform
152 denial of service against any system which verifies signedData messages using
154 This issue was reported to OpenSSL by Johannes Bauer.
158 *) Race condition handling NewSessionTicket
160 If a NewSessionTicket is received by a multi-threaded client when attempting to
161 reuse a previous ticket then a race condition can occur potentially leading to
162 a double free of the ticket data.
166 *) Removed support for the two export grade static DH ciphersuites
167 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
168 were newly added (along with a number of other static DH ciphersuites) to
169 1.0.2. However the two export ones have *never* worked since they were
170 introduced. It seems strange in any case to be adding new export
171 ciphersuites, and given "logjam" it also does not seem correct to fix them.
174 *) Only support 256-bit or stronger elliptic curves with the
175 'ecdh_auto' setting (server) or by default (client). Of supported
176 curves, prefer P-256 (both).
179 *) Reject DH handshakes with parameters shorter than 768 bits.
180 [Kurt Roeckx and Emilia Kasper]
182 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
184 *) ClientHello sigalgs DoS fix
186 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
187 invalid signature algorithms extension a NULL pointer dereference will
188 occur. This can be exploited in a DoS attack against the server.
190 This issue was was reported to OpenSSL by David Ramos of Stanford
193 [Stephen Henson and Matt Caswell]
195 *) Multiblock corrupted pointer fix
197 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
198 feature only applies on 64 bit x86 architecture platforms that support AES
199 NI instructions. A defect in the implementation of "multiblock" can cause
200 OpenSSL's internal write buffer to become incorrectly set to NULL when
201 using non-blocking IO. Typically, when the user application is using a
202 socket BIO for writing, this will only result in a failed connection.
203 However if some other BIO is used then it is likely that a segmentation
204 fault will be triggered, thus enabling a potential DoS attack.
206 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
210 *) Segmentation fault in DTLSv1_listen fix
212 The DTLSv1_listen function is intended to be stateless and processes the
213 initial ClientHello from many peers. It is common for user code to loop
214 over the call to DTLSv1_listen until a valid ClientHello is received with
215 an associated cookie. A defect in the implementation of DTLSv1_listen means
216 that state is preserved in the SSL object from one invocation to the next
217 that can lead to a segmentation fault. Errors processing the initial
218 ClientHello can trigger this scenario. An example of such an error could be
219 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
222 This issue was reported to OpenSSL by Per Allansson.
226 *) Segmentation fault in ASN1_TYPE_cmp fix
228 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
229 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
230 certificate signature algorithm consistency this can be used to crash any
231 certificate verification operation and exploited in a DoS attack. Any
232 application which performs certificate verification is vulnerable including
233 OpenSSL clients and servers which enable client authentication.
237 *) Segmentation fault for invalid PSS parameters fix
239 The signature verification routines will crash with a NULL pointer
240 dereference if presented with an ASN.1 signature using the RSA PSS
241 algorithm and invalid parameters. Since these routines are used to verify
242 certificate signature algorithms this can be used to crash any
243 certificate verification operation and exploited in a DoS attack. Any
244 application which performs certificate verification is vulnerable including
245 OpenSSL clients and servers which enable client authentication.
247 This issue was was reported to OpenSSL by Brian Carpenter.
251 *) ASN.1 structure reuse memory corruption fix
253 Reusing a structure in ASN.1 parsing may allow an attacker to cause
254 memory corruption via an invalid write. Such reuse is and has been
255 strongly discouraged and is believed to be rare.
257 Applications that parse structures containing CHOICE or ANY DEFINED BY
258 components may be affected. Certificate parsing (d2i_X509 and related
259 functions) are however not affected. OpenSSL clients and servers are
264 *) PKCS7 NULL pointer dereferences fix
266 The PKCS#7 parsing code does not handle missing outer ContentInfo
267 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
268 missing content and trigger a NULL pointer dereference on parsing.
270 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
271 otherwise parse PKCS#7 structures from untrusted sources are
272 affected. OpenSSL clients and servers are not affected.
274 This issue was reported to OpenSSL by Michal Zalewski (Google).
278 *) DoS via reachable assert in SSLv2 servers fix
280 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
281 servers that both support SSLv2 and enable export cipher suites by sending
282 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
284 This issue was discovered by Sean Burford (Google) and Emilia Käsper
285 (OpenSSL development team).
289 *) Empty CKE with client auth and DHE fix
291 If client auth is used then a server can seg fault in the event of a DHE
292 ciphersuite being selected and a zero length ClientKeyExchange message
293 being sent by the client. This could be exploited in a DoS attack.
297 *) Handshake with unseeded PRNG fix
299 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
300 with an unseeded PRNG. The conditions are:
301 - The client is on a platform where the PRNG has not been seeded
302 automatically, and the user has not seeded manually
303 - A protocol specific client method version has been used (i.e. not
304 SSL_client_methodv23)
305 - A ciphersuite is used that does not require additional random data from
306 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
308 If the handshake succeeds then the client random that has been used will
309 have been generated from a PRNG with insufficient entropy and therefore the
310 output may be predictable.
312 For example using the following command with an unseeded openssl will
313 succeed on an unpatched platform:
315 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
319 *) Use After Free following d2i_ECPrivatekey error fix
321 A malformed EC private key file consumed via the d2i_ECPrivateKey function
322 could cause a use after free condition. This, in turn, could cause a double
323 free in several private key parsing functions (such as d2i_PrivateKey
324 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
325 for applications that receive EC private keys from untrusted
326 sources. This scenario is considered rare.
328 This issue was discovered by the BoringSSL project and fixed in their
333 *) X509_to_X509_REQ NULL pointer deref fix
335 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
336 the certificate key is invalid. This function is rarely used in practice.
338 This issue was discovered by Brian Carpenter.
342 *) Removed the export ciphers from the DEFAULT ciphers
345 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
347 *) Change RSA and DH/DSA key generation apps to generate 2048-bit
351 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
352 ARMv5 through ARMv8, as opposite to "locking" it to single one.
353 So far those who have to target multiple plaforms would compromise
354 and argue that binary targeting say ARMv5 would still execute on
355 ARMv8. "Universal" build resolves this compromise by providing
356 near-optimal performance even on newer platforms.
359 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
360 (other platforms pending).
361 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
363 *) Add support for the SignedCertificateTimestampList certificate and
364 OCSP response extensions from RFC6962.
367 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
368 for corner cases. (Certain input points at infinity could lead to
369 bogus results, with non-infinity inputs mapped to infinity too.)
372 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
373 This covers AES, SHA256/512 and GHASH. "Initial" means that most
374 common cases are optimized and there still is room for further
375 improvements. Vector Permutation AES for Altivec is also added.
378 *) Add support for little-endian ppc64 Linux target.
379 [Marcelo Cerri (IBM)]
381 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
382 SHA1, SHA256 and GHASH. "Initial" means that most common cases
383 are optimized and there still is room for further improvements.
384 Both 32- and 64-bit modes are supported.
385 [Andy Polyakov, Ard Biesheuvel (Linaro)]
387 *) Improved ARMv7 NEON support.
390 *) Support for SPARC Architecture 2011 crypto extensions, first
391 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
392 SHA256/512, MD5, GHASH and modular exponentiation.
393 [Andy Polyakov, David Miller]
395 *) Accelerated modular exponentiation for Intel processors, a.k.a.
397 [Shay Gueron & Vlad Krasnov (Intel Corp)]
399 *) Support for new and upcoming Intel processors, including AVX2,
400 BMI and SHA ISA extensions. This includes additional "stitched"
401 implementations, AESNI-SHA256 and GCM, and multi-buffer support
404 This work was sponsored by Intel Corp.
407 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
408 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
409 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
412 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
413 this fixes a limiation in previous versions of OpenSSL.
416 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
417 MGF1 digest and OAEP label.
420 *) Add EVP support for key wrapping algorithms, to avoid problems with
421 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
422 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
423 algorithms and include tests cases.
426 *) Add functions to allocate and set the fields of an ECDSA_METHOD
428 [Douglas E. Engert, Steve Henson]
430 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
431 difference in days and seconds between two tm or ASN1_TIME structures.
434 *) Add -rev test option to s_server to just reverse order of characters
435 received by client and send back to server. Also prints an abbreviated
436 summary of the connection parameters.
439 *) New option -brief for s_client and s_server to print out a brief summary
440 of connection parameters.
443 *) Add callbacks for arbitrary TLS extensions.
444 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
446 *) New option -crl_download in several openssl utilities to download CRLs
447 from CRLDP extension in certificates.
450 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
453 *) New function X509_CRL_diff to generate a delta CRL from the difference
454 of two full CRLs. Add support to "crl" utility.
457 *) New functions to set lookup_crls function and to retrieve
458 X509_STORE from X509_STORE_CTX.
461 *) Print out deprecated issuer and subject unique ID fields in
465 *) Extend OCSP I/O functions so they can be used for simple general purpose
466 HTTP as well as OCSP. New wrapper function which can be used to download
467 CRLs using the OCSP API.
470 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
473 *) SSL_CONF* functions. These provide a common framework for application
474 configuration using configuration files or command lines.
477 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
478 message callback and prints the results. Needs compile time option
479 "enable-ssl-trace". New options to s_client and s_server to enable
483 *) New ctrl and macro to retrieve supported points extensions.
484 Print out extension in s_server and s_client.
487 *) New functions to retrieve certificate signature and signature
491 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
495 *) New Suite B modes for TLS code. These use and enforce the requirements
496 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
497 only use Suite B curves. The Suite B modes can be set by using the
498 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
501 *) New chain verification flags for Suite B levels of security. Check
502 algorithms are acceptable when flags are set in X509_verify_cert.
505 *) Make tls1_check_chain return a set of flags indicating checks passed
506 by a certificate chain. Add additional tests to handle client
507 certificates: checks for matching certificate type and issuer name
511 *) If an attempt is made to use a signature algorithm not in the peer
512 preference list abort the handshake. If client has no suitable
513 signature algorithms in response to a certificate request do not
517 *) If server EC tmp key is not in client preference list abort handshake.
520 *) Add support for certificate stores in CERT structure. This makes it
521 possible to have different stores per SSL structure or one store in
522 the parent SSL_CTX. Include distint stores for certificate chain
523 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
524 to build and store a certificate chain in CERT structure: returing
525 an error if the chain cannot be built: this will allow applications
526 to test if a chain is correctly configured.
528 Note: if the CERT based stores are not set then the parent SSL_CTX
529 store is used to retain compatibility with existing behaviour.
533 *) New function ssl_set_client_disabled to set a ciphersuite disabled
534 mask based on the current session, check mask when sending client
535 hello and checking the requested ciphersuite.
538 *) New ctrls to retrieve and set certificate types in a certificate
539 request message. Print out received values in s_client. If certificate
540 types is not set with custom values set sensible values based on
541 supported signature algorithms.
544 *) Support for distinct client and server supported signature algorithms.
547 *) Add certificate callback. If set this is called whenever a certificate
548 is required by client or server. An application can decide which
549 certificate chain to present based on arbitrary criteria: for example
550 supported signature algorithms. Add very simple example to s_server.
551 This fixes many of the problems and restrictions of the existing client
552 certificate callback: for example you can now clear an existing
553 certificate and specify the whole chain.
556 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
557 the certificate can be used for (if anything). Set valid_flags field
558 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
559 to have similar checks in it.
561 Add new "cert_flags" field to CERT structure and include a "strict mode".
562 This enforces some TLS certificate requirements (such as only permitting
563 certificate signature algorithms contained in the supported algorithms
564 extension) which some implementations ignore: this option should be used
565 with caution as it could cause interoperability issues.
568 *) Update and tidy signature algorithm extension processing. Work out
569 shared signature algorithms based on preferences and peer algorithms
570 and print them out in s_client and s_server. Abort handshake if no
571 shared signature algorithms.
574 *) Add new functions to allow customised supported signature algorithms
575 for SSL and SSL_CTX structures. Add options to s_client and s_server
579 *) New function SSL_certs_clear() to delete all references to certificates
580 from an SSL structure. Before this once a certificate had been added
581 it couldn't be removed.
584 *) Integrate hostname, email address and IP address checking with certificate
585 verification. New verify options supporting checking in opensl utility.
588 *) Fixes and wildcard matching support to hostname and email checking
589 functions. Add manual page.
590 [Florian Weimer (Red Hat Product Security Team)]
592 *) New functions to check a hostname email or IP address against a
593 certificate. Add options x509 utility to print results of checks against
597 *) Fix OCSP checking.
598 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
600 *) Initial experimental support for explicitly trusted non-root CAs.
601 OpenSSL still tries to build a complete chain to a root but if an
602 intermediate CA has a trust setting included that is used. The first
603 setting is used: whether to trust (e.g., -addtrust option to the x509
607 *) Add -trusted_first option which attempts to find certificates in the
608 trusted store even if an untrusted chain is also supplied.
611 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
612 platform support for Linux and Android.
615 *) Support for linux-x32, ILP32 environment in x86_64 framework.
618 *) Experimental multi-implementation support for FIPS capable OpenSSL.
619 When in FIPS mode the approved implementations are used as normal,
620 when not in FIPS mode the internal unapproved versions are used instead.
621 This means that the FIPS capable OpenSSL isn't forced to use the
622 (often lower perfomance) FIPS implementations outside FIPS mode.
625 *) Transparently support X9.42 DH parameters when calling
626 PEM_read_bio_DHparameters. This means existing applications can handle
627 the new parameter format automatically.
630 *) Initial experimental support for X9.42 DH parameter format: mainly
631 to support use of 'q' parameter for RFC5114 parameters.
634 *) Add DH parameters from RFC5114 including test data to dhtest.
637 *) Support for automatic EC temporary key parameter selection. If enabled
638 the most preferred EC parameters are automatically used instead of
639 hardcoded fixed parameters. Now a server just has to call:
640 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
641 support ECDH and use the most appropriate parameters.
644 *) Enhance and tidy EC curve and point format TLS extension code. Use
645 static structures instead of allocation if default values are used.
646 New ctrls to set curves we wish to support and to retrieve shared curves.
647 Print out shared curves in s_server. New options to s_server and s_client
648 to set list of supported curves.
651 *) New ctrls to retrieve supported signature algorithms and
652 supported curve values as an array of NIDs. Extend openssl utility
653 to print out received values.
656 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
657 between NIDs and the more common NIST names such as "P-256". Enhance
658 ecparam utility and ECC method to recognise the NIST names for curves.
661 *) Enhance SSL/TLS certificate chain handling to support different
662 chains for each certificate instead of one chain in the parent SSL_CTX.
665 *) Support for fixed DH ciphersuite client authentication: where both
666 server and client use DH certificates with common parameters.
669 *) Support for fixed DH ciphersuites: those requiring DH server
673 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
675 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
676 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
677 X509_CINF_get_signature were reverted post internal team review.
679 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
681 *) Build fixes for the Windows and OpenVMS platforms
682 [Matt Caswell and Richard Levitte]
684 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
686 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
687 message can cause a segmentation fault in OpenSSL due to a NULL pointer
688 dereference. This could lead to a Denial Of Service attack. Thanks to
689 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
693 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
694 dtls1_buffer_record function under certain conditions. In particular this
695 could occur if an attacker sent repeated DTLS records with the same
696 sequence number but for the next epoch. The memory leak could be exploited
697 by an attacker in a Denial of Service attack through memory exhaustion.
698 Thanks to Chris Mueller for reporting this issue.
702 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
703 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
704 method would be set to NULL which could later result in a NULL pointer
705 dereference. Thanks to Frank Schmirler for reporting this issue.
709 *) Abort handshake if server key exchange message is omitted for ephemeral
712 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
713 reporting this issue.
717 *) Remove non-export ephemeral RSA code on client and server. This code
718 violated the TLS standard by allowing the use of temporary RSA keys in
719 non-export ciphersuites and could be used by a server to effectively
720 downgrade the RSA key length used to a value smaller than the server
721 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
722 INRIA or reporting this issue.
726 *) Fixed issue where DH client certificates are accepted without verification.
727 An OpenSSL server will accept a DH certificate for client authentication
728 without the certificate verify message. This effectively allows a client to
729 authenticate without the use of a private key. This only affects servers
730 which trust a client certificate authority which issues certificates
731 containing DH keys: these are extremely rare and hardly ever encountered.
732 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
737 *) Ensure that the session ID context of an SSL is updated when its
738 SSL_CTX is updated via SSL_set_SSL_CTX.
740 The session ID context is typically set from the parent SSL_CTX,
741 and can vary with the CTX.
744 *) Fix various certificate fingerprint issues.
746 By using non-DER or invalid encodings outside the signed portion of a
747 certificate the fingerprint can be changed without breaking the signature.
748 Although no details of the signed portion of the certificate can be changed
749 this can cause problems with some applications: e.g. those using the
750 certificate fingerprint for blacklists.
752 1. Reject signatures with non zero unused bits.
754 If the BIT STRING containing the signature has non zero unused bits reject
755 the signature. All current signature algorithms require zero unused bits.
757 2. Check certificate algorithm consistency.
759 Check the AlgorithmIdentifier inside TBS matches the one in the
760 certificate signature. NB: this will result in signature failure
761 errors for some broken certificates.
763 Thanks to Konrad Kraszewski from Google for reporting this issue.
765 3. Check DSA/ECDSA signatures use DER.
767 Reencode DSA/ECDSA signatures and compare with the original received
768 signature. Return an error if there is a mismatch.
770 This will reject various cases including garbage after signature
771 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
772 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
773 (negative or with leading zeroes).
775 Further analysis was conducted and fixes were developed by Stephen Henson
776 of the OpenSSL core team.
781 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
782 results on some platforms, including x86_64. This bug occurs at random
783 with a very low probability, and is not known to be exploitable in any
784 way, though its exact impact is difficult to determine. Thanks to Pieter
785 Wuille (Blockstream) who reported this issue and also suggested an initial
786 fix. Further analysis was conducted by the OpenSSL development team and
787 Adam Langley of Google. The final fix was developed by Andy Polyakov of
788 the OpenSSL core team.
792 *) Do not resume sessions on the server if the negotiated protocol
793 version does not match the session's version. Resuming with a different
794 version, while not strictly forbidden by the RFC, is of questionable
795 sanity and breaks all known clients.
796 [David Benjamin, Emilia Käsper]
798 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
799 early CCS messages during renegotiation. (Note that because
800 renegotiation is encrypted, this early CCS was not exploitable.)
803 *) Tighten client-side session ticket handling during renegotiation:
804 ensure that the client only accepts a session ticket if the server sends
805 the extension anew in the ServerHello. Previously, a TLS client would
806 reuse the old extension state and thus accept a session ticket if one was
807 announced in the initial ServerHello.
809 Similarly, ensure that the client requires a session ticket if one
810 was advertised in the ServerHello. Previously, a TLS client would
811 ignore a missing NewSessionTicket message.
814 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
818 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
819 sends a carefully crafted handshake message, to cause OpenSSL to fail
820 to free up to 64k of memory causing a memory leak. This could be
821 exploited in a Denial Of Service attack. This issue affects OpenSSL
822 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
823 whether SRTP is used or configured. Implementations of OpenSSL that
824 have been compiled with OPENSSL_NO_SRTP defined are not affected.
826 The fix was developed by the OpenSSL team.
830 *) Session Ticket Memory Leak.
832 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
833 integrity of that ticket is first verified. In the event of a session
834 ticket integrity check failing, OpenSSL will fail to free memory
835 causing a memory leak. By sending a large number of invalid session
836 tickets an attacker could exploit this issue in a Denial Of Service
841 *) Build option no-ssl3 is incomplete.
843 When OpenSSL is configured with "no-ssl3" as a build option, servers
844 could accept and complete a SSL 3.0 handshake, and clients could be
845 configured to send them.
847 [Akamai and the OpenSSL team]
849 *) Add support for TLS_FALLBACK_SCSV.
850 Client applications doing fallback retries should call
851 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
853 [Adam Langley, Bodo Moeller]
855 *) Add additional DigestInfo checks.
857 Reencode DigestInto in DER and check against the original when
858 verifying RSA signature: this will reject any improperly encoded
859 DigestInfo structures.
861 Note: this is a precautionary measure and no attacks are currently known.
865 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
867 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
868 SRP code can be overrun an internal buffer. Add sanity check that
869 g, A, B < N to SRP code.
871 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
872 Group for discovering this issue.
876 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
877 TLS 1.0 instead of higher protocol versions when the ClientHello message
878 is badly fragmented. This allows a man-in-the-middle attacker to force a
879 downgrade to TLS 1.0 even if both the server and the client support a
880 higher protocol version, by modifying the client's TLS records.
882 Thanks to David Benjamin and Adam Langley (Google) for discovering and
883 researching this issue.
887 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
888 to a denial of service attack. A malicious server can crash the client
889 with a null pointer dereference (read) by specifying an anonymous (EC)DH
890 ciphersuite and sending carefully crafted handshake messages.
892 Thanks to Felix Gröbert (Google) for discovering and researching this
897 *) By sending carefully crafted DTLS packets an attacker could cause openssl
898 to leak memory. This can be exploited through a Denial of Service attack.
899 Thanks to Adam Langley for discovering and researching this issue.
903 *) An attacker can force openssl to consume large amounts of memory whilst
904 processing DTLS handshake messages. This can be exploited through a
905 Denial of Service attack.
906 Thanks to Adam Langley for discovering and researching this issue.
910 *) An attacker can force an error condition which causes openssl to crash
911 whilst processing DTLS packets due to memory being freed twice. This
912 can be exploited through a Denial of Service attack.
913 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
918 *) If a multithreaded client connects to a malicious server using a resumed
919 session and the server sends an ec point format extension it could write
920 up to 255 bytes to freed memory.
922 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
927 *) A malicious server can crash an OpenSSL client with a null pointer
928 dereference (read) by specifying an SRP ciphersuite even though it was not
929 properly negotiated with the client. This can be exploited through a
930 Denial of Service attack.
932 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
933 discovering and researching this issue.
937 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
938 X509_name_oneline, X509_name_print_ex et al. to leak some information
939 from the stack. Applications may be affected if they echo pretty printing
940 output to the attacker.
942 Thanks to Ivan Fratric (Google) for discovering this issue.
944 [Emilia Käsper, and Steve Henson]
946 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
947 for corner cases. (Certain input points at infinity could lead to
948 bogus results, with non-infinity inputs mapped to infinity too.)
951 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
953 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
954 handshake can force the use of weak keying material in OpenSSL
955 SSL/TLS clients and servers.
957 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
958 researching this issue. (CVE-2014-0224)
959 [KIKUCHI Masashi, Steve Henson]
961 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
962 OpenSSL DTLS client the code can be made to recurse eventually crashing
965 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
967 [Imre Rad, Steve Henson]
969 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
970 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
971 client or server. This is potentially exploitable to run arbitrary
972 code on a vulnerable client or server.
974 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
975 [Jüri Aedla, Steve Henson]
977 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
978 are subject to a denial of service attack.
980 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
981 this issue. (CVE-2014-3470)
982 [Felix Gröbert, Ivan Fratric, Steve Henson]
984 *) Harmonize version and its documentation. -f flag is used to display
986 [mancha <mancha1@zoho.com>]
988 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
989 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
990 [mancha <mancha1@zoho.com>]
992 *) Fix some double frees. These are not thought to be exploitable.
993 [mancha <mancha1@zoho.com>]
995 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
997 *) A missing bounds check in the handling of the TLS heartbeat extension
998 can be used to reveal up to 64k of memory to a connected client or
1001 Thanks for Neel Mehta of Google Security for discovering this bug and to
1002 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1003 preparing the fix (CVE-2014-0160)
1004 [Adam Langley, Bodo Moeller]
1006 *) Fix for the attack described in the paper "Recovering OpenSSL
1007 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1008 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1009 http://eprint.iacr.org/2014/140
1011 Thanks to Yuval Yarom and Naomi Benger for discovering this
1012 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1013 [Yuval Yarom and Naomi Benger]
1015 *) TLS pad extension: draft-agl-tls-padding-03
1017 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
1018 TLS client Hello record length value would otherwise be > 255 and
1019 less that 512 pad with a dummy extension containing zeroes so it
1020 is at least 512 bytes long.
1022 [Adam Langley, Steve Henson]
1024 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
1026 *) Fix for TLS record tampering bug. A carefully crafted invalid
1027 handshake could crash OpenSSL with a NULL pointer exception.
1028 Thanks to Anton Johansson for reporting this issues.
1031 *) Keep original DTLS digest and encryption contexts in retransmission
1032 structures so we can use the previous session parameters if they need
1033 to be resent. (CVE-2013-6450)
1036 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
1037 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
1038 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
1039 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
1040 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
1041 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
1042 [Rob Stradling, Adam Langley]
1044 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
1046 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
1047 supporting platforms or when small records were transferred.
1048 [Andy Polyakov, Steve Henson]
1050 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
1052 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1054 This addresses the flaw in CBC record processing discovered by
1055 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1056 at: http://www.isg.rhul.ac.uk/tls/
1058 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1059 Security Group at Royal Holloway, University of London
1060 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1061 Emilia Käsper for the initial patch.
1063 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1065 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
1066 ciphersuites which can be exploited in a denial of service attack.
1067 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
1068 and detecting this bug and to Wolfgang Ettlinger
1069 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
1073 *) Return an error when checking OCSP signatures when key is NULL.
1074 This fixes a DoS attack. (CVE-2013-0166)
1077 *) Make openssl verify return errors.
1078 [Chris Palmer <palmer@google.com> and Ben Laurie]
1080 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1081 the right response is stapled. Also change SSL_get_certificate()
1082 so it returns the certificate actually sent.
1083 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1084 [Rob Stradling <rob.stradling@comodo.com>]
1086 *) Fix possible deadlock when decoding public keys.
1089 *) Don't use TLS 1.0 record version number in initial client hello
1093 Changes between 1.0.1b and 1.0.1c [10 May 2012]
1095 *) Sanity check record length before skipping explicit IV in TLS
1096 1.2, 1.1 and DTLS to fix DoS attack.
1098 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1099 fuzzing as a service testing platform.
1103 *) Initialise tkeylen properly when encrypting CMS messages.
1104 Thanks to Solar Designer of Openwall for reporting this issue.
1107 *) In FIPS mode don't try to use composite ciphers as they are not
1111 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
1113 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
1114 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
1115 mean any application compiled against OpenSSL 1.0.0 headers setting
1116 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
1117 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
1118 0x10000000L Any application which was previously compiled against
1119 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
1120 will need to be recompiled as a result. Letting be results in
1121 inability to disable specifically TLS 1.1 and in client context,
1122 in unlike event, limit maximum offered version to TLS 1.0 [see below].
1125 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
1126 disable just protocol X, but all protocols above X *if* there are
1127 protocols *below* X still enabled. In more practical terms it means
1128 that if application wants to disable TLS1.0 in favor of TLS1.1 and
1129 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
1130 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
1134 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
1136 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1137 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1138 in CRYPTO_realloc_clean.
1140 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1141 issue and to Adam Langley <agl@chromium.org> for fixing it.
1143 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1145 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
1148 *) Workarounds for some broken servers that "hang" if a client hello
1149 record length exceeds 255 bytes.
1151 1. Do not use record version number > TLS 1.0 in initial client
1152 hello: some (but not all) hanging servers will now work.
1153 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
1154 the number of ciphers sent in the client hello. This should be
1155 set to an even number, such as 50, for example by passing:
1156 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
1157 Most broken servers should now work.
1158 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
1159 TLS 1.2 client support entirely.
1162 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
1165 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
1167 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
1168 STRING form instead of a DigestInfo.
1171 *) The format used for MDC2 RSA signatures is inconsistent between EVP
1172 and the RSA_sign/RSA_verify functions. This was made more apparent when
1173 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
1174 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
1175 the correct format in RSA_verify so both forms transparently work.
1178 *) Some servers which support TLS 1.0 can choke if we initially indicate
1179 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
1180 encrypted premaster secret. As a workaround use the maximum pemitted
1181 client version in client hello, this should keep such servers happy
1182 and still work with previous versions of OpenSSL.
1185 *) Add support for TLS/DTLS heartbeats.
1186 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1188 *) Add support for SCTP.
1189 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1191 *) Improved PRNG seeding for VOS.
1192 [Paul Green <Paul.Green@stratus.com>]
1194 *) Extensive assembler packs updates, most notably:
1196 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
1197 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
1198 - x86_64: bit-sliced AES implementation;
1199 - ARM: NEON support, contemporary platforms optimizations;
1200 - s390x: z196 support;
1201 - *: GHASH and GF(2^m) multiplication implementations;
1205 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
1206 (removal of unnecessary code)
1207 [Peter Sylvester <peter.sylvester@edelweb.fr>]
1209 *) Add TLS key material exporter from RFC 5705.
1212 *) Add DTLS-SRTP negotiation from RFC 5764.
1215 *) Add Next Protocol Negotiation,
1216 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
1217 disabled with a no-npn flag to config or Configure. Code donated
1219 [Adam Langley <agl@google.com> and Ben Laurie]
1221 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
1222 NIST-P256, NIST-P521, with constant-time single point multiplication on
1223 typical inputs. Compiler support for the nonstandard type __uint128_t is
1224 required to use this (present in gcc 4.4 and later, for 64-bit builds).
1225 Code made available under Apache License version 2.0.
1227 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
1228 line to include this in your build of OpenSSL, and run "make depend" (or
1229 "make update"). This enables the following EC_METHODs:
1231 EC_GFp_nistp224_method()
1232 EC_GFp_nistp256_method()
1233 EC_GFp_nistp521_method()
1235 EC_GROUP_new_by_curve_name() will automatically use these (while
1236 EC_GROUP_new_curve_GFp() currently prefers the more flexible
1238 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
1240 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
1241 all platforms. Move ssize_t definition from e_os.h to the public
1242 header file e_os2.h as it now appears in public header file cms.h
1245 *) New -sigopt option to the ca, req and x509 utilities. Additional
1246 signature parameters can be passed using this option and in
1250 *) Add RSA PSS signing function. This will generate and set the
1251 appropriate AlgorithmIdentifiers for PSS based on those in the
1252 corresponding EVP_MD_CTX structure. No application support yet.
1255 *) Support for companion algorithm specific ASN1 signing routines.
1256 New function ASN1_item_sign_ctx() signs a pre-initialised
1257 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
1258 the appropriate parameters.
1261 *) Add new algorithm specific ASN1 verification initialisation function
1262 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
1263 handling will be the same no matter what EVP_PKEY_METHOD is used.
1264 Add a PSS handler to support verification of PSS signatures: checked
1265 against a number of sample certificates.
1268 *) Add signature printing for PSS. Add PSS OIDs.
1269 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
1271 *) Add algorithm specific signature printing. An individual ASN1 method
1272 can now print out signatures instead of the standard hex dump.
1274 More complex signatures (e.g. PSS) can print out more meaningful
1275 information. Include DSA version that prints out the signature
1279 *) Password based recipient info support for CMS library: implementing
1283 *) Split password based encryption into PBES2 and PBKDF2 functions. This
1284 neatly separates the code into cipher and PBE sections and is required
1285 for some algorithms that split PBES2 into separate pieces (such as
1286 password based CMS).
1289 *) Session-handling fixes:
1290 - Fix handling of connections that are resuming with a session ID,
1291 but also support Session Tickets.
1292 - Fix a bug that suppressed issuing of a new ticket if the client
1293 presented a ticket with an expired session.
1294 - Try to set the ticket lifetime hint to something reasonable.
1295 - Make tickets shorter by excluding irrelevant information.
1296 - On the client side, don't ignore renewed tickets.
1297 [Adam Langley, Bodo Moeller (Google)]
1299 *) Fix PSK session representation.
1302 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
1304 This work was sponsored by Intel.
1307 *) Add GCM support to TLS library. Some custom code is needed to split
1308 the IV between the fixed (from PRF) and explicit (from TLS record)
1309 portions. This adds all GCM ciphersuites supported by RFC5288 and
1310 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
1311 add a special AESGCM string for GCM only.
1314 *) Expand range of ctrls for AES GCM. Permit setting invocation
1315 field on decrypt and retrieval of invocation field only on encrypt.
1318 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
1319 As required by RFC5289 these ciphersuites cannot be used if for
1320 versions of TLS earlier than 1.2.
1323 *) For FIPS capable OpenSSL interpret a NULL default public key method
1324 as unset and return the appopriate default but do *not* set the default.
1325 This means we can return the appopriate method in applications that
1326 swicth between FIPS and non-FIPS modes.
1329 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
1330 ENGINE is used then we cannot handle that in the FIPS module so we
1331 keep original code iff non-FIPS operations are allowed.
1334 *) Add -attime option to openssl utilities.
1335 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
1337 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
1340 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
1341 FIPS EC methods unconditionally for now.
1344 *) New build option no-ec2m to disable characteristic 2 code.
1347 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
1348 all cases can be covered as some introduce binary incompatibilities.
1351 *) Redirect RSA operations to FIPS module including keygen,
1352 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
1355 *) Add similar low level API blocking to ciphers.
1358 *) Low level digest APIs are not approved in FIPS mode: any attempt
1359 to use these will cause a fatal error. Applications that *really* want
1360 to use them can use the private_* version instead.
1363 *) Redirect cipher operations to FIPS module for FIPS builds.
1366 *) Redirect digest operations to FIPS module for FIPS builds.
1369 *) Update build system to add "fips" flag which will link in fipscanister.o
1370 for static and shared library builds embedding a signature if needed.
1373 *) Output TLS supported curves in preference order instead of numerical
1374 order. This is currently hardcoded for the highest order curves first.
1375 This should be configurable so applications can judge speed vs strength.
1378 *) Add TLS v1.2 server support for client authentication.
1381 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
1385 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
1386 FIPS modules versions.
1389 *) Add TLS v1.2 client side support for client authentication. Keep cache
1390 of handshake records longer as we don't know the hash algorithm to use
1391 until after the certificate request message is received.
1394 *) Initial TLS v1.2 client support. Add a default signature algorithms
1395 extension including all the algorithms we support. Parse new signature
1396 format in client key exchange. Relax some ECC signing restrictions for
1397 TLS v1.2 as indicated in RFC5246.
1400 *) Add server support for TLS v1.2 signature algorithms extension. Switch
1401 to new signature format when needed using client digest preference.
1402 All server ciphersuites should now work correctly in TLS v1.2. No client
1403 support yet and no support for client certificates.
1406 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
1407 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
1408 ciphersuites. At present only RSA key exchange ciphersuites work with
1409 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
1410 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
1411 and version checking.
1414 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
1415 with this defined it will not be affected by any changes to ssl internal
1416 structures. Add several utility functions to allow openssl application
1417 to work with OPENSSL_NO_SSL_INTERN defined.
1421 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
1423 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
1426 *) Permit abbreviated handshakes when renegotiating using the function
1427 SSL_renegotiate_abbreviated().
1428 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1430 *) Add call to ENGINE_register_all_complete() to
1431 ENGINE_load_builtin_engines(), so some implementations get used
1432 automatically instead of needing explicit application support.
1435 *) Add support for TLS key exporter as described in RFC5705.
1436 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
1438 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
1439 a few changes are required:
1441 Add SSL_OP_NO_TLSv1_1 flag.
1442 Add TLSv1_1 methods.
1443 Update version checking logic to handle version 1.1.
1444 Add explicit IV handling (ported from DTLS code).
1445 Add command line options to s_client/s_server.
1448 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
1450 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1451 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1452 content decryption and always return the same error. Note: this attack
1453 needs on average 2^20 messages so it only affects automated senders. The
1454 old behaviour can be reenabled in the CMS code by setting the
1455 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1456 an MMA defence is not necessary.
1457 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1458 this issue. (CVE-2012-0884)
1461 *) Fix CVE-2011-4619: make sure we really are receiving a
1462 client hello before rejecting multiple SGC restarts. Thanks to
1463 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1466 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
1468 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1469 Thanks to Antonio Martin, Enterprise Secure Access Research and
1470 Development, Cisco Systems, Inc. for discovering this bug and
1471 preparing a fix. (CVE-2012-0050)
1474 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
1476 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1477 of the Vaudenay padding oracle attack on CBC mode encryption
1478 which enables an efficient plaintext recovery attack against
1479 the OpenSSL implementation of DTLS. Their attack exploits timing
1480 differences arising during decryption processing. A research
1481 paper describing this attack can be found at:
1482 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1483 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1484 Security Group at Royal Holloway, University of London
1485 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1486 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1487 for preparing the fix. (CVE-2011-4108)
1488 [Robin Seggelmann, Michael Tuexen]
1490 *) Clear bytes used for block padding of SSL 3.0 records.
1492 [Adam Langley (Google)]
1494 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1495 Kadianakis <desnacked@gmail.com> for discovering this issue and
1496 Adam Langley for preparing the fix. (CVE-2011-4619)
1497 [Adam Langley (Google)]
1499 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
1500 [Andrey Kulikov <amdeich@gmail.com>]
1502 *) Prevent malformed RFC3779 data triggering an assertion failure.
1503 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1504 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1505 [Rob Austein <sra@hactrn.net>]
1507 *) Improved PRNG seeding for VOS.
1508 [Paul Green <Paul.Green@stratus.com>]
1510 *) Fix ssl_ciph.c set-up race.
1511 [Adam Langley (Google)]
1513 *) Fix spurious failures in ecdsatest.c.
1514 [Emilia Käsper (Google)]
1516 *) Fix the BIO_f_buffer() implementation (which was mixing different
1517 interpretations of the '..._len' fields).
1518 [Adam Langley (Google)]
1520 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1521 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1522 threads won't reuse the same blinding coefficients.
1524 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1525 lock to call BN_BLINDING_invert_ex, and avoids one use of
1526 BN_BLINDING_update for each BN_BLINDING structure (previously,
1527 the last update always remained unused).
1528 [Emilia Käsper (Google)]
1530 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
1531 [Bob Buckholz (Google)]
1533 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
1535 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
1536 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
1537 [Kaspar Brand <ossl@velox.ch>]
1539 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1540 for multi-threaded use of ECDH. (CVE-2011-3210)
1541 [Adam Langley (Google)]
1543 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1546 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
1547 signature public key algorithm by using OID xref utilities instead.
1548 Before this you could only use some ECC ciphersuites with SHA1 only.
1551 *) Add protection against ECDSA timing attacks as mentioned in the paper
1552 by Billy Bob Brumley and Nicola Tuveri, see:
1554 http://eprint.iacr.org/2011/232.pdf
1556 [Billy Bob Brumley and Nicola Tuveri]
1558 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
1560 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1561 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1563 *) Fix bug in string printing code: if *any* escaping is enabled we must
1564 escape the escape character (backslash) or the resulting string is
1568 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
1570 *) Disable code workaround for ancient and obsolete Netscape browsers
1571 and servers: an attacker can use it in a ciphersuite downgrade attack.
1572 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1575 *) Fixed J-PAKE implementation error, originally discovered by
1576 Sebastien Martini, further info and confirmation from Stefan
1577 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1580 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
1582 *) Fix extension code to avoid race conditions which can result in a buffer
1583 overrun vulnerability: resumed sessions must not be modified as they can
1584 be shared by multiple threads. CVE-2010-3864
1587 *) Fix WIN32 build system to correctly link an ENGINE directory into
1591 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
1593 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
1595 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
1597 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
1599 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
1600 context. The operation can be customised via the ctrl mechanism in
1601 case ENGINEs want to include additional functionality.
1604 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
1607 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
1608 output hashes compatible with older versions of OpenSSL.
1609 [Willy Weisz <weisz@vcpc.univie.ac.at>]
1611 *) Fix compression algorithm handling: if resuming a session use the
1612 compression algorithm of the resumed session instead of determining
1613 it from client hello again. Don't allow server to change algorithm.
1616 *) Add load_crls() function to apps tidying load_certs() too. Add option
1617 to verify utility to allow additional CRLs to be included.
1620 *) Update OCSP request code to permit adding custom headers to the request:
1621 some responders need this.
1624 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
1626 [Julia Lawall <julia@diku.dk>]
1628 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
1629 needlessly dereferenced structures, used obsolete functions and
1630 didn't handle all updated verify codes correctly.
1633 *) Disable MD2 in the default configuration.
1636 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
1637 indicate the initial BIO being pushed or popped. This makes it possible
1638 to determine whether the BIO is the one explicitly called or as a result
1639 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
1640 it handles reference counts correctly and doesn't zero out the I/O bio
1641 when it is not being explicitly popped. WARNING: applications which
1642 included workarounds for the old buggy behaviour will need to be modified
1643 or they could free up already freed BIOs.
1646 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
1647 renaming to all platforms (within the 0.9.8 branch, this was
1648 done conditionally on Netware platforms to avoid a name clash).
1649 [Guenter <lists@gknw.net>]
1651 *) Add ECDHE and PSK support to DTLS.
1652 [Michael Tuexen <tuexen@fh-muenster.de>]
1654 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
1658 *) Add "missing" function EVP_MD_flags() (without this the only way to
1659 retrieve a digest flags is by accessing the structure directly. Update
1660 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
1661 or cipher is registered as in the "from" argument. Print out all
1662 registered digests in the dgst usage message instead of manually
1663 attempting to work them out.
1666 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
1667 this allows the use of compression and extensions. Change default cipher
1668 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
1669 by default unless an application cipher string requests it.
1672 *) Alter match criteria in PKCS12_parse(). It used to try to use local
1673 key ids to find matching certificates and keys but some PKCS#12 files
1674 don't follow the (somewhat unwritten) rules and this strategy fails.
1675 Now just gather all certificates together and the first private key
1676 then look for the first certificate that matches the key.
1679 *) Support use of registered digest and cipher names for dgst and cipher
1680 commands instead of having to add each one as a special case. So now
1687 openssl dgst -sha256 foo
1689 and this works for ENGINE based algorithms too.
1693 *) Update Gost ENGINE to support parameter files.
1694 [Victor B. Wagner <vitus@cryptocom.ru>]
1696 *) Support GeneralizedTime in ca utility.
1697 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
1699 *) Enhance the hash format used for certificate directory links. The new
1700 form uses the canonical encoding (meaning equivalent names will work
1701 even if they aren't identical) and uses SHA1 instead of MD5. This form
1702 is incompatible with the older format and as a result c_rehash should
1703 be used to rebuild symbolic links.
1706 *) Make PKCS#8 the default write format for private keys, replacing the
1707 traditional format. This form is standardised, more secure and doesn't
1708 include an implicit MD5 dependency.
1711 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
1712 committed to OpenSSL should pass this lot as a minimum.
1715 *) Add session ticket override functionality for use by EAP-FAST.
1716 [Jouni Malinen <j@w1.fi>]
1718 *) Modify HMAC functions to return a value. Since these can be implemented
1719 in an ENGINE errors can occur.
1722 *) Type-checked OBJ_bsearch_ex.
1725 *) Type-checked OBJ_bsearch. Also some constification necessitated
1726 by type-checking. Still to come: TXT_DB, bsearch(?),
1727 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
1731 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
1732 seconds to a tm structure directly, instead of going through OS
1733 specific date routines. This avoids any issues with OS routines such
1734 as the year 2038 bug. New *_adj() functions for ASN1 time structures
1735 and X509_time_adj_ex() to cover the extended range. The existing
1736 X509_time_adj() is still usable and will no longer have any date issues.
1739 *) Delta CRL support. New use deltas option which will attempt to locate
1740 and search any appropriate delta CRLs available.
1742 This work was sponsored by Google.
1745 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
1746 code and add additional score elements. Validate alternate CRL paths
1747 as part of the CRL checking and indicate a new error "CRL path validation
1748 error" in this case. Applications wanting additional details can use
1749 the verify callback and check the new "parent" field. If this is not
1750 NULL CRL path validation is taking place. Existing applications wont
1751 see this because it requires extended CRL support which is off by
1754 This work was sponsored by Google.
1757 *) Support for freshest CRL extension.
1759 This work was sponsored by Google.
1762 *) Initial indirect CRL support. Currently only supported in the CRLs
1763 passed directly and not via lookup. Process certificate issuer
1764 CRL entry extension and lookup CRL entries by bother issuer name
1765 and serial number. Check and process CRL issuer entry in IDP extension.
1767 This work was sponsored by Google.
1770 *) Add support for distinct certificate and CRL paths. The CRL issuer
1771 certificate is validated separately in this case. Only enabled if
1772 an extended CRL support flag is set: this flag will enable additional
1773 CRL functionality in future.
1775 This work was sponsored by Google.
1778 *) Add support for policy mappings extension.
1780 This work was sponsored by Google.
1783 *) Fixes to pathlength constraint, self issued certificate handling,
1784 policy processing to align with RFC3280 and PKITS tests.
1786 This work was sponsored by Google.
1789 *) Support for name constraints certificate extension. DN, email, DNS
1790 and URI types are currently supported.
1792 This work was sponsored by Google.
1795 *) To cater for systems that provide a pointer-based thread ID rather
1796 than numeric, deprecate the current numeric thread ID mechanism and
1797 replace it with a structure and associated callback type. This
1798 mechanism allows a numeric "hash" to be extracted from a thread ID in
1799 either case, and on platforms where pointers are larger than 'long',
1800 mixing is done to help ensure the numeric 'hash' is usable even if it
1801 can't be guaranteed unique. The default mechanism is to use "&errno"
1802 as a pointer-based thread ID to distinguish between threads.
1804 Applications that want to provide their own thread IDs should now use
1805 CRYPTO_THREADID_set_callback() to register a callback that will call
1806 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
1808 Note that ERR_remove_state() is now deprecated, because it is tied
1809 to the assumption that thread IDs are numeric. ERR_remove_state(0)
1810 to free the current thread's error state should be replaced by
1811 ERR_remove_thread_state(NULL).
1813 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
1814 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
1815 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
1816 application was previously providing a numeric thread callback that
1817 was inappropriate for distinguishing threads, then uniqueness might
1818 have been obtained with &errno that happened immediately in the
1819 intermediate development versions of OpenSSL; this is no longer the
1820 case, the numeric thread callback will now override the automatic use
1822 [Geoff Thorpe, with help from Bodo Moeller]
1824 *) Initial support for different CRL issuing certificates. This covers a
1825 simple case where the self issued certificates in the chain exist and
1826 the real CRL issuer is higher in the existing chain.
1828 This work was sponsored by Google.
1831 *) Removed effectively defunct crypto/store from the build.
1834 *) Revamp of STACK to provide stronger type-checking. Still to come:
1835 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
1836 ASN1_STRING, CONF_VALUE.
1839 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
1840 RAM on SSL connections. This option can save about 34k per idle SSL.
1843 *) Revamp of LHASH to provide stronger type-checking. Still to come:
1844 STACK, TXT_DB, bsearch, qsort.
1847 *) Initial support for Cryptographic Message Syntax (aka CMS) based
1848 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
1849 support for data, signedData, compressedData, digestedData and
1850 encryptedData, envelopedData types included. Scripts to check against
1851 RFC4134 examples draft and interop and consistency checks of many
1852 content types and variants.
1855 *) Add options to enc utility to support use of zlib compression BIO.
1858 *) Extend mk1mf to support importing of options and assembly language
1859 files from Configure script, currently only included in VC-WIN32.
1860 The assembly language rules can now optionally generate the source
1861 files from the associated perl scripts.
1864 *) Implement remaining functionality needed to support GOST ciphersuites.
1865 Interop testing has been performed using CryptoPro implementations.
1866 [Victor B. Wagner <vitus@cryptocom.ru>]
1868 *) s390x assembler pack.
1871 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
1875 *) Implement Opaque PRF Input TLS extension as specified in
1876 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
1877 official specification yet and no extension type assignment by
1878 IANA exists, this extension (for now) will have to be explicitly
1879 enabled when building OpenSSL by providing the extension number
1880 to use. For example, specify an option
1882 -DTLSEXT_TYPE_opaque_prf_input=0x9527
1884 to the "config" or "Configure" script to enable the extension,
1885 assuming extension number 0x9527 (which is a completely arbitrary
1886 and unofficial assignment based on the MD5 hash of the Internet
1887 Draft). Note that by doing so, you potentially lose
1888 interoperability with other TLS implementations since these might
1889 be using the same extension number for other purposes.
1891 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
1892 opaque PRF input value to use in the handshake. This will create
1893 an interal copy of the length-'len' string at 'src', and will
1894 return non-zero for success.
1896 To get more control and flexibility, provide a callback function
1899 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
1900 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
1904 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
1907 Callback function 'cb' will be called in handshakes, and is
1908 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
1909 Argument 'arg' is for application purposes (the value as given to
1910 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
1911 be provided to the callback function). The callback function
1912 has to return non-zero to report success: usually 1 to use opaque
1913 PRF input just if possible, or 2 to enforce use of the opaque PRF
1914 input. In the latter case, the library will abort the handshake
1915 if opaque PRF input is not successfully negotiated.
1917 Arguments 'peerinput' and 'len' given to the callback function
1918 will always be NULL and 0 in the case of a client. A server will
1919 see the client's opaque PRF input through these variables if
1920 available (NULL and 0 otherwise). Note that if the server
1921 provides an opaque PRF input, the length must be the same as the
1922 length of the client's opaque PRF input.
1924 Note that the callback function will only be called when creating
1925 a new session (session resumption can resume whatever was
1926 previously negotiated), and will not be called in SSL 2.0
1927 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
1928 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
1929 for applications that need to enforce opaque PRF input.
1933 *) Update ssl code to support digests other than SHA1+MD5 for handshake
1936 [Victor B. Wagner <vitus@cryptocom.ru>]
1938 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1939 RFC4507bis. The encrypted ticket format is an encrypted encoded
1940 SSL_SESSION structure, that way new session features are automatically
1943 If a client application caches session in an SSL_SESSION structure
1944 support is transparent because tickets are now stored in the encoded
1947 The SSL_CTX structure automatically generates keys for ticket
1948 protection in servers so again support should be possible
1949 with no application modification.
1951 If a client or server wishes to disable RFC4507 support then the option
1952 SSL_OP_NO_TICKET can be set.
1954 Add a TLS extension debugging callback to allow the contents of any client
1955 or server extensions to be examined.
1957 This work was sponsored by Google.
1960 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
1961 OpenSSL should now compile cleanly on gcc 4.2
1962 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
1964 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
1965 support including streaming MAC support: this is required for GOST
1966 ciphersuite support.
1967 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
1969 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1970 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1971 to output in BER and PEM format.
1974 *) Experimental support for use of HMAC via EVP_PKEY interface. This
1975 allows HMAC to be handled via the EVP_DigestSign*() interface. The
1976 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
1977 ENGINE support for HMAC keys which are unextractable. New -mac and
1978 -macopt options to dgst utility.
1981 *) New option -sigopt to dgst utility. Update dgst to use
1982 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
1983 alternative signing paramaters such as X9.31 or PSS in the dgst
1987 *) Change ssl_cipher_apply_rule(), the internal function that does
1988 the work each time a ciphersuite string requests enabling
1989 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
1990 removing ("!foo+bar") a class of ciphersuites: Now it maintains
1991 the order of disabled ciphersuites such that those ciphersuites
1992 that most recently went from enabled to disabled not only stay
1993 in order with respect to each other, but also have higher priority
1994 than other disabled ciphersuites the next time ciphersuites are
1997 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
1998 the same ciphersuites as with "HIGH" alone, but in a specific
1999 order where the PSK ciphersuites come first (since they are the
2000 most recently disabled ciphersuites when "HIGH" is parsed).
2002 Also, change ssl_create_cipher_list() (using this new
2003 funcionality) such that between otherwise identical
2004 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
2008 *) Change ssl_create_cipher_list() so that it automatically
2009 arranges the ciphersuites in reasonable order before starting
2010 to process the rule string. Thus, the definition for "DEFAULT"
2011 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
2012 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
2013 This makes it much easier to arrive at a reasonable default order
2014 in applications for which anonymous ciphers are OK (meaning
2015 that you can't actually use DEFAULT).
2016 [Bodo Moeller; suggested by Victor Duchovni]
2018 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
2019 processing) into multiple integers instead of setting
2020 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
2021 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
2022 (These masks as well as the individual bit definitions are hidden
2023 away into the non-exported interface ssl/ssl_locl.h, so this
2024 change to the definition of the SSL_CIPHER structure shouldn't
2025 affect applications.) This give us more bits for each of these
2026 categories, so there is no longer a need to coagulate AES128 and
2027 AES256 into a single algorithm bit, and to coagulate Camellia128
2028 and Camellia256 into a single algorithm bit, which has led to all
2031 Thus, among other things, the kludge introduced in 0.9.7m and
2032 0.9.8e for masking out AES256 independently of AES128 or masking
2033 out Camellia256 independently of AES256 is not needed here in 0.9.9.
2035 With the change, we also introduce new ciphersuite aliases that
2036 so far were missing: "AES128", "AES256", "CAMELLIA128", and
2040 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
2041 Use the leftmost N bytes of the signature input if the input is
2042 larger than the prime q (with N being the size in bytes of q).
2045 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
2046 it yet and it is largely untested.
2049 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
2052 *) Initial incomplete changes to avoid need for function casts in OpenSSL
2053 some compilers (gcc 4.2 and later) reject their use. Safestack is
2054 reimplemented. Update ASN1 to avoid use of legacy functions.
2057 *) Win32/64 targets are linked with Winsock2.
2060 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
2061 to external functions. This can be used to increase CRL handling
2062 efficiency especially when CRLs are very large by (for example) storing
2063 the CRL revoked certificates in a database.
2066 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
2067 new CRLs added to a directory can be used. New command line option
2068 -verify_return_error to s_client and s_server. This causes real errors
2069 to be returned by the verify callback instead of carrying on no matter
2070 what. This reflects the way a "real world" verify callback would behave.
2073 *) GOST engine, supporting several GOST algorithms and public key formats.
2074 Kindly donated by Cryptocom.
2077 *) Partial support for Issuing Distribution Point CRL extension. CRLs
2078 partitioned by DP are handled but no indirect CRL or reason partitioning
2079 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
2080 selected via a scoring technique which handles IDP and AKID in CRLs.
2083 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
2084 will ultimately be used for all verify operations: this will remove the
2085 X509_STORE dependency on certificate verification and allow alternative
2086 lookup methods. X509_STORE based implementations of these two callbacks.
2089 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
2090 Modify get_crl() to find a valid (unexpired) CRL if possible.
2093 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
2094 this would be called X509_CRL_cmp() but that name is already used by
2095 a function that just compares CRL issuer names. Cache several CRL
2096 extensions in X509_CRL structure and cache CRLDP in X509.
2099 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
2100 this maps equivalent X509_NAME structures into a consistent structure.
2101 Name comparison can then be performed rapidly using memcmp().
2104 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
2108 *) Allow digests to supply their own micalg string for S/MIME type using
2109 the ctrl EVP_MD_CTRL_MICALG.
2112 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
2113 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
2114 ctrl. It can then customise the structure before and/or after signing
2118 *) New function OBJ_add_sigid() to allow application defined signature OIDs
2119 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
2120 to free up any added signature OIDs.
2123 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
2124 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
2125 digest and cipher tables. New options added to openssl utility:
2126 list-message-digest-algorithms and list-cipher-algorithms.
2129 *) Change the array representation of binary polynomials: the list
2130 of degrees of non-zero coefficients is now terminated with -1.
2131 Previously it was terminated with 0, which was also part of the
2132 value; thus, the array representation was not applicable to
2133 polynomials where t^0 has coefficient zero. This change makes
2134 the array representation useful in a more general context.
2137 *) Various modifications and fixes to SSL/TLS cipher string
2138 handling. For ECC, the code now distinguishes between fixed ECDH
2139 with RSA certificates on the one hand and with ECDSA certificates
2140 on the other hand, since these are separate ciphersuites. The
2141 unused code for Fortezza ciphersuites has been removed.
2143 For consistency with EDH, ephemeral ECDH is now called "EECDH"
2144 (not "ECDHE"). For consistency with the code for DH
2145 certificates, use of ECDH certificates is now considered ECDH
2146 authentication, not RSA or ECDSA authentication (the latter is
2147 merely the CA's signing algorithm and not actively used in the
2150 The temporary ciphersuite alias "ECCdraft" is no longer
2151 available, and ECC ciphersuites are no longer excluded from "ALL"
2152 and "DEFAULT". The following aliases now exist for RFC 4492
2153 ciphersuites, most of these by analogy with the DH case:
2155 kECDHr - ECDH cert, signed with RSA
2156 kECDHe - ECDH cert, signed with ECDSA
2157 kECDH - ECDH cert (signed with either RSA or ECDSA)
2158 kEECDH - ephemeral ECDH
2159 ECDH - ECDH cert or ephemeral ECDH
2165 AECDH - anonymous ECDH
2166 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
2170 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
2171 Use correct micalg parameters depending on digest(s) in signed message.
2174 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
2175 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
2178 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
2179 an engine to register a method. Add ENGINE lookups for methods and
2180 functional reference processing.
2183 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
2184 EVP_{Sign,Verify}* which allow an application to customise the signature
2188 *) New -resign option to smime utility. This adds one or more signers
2189 to an existing PKCS#7 signedData structure. Also -md option to use an
2190 alternative message digest algorithm for signing.
2193 *) Tidy up PKCS#7 routines and add new functions to make it easier to
2194 create PKCS7 structures containing multiple signers. Update smime
2195 application to support multiple signers.
2198 *) New -macalg option to pkcs12 utility to allow setting of an alternative
2202 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
2203 Reorganize PBE internals to lookup from a static table using NIDs,
2204 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
2205 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
2206 PRF which will be automatically used with PBES2.
2209 *) Replace the algorithm specific calls to generate keys in "req" with the
2213 *) Update PKCS#7 enveloped data routines to use new API. This is now
2214 supported by any public key method supporting the encrypt operation. A
2215 ctrl is added to allow the public key algorithm to examine or modify
2216 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
2220 *) Add a ctrl to asn1 method to allow a public key algorithm to express
2221 a default digest type to use. In most cases this will be SHA1 but some
2222 algorithms (such as GOST) need to specify an alternative digest. The
2223 return value indicates how strong the prefernce is 1 means optional and
2224 2 is mandatory (that is it is the only supported type). Modify
2225 ASN1_item_sign() to accept a NULL digest argument to indicate it should
2226 use the default md. Update openssl utilities to use the default digest
2227 type for signing if it is not explicitly indicated.
2230 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
2231 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
2232 signing method from the key type. This effectively removes the link
2233 between digests and public key types.
2236 *) Add an OID cross reference table and utility functions. Its purpose is to
2237 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
2238 rsaEncryption. This will allow some of the algorithm specific hackery
2239 needed to use the correct OID to be removed.
2242 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
2243 structures for PKCS7_sign(). They are now set up by the relevant public
2247 *) Add provisional EC pkey method with support for ECDSA and ECDH.
2250 *) Add support for key derivation (agreement) in the API, DH method and
2254 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
2255 public and private key formats. As a side effect these add additional
2256 command line functionality not previously available: DSA signatures can be
2257 generated and verified using pkeyutl and DH key support and generation in
2262 [Oliver Tappe <zooey@hirschkaefer.de>]
2264 *) New make target "install_html_docs" installs HTML renditions of the
2266 [Oliver Tappe <zooey@hirschkaefer.de>]
2268 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
2269 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
2270 support key and parameter generation and add initial key generation
2271 functionality for RSA.
2274 *) Add functions for main EVP_PKEY_method operations. The undocumented
2275 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
2276 EVP_PKEY_{encrypt,decrypt}_old.
2279 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
2280 key API, doesn't do much yet.
2283 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
2284 public key algorithms. New option to openssl utility:
2285 "list-public-key-algorithms" to print out info.
2288 *) Implement the Supported Elliptic Curves Extension for
2289 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2292 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
2293 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
2296 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
2297 utilities such as rsa, dsa, dsaparam etc except they process any key
2301 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
2302 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
2303 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
2307 *) Initial support for pluggable public key ASN1.
2308 De-spaghettify the public key ASN1 handling. Move public and private
2309 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
2310 algorithm specific handling to a single module within the relevant
2311 algorithm directory. Add functions to allow (near) opaque processing
2312 of public and private key structures.
2315 *) Implement the Supported Point Formats Extension for
2316 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2319 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
2320 for the psk identity [hint] and the psk callback functions to the
2321 SSL_SESSION, SSL and SSL_CTX structure.
2324 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
2328 SSL_CTX_use_psk_identity_hint
2329 SSL_get_psk_identity_hint
2330 SSL_get_psk_identity
2331 SSL_use_psk_identity_hint
2333 [Mika Kousa and Pasi Eronen of Nokia Corporation]
2335 *) Add RFC 3161 compliant time stamp request creation, response generation
2336 and response verification functionality.
2337 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
2339 *) Add initial support for TLS extensions, specifically for the server_name
2340 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2341 have new members for a host name. The SSL data structure has an
2342 additional member SSL_CTX *initial_ctx so that new sessions can be
2343 stored in that context to allow for session resumption, even after the
2344 SSL has been switched to a new SSL_CTX in reaction to a client's
2345 server_name extension.
2347 New functions (subject to change):
2349 SSL_get_servername()
2350 SSL_get_servername_type()
2353 New CTRL codes and macros (subject to change):
2355 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2356 - SSL_CTX_set_tlsext_servername_callback()
2357 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2358 - SSL_CTX_set_tlsext_servername_arg()
2359 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2361 openssl s_client has a new '-servername ...' option.
2363 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2364 '-key2 ...', '-servername_fatal' (subject to change). This allows
2365 testing the HostName extension for a specific single host name ('-cert'
2366 and '-key' remain fallbacks for handshakes without HostName
2367 negotiation). If the unrecogninzed_name alert has to be sent, this by
2368 default is a warning; it becomes fatal with the '-servername_fatal'
2371 [Peter Sylvester, Remy Allais, Christophe Renou]
2373 *) Whirlpool hash implementation is added.
2376 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
2377 bn(64,32). Because of instruction set limitations it doesn't have
2378 any negative impact on performance. This was done mostly in order
2379 to make it possible to share assembler modules, such as bn_mul_mont
2380 implementations, between 32- and 64-bit builds without hassle.
2383 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
2384 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
2388 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
2389 dedicated Montgomery multiplication procedure, is introduced.
2390 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
2391 "64-bit" performance on certain 32-bit targets.
2394 *) New option SSL_OP_NO_COMP to disable use of compression selectively
2395 in SSL structures. New SSL ctrl to set maximum send fragment size.
2396 Save memory by seeting the I/O buffer sizes dynamically instead of
2397 using the maximum available value.
2400 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
2401 in addition to the text details.
2404 *) Very, very preliminary EXPERIMENTAL support for printing of general
2405 ASN1 structures. This currently produces rather ugly output and doesn't
2406 handle several customised structures at all.
2409 *) Integrated support for PVK file format and some related formats such
2410 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
2411 these in the 'rsa' and 'dsa' utilities.
2414 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
2417 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
2418 place for the (very old) "NETSCAPE" format certificates which are now
2419 handled using new ASN1 code equivalents.
2422 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
2423 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
2424 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
2427 *) Modify CRL distribution points extension code to print out previously
2428 unsupported fields. Enhance extension setting code to allow setting of
2432 *) Add print and set support for Issuing Distribution Point CRL extension.
2435 *) Change 'Configure' script to enable Camellia by default.
2438 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
2440 *) When rejecting SSL/TLS records due to an incorrect version number, never
2441 update s->server with a new major version number. As of
2442 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
2443 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
2444 the previous behavior could result in a read attempt at NULL when
2445 receiving specific incorrect SSL/TLS records once record payload
2446 protection is active. (CVE-2010-0740)
2447 [Bodo Moeller, Adam Langley <agl@chromium.org>]
2449 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
2450 could be crashed if the relevant tables were not present (e.g. chrooted).
2451 [Tomas Hoger <thoger@redhat.com>]
2453 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
2455 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
2456 [Martin Olsson, Neel Mehta]
2458 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
2459 accommodate for stack sorting, always a write lock!).
2462 *) On some versions of WIN32 Heap32Next is very slow. This can cause
2463 excessive delays in the RAND_poll(): over a minute. As a workaround
2464 include a time check in the inner Heap32Next loop too.
2467 *) The code that handled flushing of data in SSL/TLS originally used the
2468 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
2469 the problem outlined in PR#1949. The fix suggested there however can
2470 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
2471 of Apache). So instead simplify the code to flush unconditionally.
2472 This should be fine since flushing with no data to flush is a no op.
2475 *) Handle TLS versions 2.0 and later properly and correctly use the
2476 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
2477 off ancient servers have a habit of sticking around for a while...
2480 *) Modify compression code so it frees up structures without using the
2481 ex_data callbacks. This works around a problem where some applications
2482 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
2483 restarting) then use compression (e.g. SSL with compression) later.
2484 This results in significant per-connection memory leaks and
2485 has caused some security issues including CVE-2008-1678 and
2489 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
2490 change when encrypting or decrypting.
2493 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
2494 connect and renegotiate with servers which do not support RI.
2495 Until RI is more widely deployed this option is enabled by default.
2498 *) Add "missing" ssl ctrls to clear options and mode.
2501 *) If client attempts to renegotiate and doesn't support RI respond with
2502 a no_renegotiation alert as required by RFC5746. Some renegotiating
2503 TLS clients will continue a connection gracefully when they receive
2504 the alert. Unfortunately OpenSSL mishandled this alert and would hang
2505 waiting for a server hello which it will never receive. Now we treat a
2506 received no_renegotiation alert as a fatal error. This is because
2507 applications requesting a renegotiation might well expect it to succeed
2508 and would have no code in place to handle the server denying it so the
2509 only safe thing to do is to terminate the connection.
2512 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
2513 peer supports secure renegotiation and 0 otherwise. Print out peer
2514 renegotiation support in s_client/s_server.
2517 *) Replace the highly broken and deprecated SPKAC certification method with
2518 the updated NID creation version. This should correctly handle UTF8.
2521 *) Implement RFC5746. Re-enable renegotiation but require the extension
2522 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
2523 turns out to be a bad idea. It has been replaced by
2524 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
2525 SSL_CTX_set_options(). This is really not recommended unless you
2526 know what you are doing.
2527 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
2529 *) Fixes to stateless session resumption handling. Use initial_ctx when
2530 issuing and attempting to decrypt tickets in case it has changed during
2531 servername handling. Use a non-zero length session ID when attempting
2532 stateless session resumption: this makes it possible to determine if
2533 a resumption has occurred immediately after receiving server hello
2534 (several places in OpenSSL subtly assume this) instead of later in
2538 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
2539 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
2540 fixes for a few places where the return code is not checked
2542 [Julia Lawall <julia@diku.dk>]
2544 *) Add --strict-warnings option to Configure script to include devteam
2545 warnings in other configurations.
2548 *) Add support for --libdir option and LIBDIR variable in makefiles. This
2549 makes it possible to install openssl libraries in locations which
2550 have names other than "lib", for example "/usr/lib64" which some
2552 [Steve Henson, based on patch from Jeremy Utley]
2554 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
2555 X690 8.9.12 and can produce some misleading textual output of OIDs.
2556 [Steve Henson, reported by Dan Kaminsky]
2558 *) Delete MD2 from algorithm tables. This follows the recommendation in
2559 several standards that it is not used in new applications due to
2560 several cryptographic weaknesses. For binary compatibility reasons
2561 the MD2 API is still compiled in by default.
2564 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
2568 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
2569 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
2571 [Guenter <lists@gknw.net>]
2573 *) Fix the server certificate chain building code to use X509_verify_cert(),
2574 it used to have an ad-hoc builder which was unable to cope with anything
2575 other than a simple chain.
2576 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
2578 *) Don't check self signed certificate signatures in X509_verify_cert()
2579 by default (a flag can override this): it just wastes time without
2580 adding any security. As a useful side effect self signed root CAs
2581 with non-FIPS digests are now usable in FIPS mode.
2584 *) In dtls1_process_out_of_seq_message() the check if the current message
2585 is already buffered was missing. For every new message was memory
2586 allocated, allowing an attacker to perform an denial of service attack
2587 with sending out of seq handshake messages until there is no memory
2588 left. Additionally every future messege was buffered, even if the
2589 sequence number made no sense and would be part of another handshake.
2590 So only messages with sequence numbers less than 10 in advance will be
2591 buffered. (CVE-2009-1378)
2592 [Robin Seggelmann, discovered by Daniel Mentz]
2594 *) Records are buffered if they arrive with a future epoch to be
2595 processed after finishing the corresponding handshake. There is
2596 currently no limitation to this buffer allowing an attacker to perform
2597 a DOS attack with sending records with future epochs until there is no
2598 memory left. This patch adds the pqueue_size() function to detemine
2599 the size of a buffer and limits the record buffer to 100 entries.
2601 [Robin Seggelmann, discovered by Daniel Mentz]
2603 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
2604 parent structure is freed. (CVE-2009-1379)
2607 *) Handle non-blocking I/O properly in SSL_shutdown() call.
2608 [Darryl Miles <darryl-mailinglists@netbauds.net>]
2611 [Ilya O. <vrghost@gmail.com>]
2613 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
2615 *) Disable renegotiation completely - this fixes a severe security
2616 problem (CVE-2009-3555) at the cost of breaking all
2617 renegotiation. Renegotiation can be re-enabled by setting
2618 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
2619 run-time. This is really not recommended unless you know what
2623 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
2625 *) Don't set val to NULL when freeing up structures, it is freed up by
2626 underlying code. If sizeof(void *) > sizeof(long) this can result in
2627 zeroing past the valid field. (CVE-2009-0789)
2628 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
2630 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
2631 checked correctly. This would allow some invalid signed attributes to
2632 appear to verify correctly. (CVE-2009-0591)
2633 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2635 *) Reject UniversalString and BMPString types with invalid lengths. This
2636 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
2637 a legal length. (CVE-2009-0590)
2640 *) Set S/MIME signing as the default purpose rather than setting it
2641 unconditionally. This allows applications to override it at the store
2645 *) Permit restricted recursion of ASN1 strings. This is needed in practice
2646 to handle some structures.
2649 *) Improve efficiency of mem_gets: don't search whole buffer each time
2651 [Jeremy Shapiro <jnshapir@us.ibm.com>]
2653 *) New -hex option for openssl rand.
2656 *) Print out UTF8String and NumericString when parsing ASN1.
2659 *) Support NumericString type for name components.
2662 *) Allow CC in the environment to override the automatically chosen
2663 compiler. Note that nothing is done to ensure flags work with the
2667 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2669 *) Properly check EVP_VerifyFinal() and similar return values
2671 [Ben Laurie, Bodo Moeller, Google Security Team]
2673 *) Enable TLS extensions by default.
2676 *) Allow the CHIL engine to be loaded, whether the application is
2677 multithreaded or not. (This does not release the developer from the
2678 obligation to set up the dynamic locking callbacks.)
2679 [Sander Temme <sander@temme.net>]
2681 *) Use correct exit code if there is an error in dgst command.
2682 [Steve Henson; problem pointed out by Roland Dirlewanger]
2684 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2685 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2688 *) Add experimental JPAKE support, including demo authentication in
2689 s_client and s_server.
2692 *) Set the comparison function in v3_addr_canonize().
2693 [Rob Austein <sra@hactrn.net>]
2695 *) Add support for XMPP STARTTLS in s_client.
2696 [Philip Paeps <philip@freebsd.org>]
2698 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
2699 to ensure that even with this option, only ciphersuites in the
2700 server's preference list will be accepted. (Note that the option
2701 applies only when resuming a session, so the earlier behavior was
2702 just about the algorithm choice for symmetric cryptography.)
2705 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
2707 *) Fix NULL pointer dereference if a DTLS server received
2708 ChangeCipherSpec as first record (CVE-2009-1386).
2711 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
2712 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2715 *) The fix in 0.9.8c that supposedly got rid of unsafe
2716 double-checked locking was incomplete for RSA blinding,
2717 addressing just one layer of what turns out to have been
2718 doubly unsafe triple-checked locking.
2720 So now fix this for real by retiring the MONT_HELPER macro
2721 in crypto/rsa/rsa_eay.c.
2723 [Bodo Moeller; problem pointed out by Marius Schilder]
2725 *) Various precautionary measures:
2727 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
2729 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
2730 (NB: This would require knowledge of the secret session ticket key
2731 to exploit, in which case you'd be SOL either way.)
2733 - Change bn_nist.c so that it will properly handle input BIGNUMs
2734 outside the expected range.
2736 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2739 [Neel Mehta, Bodo Moeller]
2741 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2742 the load fails. Useful for distros.
2743 [Ben Laurie and the FreeBSD team]
2745 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2748 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2751 *) Expand ENGINE to support engine supplied SSL client certificate functions.
2753 This work was sponsored by Logica.
2756 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
2757 keystores. Support for SSL/TLS client authentication too.
2758 Not compiled unless enable-capieng specified to Configure.
2760 This work was sponsored by Logica.
2763 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
2764 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
2765 attribute creation routines such as certifcate requests and PKCS#12
2769 Changes between 0.9.8g and 0.9.8h [28 May 2008]
2771 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
2772 handshake which could lead to a cilent crash as found using the
2773 Codenomicon TLS test suite (CVE-2008-1672)
2774 [Steve Henson, Mark Cox]
2776 *) Fix double free in TLS server name extensions which could lead to
2777 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
2780 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
2782 Clear the error queue to ensure that error entries left from
2783 older function calls do not interfere with the correct operation.
2784 [Lutz Jaenicke, Erik de Castro Lopo]
2786 *) Remove root CA certificates of commercial CAs:
2788 The OpenSSL project does not recommend any specific CA and does not
2789 have any policy with respect to including or excluding any CA.
2790 Therefore it does not make any sense to ship an arbitrary selection
2791 of root CA certificates with the OpenSSL software.
2794 *) RSA OAEP patches to fix two separate invalid memory reads.
2795 The first one involves inputs when 'lzero' is greater than
2796 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
2797 before the beginning of from). The second one involves inputs where
2798 the 'db' section contains nothing but zeroes (there is a one-byte
2799 invalid read after the end of 'db').
2800 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2802 *) Partial backport from 0.9.9-dev:
2804 Introduce bn_mul_mont (dedicated Montgomery multiplication
2805 procedure) as a candidate for BIGNUM assembler implementation.
2806 While 0.9.9-dev uses assembler for various architectures, only
2807 x86_64 is available by default here in the 0.9.8 branch, and
2808 32-bit x86 is available through a compile-time setting.
2810 To try the 32-bit x86 assembler implementation, use Configure
2811 option "enable-montasm" (which exists only for this backport).
2813 As "enable-montasm" for 32-bit x86 disclaims code stability
2814 anyway, in this constellation we activate additional code
2815 backported from 0.9.9-dev for further performance improvements,
2816 namely BN_from_montgomery_word. (To enable this otherwise,
2817 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
2819 [Andy Polyakov (backport partially by Bodo Moeller)]
2821 *) Add TLS session ticket callback. This allows an application to set
2822 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
2823 values. This is useful for key rollover for example where several key
2824 sets may exist with different names.
2827 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
2828 This was broken until now in 0.9.8 releases, such that the only way
2829 a registered ENGINE could be used (assuming it initialises
2830 successfully on the host) was to explicitly set it as the default
2831 for the relevant algorithms. This is in contradiction with 0.9.7
2832 behaviour and the documentation. With this fix, when an ENGINE is
2833 registered into a given algorithm's table of implementations, the
2834 'uptodate' flag is reset so that auto-discovery will be used next
2835 time a new context for that algorithm attempts to select an
2837 [Ian Lister (tweaked by Geoff Thorpe)]
2839 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
2840 implemention in the following ways:
2842 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
2845 Lack of BER streaming support means one pass streaming processing is
2846 only supported if data is detached: setting the streaming flag is
2847 ignored for embedded content.
2849 CMS support is disabled by default and must be explicitly enabled
2850 with the enable-cms configuration option.
2853 *) Update the GMP engine glue to do direct copies between BIGNUM and
2854 mpz_t when openssl and GMP use the same limb size. Otherwise the
2855 existing "conversion via a text string export" trick is still used.
2856 [Paul Sheer <paulsheer@gmail.com>]
2858 *) Zlib compression BIO. This is a filter BIO which compressed and
2859 uncompresses any data passed through it.
2862 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
2863 RFC3394 compatible AES key wrapping.
2866 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
2867 sets string data without copying. X509_ALGOR_set0() and
2868 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
2869 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
2870 from an X509_ATTRIBUTE structure optionally checking it occurs only
2871 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
2875 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
2876 to get the expected BN_FLG_CONSTTIME behavior.
2877 [Bodo Moeller (Google)]
2881 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
2882 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
2883 - added some more tests to do_tests.pl
2884 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
2885 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
2886 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
2887 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
2888 - various changes to netware.pl to enable gcc-cross builds on Win32
2890 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
2891 - various changes to fix missing prototype warnings
2892 - fixed x86nasm.pl to create correct asm files for NASM COFF output
2893 - added AES, WHIRLPOOL and CPUID assembler code to build files
2894 - added missing AES assembler make rules to mk1mf.pl
2895 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
2896 [Guenter Knauf <eflash@gmx.net>]
2898 *) Implement certificate status request TLS extension defined in RFC3546.
2899 A client can set the appropriate parameters and receive the encoded
2900 OCSP response via a callback. A server can query the supplied parameters
2901 and set the encoded OCSP response in the callback. Add simplified examples
2902 to s_client and s_server.
2905 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
2907 *) Fix various bugs:
2908 + Binary incompatibility of ssl_ctx_st structure
2909 + DTLS interoperation with non-compliant servers
2910 + Don't call get_session_cb() without proposed session
2911 + Fix ia64 assembler code
2912 [Andy Polyakov, Steve Henson]
2914 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
2916 *) DTLS Handshake overhaul. There were longstanding issues with
2917 OpenSSL DTLS implementation, which were making it impossible for
2918 RFC 4347 compliant client to communicate with OpenSSL server.
2919 Unfortunately just fixing these incompatibilities would "cut off"
2920 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
2921 server keeps tolerating non RFC compliant syntax. The opposite is
2922 not true, 0.9.8f client can not communicate with earlier server.
2923 This update even addresses CVE-2007-4995.
2926 *) Changes to avoid need for function casts in OpenSSL: some compilers
2927 (gcc 4.2 and later) reject their use.
2928 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
2931 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2932 RFC4507bis. The encrypted ticket format is an encrypted encoded
2933 SSL_SESSION structure, that way new session features are automatically
2936 If a client application caches session in an SSL_SESSION structure
2937 support is transparent because tickets are now stored in the encoded
2940 The SSL_CTX structure automatically generates keys for ticket
2941 protection in servers so again support should be possible
2942 with no application modification.
2944 If a client or server wishes to disable RFC4507 support then the option
2945 SSL_OP_NO_TICKET can be set.
2947 Add a TLS extension debugging callback to allow the contents of any client
2948 or server extensions to be examined.
2950 This work was sponsored by Google.
2953 *) Add initial support for TLS extensions, specifically for the server_name
2954 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2955 have new members for a host name. The SSL data structure has an
2956 additional member SSL_CTX *initial_ctx so that new sessions can be
2957 stored in that context to allow for session resumption, even after the
2958 SSL has been switched to a new SSL_CTX in reaction to a client's
2959 server_name extension.
2961 New functions (subject to change):
2963 SSL_get_servername()
2964 SSL_get_servername_type()
2967 New CTRL codes and macros (subject to change):
2969 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2970 - SSL_CTX_set_tlsext_servername_callback()
2971 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2972 - SSL_CTX_set_tlsext_servername_arg()
2973 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2975 openssl s_client has a new '-servername ...' option.
2977 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2978 '-key2 ...', '-servername_fatal' (subject to change). This allows
2979 testing the HostName extension for a specific single host name ('-cert'
2980 and '-key' remain fallbacks for handshakes without HostName
2981 negotiation). If the unrecogninzed_name alert has to be sent, this by
2982 default is a warning; it becomes fatal with the '-servername_fatal'
2985 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2987 *) Add AES and SSE2 assembly language support to VC++ build.
2990 *) Mitigate attack on final subtraction in Montgomery reduction.
2993 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2994 (which previously caused an internal error).
2997 *) Squeeze another 10% out of IGE mode when in != out.
3000 *) AES IGE mode speedup.
3001 [Dean Gaudet (Google)]
3003 *) Add the Korean symmetric 128-bit cipher SEED (see
3004 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
3005 add SEED ciphersuites from RFC 4162:
3007 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
3008 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
3009 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
3010 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
3012 To minimize changes between patchlevels in the OpenSSL 0.9.8
3013 series, SEED remains excluded from compilation unless OpenSSL
3014 is configured with 'enable-seed'.
3015 [KISA, Bodo Moeller]
3017 *) Mitigate branch prediction attacks, which can be practical if a
3018 single processor is shared, allowing a spy process to extract
3019 information. For detailed background information, see
3020 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
3021 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
3022 and Necessary Software Countermeasures"). The core of the change
3023 are new versions BN_div_no_branch() and
3024 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
3025 respectively, which are slower, but avoid the security-relevant
3026 conditional branches. These are automatically called by BN_div()
3027 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
3028 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
3029 remove a conditional branch.
3031 BN_FLG_CONSTTIME is the new name for the previous
3032 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
3033 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
3034 in the exponent causes BN_mod_exp_mont() to use the alternative
3035 implementation in BN_mod_exp_mont_consttime().) The old name
3036 remains as a deprecated alias.
3038 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
3039 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
3040 constant-time implementations for more than just exponentiation.
3041 Here too the old name is kept as a deprecated alias.
3043 BN_BLINDING_new() will now use BN_dup() for the modulus so that
3044 the BN_BLINDING structure gets an independent copy of the
3045 modulus. This means that the previous "BIGNUM *m" argument to
3046 BN_BLINDING_new() and to BN_BLINDING_create_param() now
3047 essentially becomes "const BIGNUM *m", although we can't actually
3048 change this in the header file before 0.9.9. It allows
3049 RSA_setup_blinding() to use BN_with_flags() on the modulus to
3050 enable BN_FLG_CONSTTIME.
3052 [Matthew D Wood (Intel Corp)]
3054 *) In the SSL/TLS server implementation, be strict about session ID
3055 context matching (which matters if an application uses a single
3056 external cache for different purposes). Previously,
3057 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
3058 set. This did ensure strict client verification, but meant that,
3059 with applications using a single external cache for quite
3060 different requirements, clients could circumvent ciphersuite
3061 restrictions for a given session ID context by starting a session
3062 in a different context.
3065 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3066 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3067 authentication-only ciphersuites.
3070 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
3071 not complete and could lead to a possible single byte overflow
3072 (CVE-2007-5135) [Ben Laurie]
3074 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
3076 *) Since AES128 and AES256 (and similarly Camellia128 and
3077 Camellia256) share a single mask bit in the logic of
3078 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3079 kludge to work properly if AES128 is available and AES256 isn't
3080 (or if Camellia128 is available and Camellia256 isn't).
3083 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
3084 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
3085 When a point or a seed is encoded in a BIT STRING, we need to
3086 prevent the removal of trailing zero bits to get the proper DER
3087 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
3088 of a NamedBitList, for which trailing 0 bits need to be removed.)
3091 *) Have SSL/TLS server implementation tolerate "mismatched" record
3092 protocol version while receiving ClientHello even if the
3093 ClientHello is fragmented. (The server can't insist on the
3094 particular protocol version it has chosen before the ServerHello
3095 message has informed the client about his choice.)
3098 *) Add RFC 3779 support.
3099 [Rob Austein for ARIN, Ben Laurie]
3101 *) Load error codes if they are not already present instead of using a
3102 static variable. This allows them to be cleanly unloaded and reloaded.
3103 Improve header file function name parsing.
3106 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
3107 or CAPABILITY handshake as required by RFCs.
3110 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
3112 *) Introduce limits to prevent malicious keys being able to
3113 cause a denial of service. (CVE-2006-2940)
3114 [Steve Henson, Bodo Moeller]
3116 *) Fix ASN.1 parsing of certain invalid structures that can result
3117 in a denial of service. (CVE-2006-2937) [Steve Henson]
3119 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3120 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3122 *) Fix SSL client code which could crash if connecting to a
3123 malicious SSLv2 server. (CVE-2006-4343)
3124 [Tavis Ormandy and Will Drewry, Google Security Team]
3126 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
3127 match only those. Before that, "AES256-SHA" would be interpreted
3128 as a pattern and match "AES128-SHA" too (since AES128-SHA got
3129 the same strength classification in 0.9.7h) as we currently only
3130 have a single AES bit in the ciphersuite description bitmap.
3131 That change, however, also applied to ciphersuite strings such as
3132 "RC4-MD5" that intentionally matched multiple ciphersuites --
3133 namely, SSL 2.0 ciphersuites in addition to the more common ones
3134 from SSL 3.0/TLS 1.0.
3136 So we change the selection algorithm again: Naming an explicit
3137 ciphersuite selects this one ciphersuite, and any other similar
3138 ciphersuite (same bitmap) from *other* protocol versions.
3139 Thus, "RC4-MD5" again will properly select both the SSL 2.0
3140 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
3142 Since SSL 2.0 does not have any ciphersuites for which the
3143 128/256 bit distinction would be relevant, this works for now.
3144 The proper fix will be to use different bits for AES128 and
3145 AES256, which would have avoided the problems from the beginning;
3146 however, bits are scarce, so we can only do this in a new release
3147 (not just a patchlevel) when we can change the SSL_CIPHER
3148 definition to split the single 'unsigned long mask' bitmap into
3149 multiple values to extend the available space.
3153 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
3155 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3156 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3158 *) Add AES IGE and biIGE modes.
3161 *) Change the Unix randomness entropy gathering to use poll() when
3162 possible instead of select(), since the latter has some
3163 undesirable limitations.
3164 [Darryl Miles via Richard Levitte and Bodo Moeller]
3166 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
3167 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
3168 cannot be implicitly activated as part of, e.g., the "AES" alias.
3169 However, please upgrade to OpenSSL 0.9.9[-dev] for
3170 non-experimental use of the ECC ciphersuites to get TLS extension
3171 support, which is required for curve and point format negotiation
3172 to avoid potential handshake problems.
3175 *) Disable rogue ciphersuites:
3177 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3178 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3179 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3181 The latter two were purportedly from
3182 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3185 Also deactivate the remaining ciphersuites from
3186 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3187 unofficial, and the ID has long expired.
3190 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3191 dual-core machines) and other potential thread-safety issues.
3194 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3195 versions), which is now available for royalty-free use
3196 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3197 Also, add Camellia TLS ciphersuites from RFC 4132.
3199 To minimize changes between patchlevels in the OpenSSL 0.9.8
3200 series, Camellia remains excluded from compilation unless OpenSSL
3201 is configured with 'enable-camellia'.
3204 *) Disable the padding bug check when compression is in use. The padding
3205 bug check assumes the first packet is of even length, this is not
3206 necessarily true if compresssion is enabled and can result in false
3207 positives causing handshake failure. The actual bug test is ancient
3208 code so it is hoped that implementations will either have fixed it by
3209 now or any which still have the bug do not support compression.
3212 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3214 *) When applying a cipher rule check to see if string match is an explicit
3215 cipher suite and only match that one cipher suite if it is.
3218 *) Link in manifests for VC++ if needed.
3219 [Austin Ziegler <halostatue@gmail.com>]
3221 *) Update support for ECC-based TLS ciphersuites according to
3222 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3223 TLS extensions, which are supported starting with the 0.9.9
3224 branch, not in the OpenSSL 0.9.8 branch).
3227 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3228 opaque EVP_CIPHER_CTX handling.
3231 *) Fixes and enhancements to zlib compression code. We now only use
3232 "zlib1.dll" and use the default __cdecl calling convention on Win32
3233 to conform with the standards mentioned here:
3234 http://www.zlib.net/DLL_FAQ.txt
3235 Static zlib linking now works on Windows and the new --with-zlib-include
3236 --with-zlib-lib options to Configure can be used to supply the location
3237 of the headers and library. Gracefully handle case where zlib library
3241 *) Several fixes and enhancements to the OID generation code. The old code
3242 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3243 handle numbers larger than ULONG_MAX, truncated printing and had a
3244 non standard OBJ_obj2txt() behaviour.
3247 *) Add support for building of engines under engine/ as shared libraries
3248 under VC++ build system.
3251 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3252 Hopefully, we will not see any false combination of paths any more.
3255 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3257 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3258 (part of SSL_OP_ALL). This option used to disable the
3259 countermeasure against man-in-the-middle protocol-version
3260 rollback in the SSL 2.0 server implementation, which is a bad
3261 idea. (CVE-2005-2969)
3263 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3264 for Information Security, National Institute of Advanced Industrial
3265 Science and Technology [AIST], Japan)]
3267 *) Add two function to clear and return the verify parameter flags.
3270 *) Keep cipherlists sorted in the source instead of sorting them at
3271 runtime, thus removing the need for a lock.
3274 *) Avoid some small subgroup attacks in Diffie-Hellman.
3275 [Nick Mathewson and Ben Laurie]
3277 *) Add functions for well-known primes.
3280 *) Extended Windows CE support.
3281 [Satoshi Nakamura and Andy Polyakov]
3283 *) Initialize SSL_METHOD structures at compile time instead of during
3284 runtime, thus removing the need for a lock.
3287 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3288 attempting to decrypt each encrypted key in turn. Add support to
3292 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3294 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3297 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3300 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3301 key into the same file any more.
3304 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
3307 *) Add -utf8 command line and config file option to 'ca'.
3308 [Stefan <stf@udoma.org]
3310 *) Removed the macro des_crypt(), as it seems to conflict with some
3311 libraries. Use DES_crypt().
3314 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
3315 involves renaming the source and generated shared-libs for
3316 both. The engines will accept the corrected or legacy ids
3317 ('ncipher' and '4758_cca' respectively) when binding. NB,
3318 this only applies when building 'shared'.
3319 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
3321 *) Add attribute functions to EVP_PKEY structure. Modify
3322 PKCS12_create() to recognize a CSP name attribute and
3323 use it. Make -CSP option work again in pkcs12 utility.
3326 *) Add new functionality to the bn blinding code:
3327 - automatic re-creation of the BN_BLINDING parameters after
3328 a fixed number of uses (currently 32)
3329 - add new function for parameter creation
3330 - introduce flags to control the update behaviour of the
3331 BN_BLINDING parameters
3332 - hide BN_BLINDING structure
3333 Add a second BN_BLINDING slot to the RSA structure to improve
3334 performance when a single RSA object is shared among several
3338 *) Add support for DTLS.
3339 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
3341 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
3342 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
3345 *) Remove buggy and incompletet DH cert support from
3346 ssl/ssl_rsa.c and ssl/s3_both.c
3349 *) Use SHA-1 instead of MD5 as the default digest algorithm for
3350 the apps/openssl applications.
3353 *) Compile clean with "-Wall -Wmissing-prototypes
3354 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
3355 DEBUG_SAFESTACK must also be set.
3358 *) Change ./Configure so that certain algorithms can be disabled by default.
3359 The new counterpiece to "no-xxx" is "enable-xxx".
3361 The patented RC5 and MDC2 algorithms will now be disabled unless
3362 "enable-rc5" and "enable-mdc2", respectively, are specified.
3364 (IDEA remains enabled despite being patented. This is because IDEA
3365 is frequently required for interoperability, and there is no license
3366 fee for non-commercial use. As before, "no-idea" can be used to
3367 avoid this algorithm.)
3371 *) Add processing of proxy certificates (see RFC 3820). This work was
3372 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
3373 EGEE (Enabling Grids for E-science in Europe).
3376 *) RC4 performance overhaul on modern architectures/implementations, such
3377 as Intel P4, IA-64 and AMD64.
3380 *) New utility extract-section.pl. This can be used specify an alternative
3381 section number in a pod file instead of having to treat each file as
3382 a separate case in Makefile. This can be done by adding two lines to the
3385 =for comment openssl_section:XXX
3387 The blank line is mandatory.
3391 *) New arguments -certform, -keyform and -pass for s_client and s_server
3392 to allow alternative format key and certificate files and passphrase
3396 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
3397 update associated structures and add various utility functions.
3399 Add new policy related verify parameters, include policy checking in
3400 standard verify code. Enhance 'smime' application with extra parameters
3401 to support policy checking and print out.
3404 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
3405 Nehemiah processors. These extensions support AES encryption in hardware
3406 as well as RNG (though RNG support is currently disabled).
3407 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
3409 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
3412 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
3413 [Andy Polyakov and a number of other people]
3415 *) Improved PowerPC platform support. Most notably BIGNUM assembler
3416 implementation contributed by IBM.
3417 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
3419 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
3420 exponent rather than 'unsigned long'. There is a corresponding change to
3421 the new 'rsa_keygen' element of the RSA_METHOD structure.
3422 [Jelte Jansen, Geoff Thorpe]
3424 *) Functionality for creating the initial serial number file is now
3425 moved from CA.pl to the 'ca' utility with a new option -create_serial.
3427 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
3428 number file to 1, which is bound to cause problems. To avoid
3429 the problems while respecting compatibility between different 0.9.7
3430 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
3431 CA.pl for serial number initialization. With the new release 0.9.8,
3432 we can fix the problem directly in the 'ca' utility.)
3435 *) Reduced header interdepencies by declaring more opaque objects in
3436 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
3437 give fewer recursive includes, which could break lazy source code - so
3438 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
3439 developers should define this symbol when building and using openssl to
3440 ensure they track the recommended behaviour, interfaces, [etc], but
3441 backwards-compatible behaviour prevails when this isn't defined.
3444 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
3447 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
3448 This will generate a random key of the appropriate length based on the
3449 cipher context. The EVP_CIPHER can provide its own random key generation
3450 routine to support keys of a specific form. This is used in the des and
3451 3des routines to generate a key of the correct parity. Update S/MIME
3452 code to use new functions and hence generate correct parity DES keys.
3453 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
3454 valid (weak or incorrect parity).
3457 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
3458 as looking them up. This is useful when the verified structure may contain
3459 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
3460 present unless the new PKCS7_NO_CRL flag is asserted.
3463 *) Extend ASN1 oid configuration module. It now additionally accepts the
3466 shortName = some long name, 1.2.3.4
3469 *) Reimplemented the BN_CTX implementation. There is now no more static
3470 limitation on the number of variables it can handle nor the depth of the
3471 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
3472 information can now expand as required, and rather than having a single
3473 static array of bignums, BN_CTX now uses a linked-list of such arrays
3474 allowing it to expand on demand whilst maintaining the usefulness of
3475 BN_CTX's "bundling".
3478 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
3479 to allow all RSA operations to function using a single BN_CTX.
3482 *) Preliminary support for certificate policy evaluation and checking. This
3483 is initially intended to pass the tests outlined in "Conformance Testing
3484 of Relying Party Client Certificate Path Processing Logic" v1.07.
3487 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
3488 remained unused and not that useful. A variety of other little bignum
3489 tweaks and fixes have also been made continuing on from the audit (see
3493 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
3494 associated ASN1, EVP and SSL functions and old ASN1 macros.
3497 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
3498 and this should never fail. So the return value from the use of
3499 BN_set_word() (which can fail due to needless expansion) is now deprecated;
3500 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
3503 *) BN_CTX_get() should return zero-valued bignums, providing the same
3504 initialised value as BN_new().
3505 [Geoff Thorpe, suggested by Ulf Möller]
3507 *) Support for inhibitAnyPolicy certificate extension.
3510 *) An audit of the BIGNUM code is underway, for which debugging code is
3511 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
3512 is considered valid when processing BIGNUMs, and causes execution to
3513 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
3514 further steps are taken to deliberately pollute unused data in BIGNUM
3515 structures to try and expose faulty code further on. For now, openssl will
3516 (in its default mode of operation) continue to tolerate the inconsistent
3517 forms that it has tolerated in the past, but authors and packagers should
3518 consider trying openssl and their own applications when compiled with
3519 these debugging symbols defined. It will help highlight potential bugs in
3520 their own code, and will improve the test coverage for OpenSSL itself. At
3521 some point, these tighter rules will become openssl's default to improve
3522 maintainability, though the assert()s and other overheads will remain only
3523 in debugging configurations. See bn.h for more details.
3524 [Geoff Thorpe, Nils Larsch, Ulf Möller]
3526 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
3527 that can only be obtained through BN_CTX_new() (which implicitly
3528 initialises it). The presence of this function only made it possible
3529 to overwrite an existing structure (and cause memory leaks).
3532 *) Because of the callback-based approach for implementing LHASH as a
3533 template type, lh_insert() adds opaque objects to hash-tables and
3534 lh_doall() or lh_doall_arg() are typically used with a destructor callback
3535 to clean up those corresponding objects before destroying the hash table
3536 (and losing the object pointers). So some over-zealous constifications in
3537 LHASH have been relaxed so that lh_insert() does not take (nor store) the
3538 objects as "const" and the lh_doall[_arg] callback wrappers are not
3539 prototyped to have "const" restrictions on the object pointers they are
3540 given (and so aren't required to cast them away any more).
3543 *) The tmdiff.h API was so ugly and minimal that our own timing utility
3544 (speed) prefers to use its own implementation. The two implementations
3545 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
3546 its object type properly exposed (MS_TM) instead of casting to/from "char
3547 *". This may still change yet if someone realises MS_TM and "ms_time_***"
3548 aren't necessarily the greatest nomenclatures - but this is what was used
3549 internally to the implementation so I've used that for now.
3552 *) Ensure that deprecated functions do not get compiled when
3553 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
3554 the self-tests were still using deprecated key-generation functions so
3555 these have been updated also.
3558 *) Reorganise PKCS#7 code to separate the digest location functionality
3559 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
3560 New function PKCS7_set_digest() to set the digest type for PKCS#7
3561 digestedData type. Add additional code to correctly generate the
3562 digestedData type and add support for this type in PKCS7 initialization
3566 *) New function PKCS7_set0_type_other() this initializes a PKCS7
3567 structure of type "other".
3570 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
3571 sure the loop does correctly stop and breaking ("division by zero")
3572 modulus operations are not performed. The (pre-generated) prime
3573 table crypto/bn/bn_prime.h was already correct, but it could not be
3574 re-generated on some platforms because of the "division by zero"
3575 situation in the script.
3576 [Ralf S. Engelschall]
3578 *) Update support for ECC-based TLS ciphersuites according to
3579 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
3580 SHA-1 now is only used for "small" curves (where the
3581 representation of a field element takes up to 24 bytes); for
3582 larger curves, the field element resulting from ECDH is directly
3583 used as premaster secret.
3584 [Douglas Stebila (Sun Microsystems Laboratories)]
3586 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
3587 curve secp160r1 to the tests.
3588 [Douglas Stebila (Sun Microsystems Laboratories)]
3590 *) Add the possibility to load symbols globally with DSO.
3591 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
3593 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
3594 control of the error stack.
3597 *) Add support for STORE in ENGINE.
3600 *) Add the STORE type. The intention is to provide a common interface
3601 to certificate and key stores, be they simple file-based stores, or
3602 HSM-type store, or LDAP stores, or...
3603 NOTE: The code is currently UNTESTED and isn't really used anywhere.
3606 *) Add a generic structure called OPENSSL_ITEM. This can be used to
3607 pass a list of arguments to any function as well as provide a way
3608 for a function to pass data back to the caller.
3611 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
3612 works like BUF_strdup() but can be used to duplicate a portion of
3613 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
3617 *) Add the function sk_find_ex() which works like sk_find(), but will
3618 return an index to an element even if an exact match couldn't be
3619 found. The index is guaranteed to point at the element where the
3620 searched-for key would be inserted to preserve sorting order.
3623 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
3624 takes an extra flags argument for optional functionality. Currently,
3625 the following flags are defined:
3627 OBJ_BSEARCH_VALUE_ON_NOMATCH
3628 This one gets OBJ_bsearch_ex() to return a pointer to the first
3629 element where the comparing function returns a negative or zero
3632 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
3633 This one gets OBJ_bsearch_ex() to return a pointer to the first
3634 element where the comparing function returns zero. This is useful
3635 if there are more than one element where the comparing function
3639 *) Make it possible to create self-signed certificates with 'openssl ca'
3640 in such a way that the self-signed certificate becomes part of the
3641 CA database and uses the same mechanisms for serial number generation
3642 as all other certificate signing. The new flag '-selfsign' enables
3643 this functionality. Adapt CA.sh and CA.pl.in.
3646 *) Add functionality to check the public key of a certificate request
3647 against a given private. This is useful to check that a certificate
3648 request can be signed by that key (self-signing).
3651 *) Make it possible to have multiple active certificates with the same
3652 subject in the CA index file. This is done only if the keyword
3653 'unique_subject' is set to 'no' in the main CA section (default
3654 if 'CA_default') of the configuration file. The value is saved
3655 with the database itself in a separate index attribute file,
3656 named like the index file with '.attr' appended to the name.
3659 *) Generate muti valued AVAs using '+' notation in config files for
3663 *) Support for nameConstraints certificate extension.
3666 *) Support for policyConstraints certificate extension.
3669 *) Support for policyMappings certificate extension.
3672 *) Make sure the default DSA_METHOD implementation only uses its
3673 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
3674 and change its own handlers to be NULL so as to remove unnecessary
3675 indirection. This lets alternative implementations fallback to the
3676 default implementation more easily.
3679 *) Support for directoryName in GeneralName related extensions
3683 *) Make it possible to link applications using Makefile.shared.
3684 Make that possible even when linking against static libraries!
3687 *) Support for single pass processing for S/MIME signing. This now
3688 means that S/MIME signing can be done from a pipe, in addition
3689 cleartext signing (multipart/signed type) is effectively streaming
3690 and the signed data does not need to be all held in memory.
3692 This is done with a new flag PKCS7_STREAM. When this flag is set
3693 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
3694 is done after the data is output (and digests calculated) in
3695 SMIME_write_PKCS7().
3698 *) Add full support for -rpath/-R, both in shared libraries and
3699 applications, at least on the platforms where it's known how
3703 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
3704 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
3705 will now compute a table of multiples of the generator that
3706 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
3707 faster (notably in the case of a single point multiplication,
3708 scalar * generator).
3709 [Nils Larsch, Bodo Moeller]
3711 *) IPv6 support for certificate extensions. The various extensions
3712 which use the IP:a.b.c.d can now take IPv6 addresses using the
3713 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
3717 *) Added an ENGINE that implements RSA by performing private key
3718 exponentiations with the GMP library. The conversions to and from
3719 GMP's mpz_t format aren't optimised nor are any montgomery forms
3720 cached, and on x86 it appears OpenSSL's own performance has caught up.
3721 However there are likely to be other architectures where GMP could
3722 provide a boost. This ENGINE is not built in by default, but it can be
3723 specified at Configure time and should be accompanied by the necessary
3724 linker additions, eg;
3725 ./config -DOPENSSL_USE_GMP -lgmp
3728 *) "openssl engine" will not display ENGINE/DSO load failure errors when
3729 testing availability of engines with "-t" - the old behaviour is
3730 produced by increasing the feature's verbosity with "-tt".
3733 *) ECDSA routines: under certain error conditions uninitialized BN objects
3734 could be freed. Solution: make sure initialization is performed early
3735 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
3739 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3740 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3741 software implementations. For DSA and DH, parameter generation can
3742 also be overriden by providing the appropriate method callbacks.
3745 *) Change the "progress" mechanism used in key-generation and
3746 primality testing to functions that take a new BN_GENCB pointer in
3747 place of callback/argument pairs. The new API functions have "_ex"
3748 postfixes and the older functions are reimplemented as wrappers for
3749 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
3750 declarations of the old functions to help (graceful) attempts to
3751 migrate to the new functions. Also, the new key-generation API
3752 functions operate on a caller-supplied key-structure and return
3753 success/failure rather than returning a key or NULL - this is to
3754 help make "keygen" another member function of RSA_METHOD etc.
3756 Example for using the new callback interface:
3758 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
3762 BN_GENCB_set(&my_cb, my_callback, my_arg);
3764 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
3765 /* For the meaning of a, b in calls to my_callback(), see the
3766 * documentation of the function that calls the callback.
3767 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
3768 * my_callback should return 1 if it wants BN_is_prime_ex()
3769 * to continue, or 0 to stop.
3774 *) Change the ZLIB compression method to be stateful, and make it
3775 available to TLS with the number defined in
3776 draft-ietf-tls-compression-04.txt.
3779 *) Add the ASN.1 structures and functions for CertificatePair, which
3780 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
3782 CertificatePair ::= SEQUENCE {
3783 forward [0] Certificate OPTIONAL,
3784 reverse [1] Certificate OPTIONAL,
3785 -- at least one of the pair shall be present -- }
3787 Also implement the PEM functions to read and write certificate
3788 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
3790 This needed to be defined, mostly for the sake of the LDAP
3791 attribute crossCertificatePair, but may prove useful elsewhere as
3795 *) Make it possible to inhibit symlinking of shared libraries in
3796 Makefile.shared, for Cygwin's sake.
3799 *) Extend the BIGNUM API by creating a function
3800 void BN_set_negative(BIGNUM *a, int neg);
3801 and a macro that behave like
3802 int BN_is_negative(const BIGNUM *a);
3804 to avoid the need to access 'a->neg' directly in applications.
3807 *) Implement fast modular reduction for pseudo-Mersenne primes
3808 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
3809 EC_GROUP_new_curve_GFp() will now automatically use this
3811 [Nils Larsch <nla@trustcenter.de>]
3813 *) Add new lock type (CRYPTO_LOCK_BN).
3816 *) Change the ENGINE framework to automatically load engines
3817 dynamically from specific directories unless they could be
3818 found to already be built in or loaded. Move all the
3819 current engines except for the cryptodev one to a new
3821 The engines in engines/ are built as shared libraries if
3822 the "shared" options was given to ./Configure or ./config.
3823 Otherwise, they are inserted in libcrypto.a.
3824 /usr/local/ssl/engines is the default directory for dynamic
3825 engines, but that can be overriden at configure time through
3826 the usual use of --prefix and/or --openssldir, and at run
3827 time with the environment variable OPENSSL_ENGINES.
3828 [Geoff Thorpe and Richard Levitte]
3830 *) Add Makefile.shared, a helper makefile to build shared
3831 libraries. Addapt Makefile.org.
3834 *) Add version info to Win32 DLLs.
3835 [Peter 'Luna' Runestig" <peter@runestig.com>]
3837 *) Add new 'medium level' PKCS#12 API. Certificates and keys
3838 can be added using this API to created arbitrary PKCS#12
3839 files while avoiding the low level API.
3841 New options to PKCS12_create(), key or cert can be NULL and
3842 will then be omitted from the output file. The encryption
3843 algorithm NIDs can be set to -1 for no encryption, the mac
3844 iteration count can be set to 0 to omit the mac.
3846 Enhance pkcs12 utility by making the -nokeys and -nocerts
3847 options work when creating a PKCS#12 file. New option -nomac
3848 to omit the mac, NONE can be set for an encryption algorithm.
3849 New code is modified to use the enhanced PKCS12_create()
3850 instead of the low level API.
3853 *) Extend ASN1 encoder to support indefinite length constructed
3854 encoding. This can output sequences tags and octet strings in
3855 this form. Modify pk7_asn1.c to support indefinite length
3856 encoding. This is experimental and needs additional code to
3857 be useful, such as an ASN1 bio and some enhanced streaming
3860 Extend template encode functionality so that tagging is passed
3861 down to the template encoder.
3864 *) Let 'openssl req' fail if an argument to '-newkey' is not
3865 recognized instead of using RSA as a default.
3868 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
3869 As these are not official, they are not included in "ALL";
3870 the "ECCdraft" ciphersuite group alias can be used to select them.
3871 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
3873 *) Add ECDH engine support.
3874 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
3876 *) Add ECDH in new directory crypto/ecdh/.
3877 [Douglas Stebila (Sun Microsystems Laboratories)]
3879 *) Let BN_rand_range() abort with an error after 100 iterations
3880 without success (which indicates a broken PRNG).
3883 *) Change BN_mod_sqrt() so that it verifies that the input value
3884 is really the square of the return value. (Previously,
3885 BN_mod_sqrt would show GIGO behaviour.)
3888 *) Add named elliptic curves over binary fields from X9.62, SECG,
3889 and WAP/WTLS; add OIDs that were still missing.
3891 [Sheueling Chang Shantz and Douglas Stebila
3892 (Sun Microsystems Laboratories)]
3894 *) Extend the EC library for elliptic curves over binary fields
3895 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
3898 EC_GF2m_simple_method
3902 EC_GROUP_new_curve_GF2m
3903 EC_GROUP_set_curve_GF2m
3904 EC_GROUP_get_curve_GF2m
3905 EC_POINT_set_affine_coordinates_GF2m
3906 EC_POINT_get_affine_coordinates_GF2m
3907 EC_POINT_set_compressed_coordinates_GF2m
3909 Point compression for binary fields is disabled by default for
3910 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
3913 As binary polynomials are represented as BIGNUMs, various members
3914 of the EC_GROUP and EC_POINT data structures can be shared
3915 between the implementations for prime fields and binary fields;
3916 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
3917 are essentially identical to their ..._GFp counterparts.
3918 (For simplicity, the '..._GFp' prefix has been dropped from
3919 various internal method names.)
3921 An internal 'field_div' method (similar to 'field_mul' and
3922 'field_sqr') has been added; this is used only for binary fields.
3924 [Sheueling Chang Shantz and Douglas Stebila
3925 (Sun Microsystems Laboratories)]
3927 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
3928 through methods ('mul', 'precompute_mult').
3930 The generic implementations (now internally called 'ec_wNAF_mul'
3931 and 'ec_wNAF_precomputed_mult') remain the default if these
3932 methods are undefined.
3934 [Sheueling Chang Shantz and Douglas Stebila
3935 (Sun Microsystems Laboratories)]
3937 *) New function EC_GROUP_get_degree, which is defined through
3938 EC_METHOD. For curves over prime fields, this returns the bit
3939 length of the modulus.
3941 [Sheueling Chang Shantz and Douglas Stebila
3942 (Sun Microsystems Laboratories)]
3944 *) New functions EC_GROUP_dup, EC_POINT_dup.
3945 (These simply call ..._new and ..._copy).
3947 [Sheueling Chang Shantz and Douglas Stebila
3948 (Sun Microsystems Laboratories)]
3950 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
3951 Polynomials are represented as BIGNUMs (where the sign bit is not
3952 used) in the following functions [macros]:
3955 BN_GF2m_sub [= BN_GF2m_add]
3956 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
3957 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
3958 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3960 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3961 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3962 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3963 BN_GF2m_cmp [= BN_ucmp]
3965 (Note that only the 'mod' functions are actually for fields GF(2^m).
3966 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3968 For some functions, an the irreducible polynomial defining a
3969 field can be given as an 'unsigned int[]' with strictly
3970 decreasing elements giving the indices of those bits that are set;
3971 i.e., p[] represents the polynomial
3972 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3974 p[0] > p[1] > ... > p[k] = 0.
3975 This applies to the following functions:
3980 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3981 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3983 BN_GF2m_mod_sqrt_arr
3984 BN_GF2m_mod_solve_quad_arr
3988 Conversion can be performed by the following functions:
3993 bntest.c has additional tests for binary polynomial arithmetic.
3995 Two implementations for BN_GF2m_mod_div() are available.
3996 The default algorithm simply uses BN_GF2m_mod_inv() and
3997 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3998 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3999 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
4001 [Sheueling Chang Shantz and Douglas Stebila
4002 (Sun Microsystems Laboratories)]
4004 *) Add new error code 'ERR_R_DISABLED' that can be used when some
4005 functionality is disabled at compile-time.
4006 [Douglas Stebila <douglas.stebila@sun.com>]
4008 *) Change default behaviour of 'openssl asn1parse' so that more
4009 information is visible when viewing, e.g., a certificate:
4011 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
4012 mode the content of non-printable OCTET STRINGs is output in a
4013 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
4014 avoid the appearance of a printable string.
4015 [Nils Larsch <nla@trustcenter.de>]
4017 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
4019 EC_GROUP_set_asn1_flag()
4020 EC_GROUP_get_asn1_flag()
4021 EC_GROUP_set_point_conversion_form()
4022 EC_GROUP_get_point_conversion_form()
4023 These control ASN1 encoding details:
4024 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
4025 has been set to OPENSSL_EC_NAMED_CURVE.
4026 - Points are encoded in uncompressed form by default; options for
4027 asn1_for are as for point2oct, namely
4028 POINT_CONVERSION_COMPRESSED
4029 POINT_CONVERSION_UNCOMPRESSED
4030 POINT_CONVERSION_HYBRID
4032 Also add 'seed' and 'seed_len' members to EC_GROUP with access
4035 EC_GROUP_get0_seed()
4036 EC_GROUP_get_seed_len()
4037 This is used only for ASN1 purposes (so far).
4038 [Nils Larsch <nla@trustcenter.de>]
4040 *) Add 'field_type' member to EC_METHOD, which holds the NID
4041 of the appropriate field type OID. The new function
4042 EC_METHOD_get_field_type() returns this value.
4043 [Nils Larsch <nla@trustcenter.de>]
4048 EC_POINT_point2hex()
4049 EC_POINT_hex2point()
4050 providing useful interfaces to EC_POINT_point2oct() and
4051 EC_POINT_oct2point().
4052 [Nils Larsch <nla@trustcenter.de>]
4054 *) Change internals of the EC library so that the functions
4055 EC_GROUP_set_generator()
4056 EC_GROUP_get_generator()
4057 EC_GROUP_get_order()
4058 EC_GROUP_get_cofactor()
4059 are implemented directly in crypto/ec/ec_lib.c and not dispatched
4060 to methods, which would lead to unnecessary code duplication when
4061 adding different types of curves.
4062 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
4064 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
4065 arithmetic, and such that modified wNAFs are generated
4066 (which avoid length expansion in many cases).
4069 *) Add a function EC_GROUP_check_discriminant() (defined via
4070 EC_METHOD) that verifies that the curve discriminant is non-zero.
4072 Add a function EC_GROUP_check() that makes some sanity tests
4073 on a EC_GROUP, its generator and order. This includes
4074 EC_GROUP_check_discriminant().
4075 [Nils Larsch <nla@trustcenter.de>]
4077 *) Add ECDSA in new directory crypto/ecdsa/.
4079 Add applications 'openssl ecparam' and 'openssl ecdsa'
4080 (these are based on 'openssl dsaparam' and 'openssl dsa').
4082 ECDSA support is also included in various other files across the
4083 library. Most notably,
4084 - 'openssl req' now has a '-newkey ecdsa:file' option;
4085 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
4086 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
4087 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
4088 them suitable for ECDSA where domain parameters must be
4089 extracted before the specific public key;
4090 - ECDSA engine support has been added.
4091 [Nils Larsch <nla@trustcenter.de>]
4093 *) Include some named elliptic curves, and add OIDs from X9.62,
4094 SECG, and WAP/WTLS. Each curve can be obtained from the new
4096 EC_GROUP_new_by_curve_name(),
4097 and the list of available named curves can be obtained with
4098 EC_get_builtin_curves().
4099 Also add a 'curve_name' member to EC_GROUP objects, which can be
4101 EC_GROUP_set_curve_name()
4102 EC_GROUP_get_curve_name()
4103 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
4105 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4106 was actually never needed) and in BN_mul(). The removal in BN_mul()
4107 required a small change in bn_mul_part_recursive() and the addition
4108 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4109 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4110 bn_sub_words() and bn_add_words() except they take arrays with
4114 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
4116 *) Cleanse PEM buffers before freeing them since they may contain
4118 [Benjamin Bennett <ben@psc.edu>]
4120 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4121 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4122 authentication-only ciphersuites.
4125 *) Since AES128 and AES256 share a single mask bit in the logic of
4126 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4127 kludge to work properly if AES128 is available and AES256 isn't.
4130 *) Expand security boundary to match 1.1.1 module.
4133 *) Remove redundant features: hash file source, editing of test vectors
4134 modify fipsld to use external fips_premain.c signature.
4137 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
4138 run algorithm test programs.
4141 *) Make algorithm test programs more tolerant of whitespace.
4144 *) Have SSL/TLS server implementation tolerate "mismatched" record
4145 protocol version while receiving ClientHello even if the
4146 ClientHello is fragmented. (The server can't insist on the
4147 particular protocol version it has chosen before the ServerHello
4148 message has informed the client about his choice.)
4151 *) Load error codes if they are not already present instead of using a
4152 static variable. This allows them to be cleanly unloaded and reloaded.
4155 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
4157 *) Introduce limits to prevent malicious keys being able to
4158 cause a denial of service. (CVE-2006-2940)
4159 [Steve Henson, Bodo Moeller]
4161 *) Fix ASN.1 parsing of certain invalid structures that can result
4162 in a denial of service. (CVE-2006-2937) [Steve Henson]
4164 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4165 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4167 *) Fix SSL client code which could crash if connecting to a
4168 malicious SSLv2 server. (CVE-2006-4343)
4169 [Tavis Ormandy and Will Drewry, Google Security Team]
4171 *) Change ciphersuite string processing so that an explicit
4172 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
4173 will no longer include "AES128-SHA"), and any other similar
4174 ciphersuite (same bitmap) from *other* protocol versions (so that
4175 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
4176 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
4177 changes from 0.9.8b and 0.9.8d.
4180 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
4182 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4183 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4185 *) Change the Unix randomness entropy gathering to use poll() when
4186 possible instead of select(), since the latter has some
4187 undesirable limitations.
4188 [Darryl Miles via Richard Levitte and Bodo Moeller]
4190 *) Disable rogue ciphersuites:
4192 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4193 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4194 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4196 The latter two were purportedly from
4197 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4200 Also deactive the remaining ciphersuites from
4201 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4202 unofficial, and the ID has long expired.
4205 *) Fix RSA blinding Heisenbug (problems sometimes occured on
4206 dual-core machines) and other potential thread-safety issues.
4209 Changes between 0.9.7i and 0.9.7j [04 May 2006]
4211 *) Adapt fipsld and the build system to link against the validated FIPS
4212 module in FIPS mode.
4215 *) Fixes for VC++ 2005 build under Windows.
4218 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
4219 from a Windows bash shell such as MSYS. It is autodetected from the
4220 "config" script when run from a VC++ environment. Modify standard VC++
4221 build to use fipscanister.o from the GNU make build.
4224 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
4226 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
4227 The value now differs depending on if you build for FIPS or not.
4228 BEWARE! A program linked with a shared FIPSed libcrypto can't be
4229 safely run with a non-FIPSed libcrypto, as it may crash because of
4230 the difference induced by this change.
4233 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
4235 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4236 (part of SSL_OP_ALL). This option used to disable the
4237 countermeasure against man-in-the-middle protocol-version
4238 rollback in the SSL 2.0 server implementation, which is a bad
4239 idea. (CVE-2005-2969)
4241 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4242 for Information Security, National Institute of Advanced Industrial
4243 Science and Technology [AIST], Japan)]
4245 *) Minimal support for X9.31 signatures and PSS padding modes. This is
4246 mainly for FIPS compliance and not fully integrated at this stage.
4249 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
4250 the exponentiation using a fixed-length exponent. (Otherwise,
4251 the information leaked through timing could expose the secret key
4252 after many signatures; cf. Bleichenbacher's attack on DSA with
4256 *) Make a new fixed-window mod_exp implementation the default for
4257 RSA, DSA, and DH private-key operations so that the sequence of
4258 squares and multiplies and the memory access pattern are
4259 independent of the particular secret key. This will mitigate
4260 cache-timing and potential related attacks.
4262 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
4263 and this is automatically used by BN_mod_exp_mont() if the new flag
4264 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
4265 will use this BN flag for private exponents unless the flag
4266 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
4267 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
4269 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
4271 *) Change the client implementation for SSLv23_method() and
4272 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
4273 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
4274 (Previously, the SSL 2.0 backwards compatible Client Hello
4275 message format would be used even with SSL_OP_NO_SSLv2.)
4278 *) Add support for smime-type MIME parameter in S/MIME messages which some
4282 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
4283 a threadsafe manner. Modify rsa code to use new function and add calls
4284 to dsa and dh code (which had race conditions before).
4287 *) Include the fixed error library code in the C error file definitions
4288 instead of fixing them up at runtime. This keeps the error code
4289 structures constant.
4292 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
4294 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
4297 *) Fixes for newer kerberos headers. NB: the casts are needed because
4298 the 'length' field is signed on one version and unsigned on another
4299 with no (?) obvious way to tell the difference, without these VC++
4300 complains. Also the "definition" of FAR (blank) is no longer included
4301 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
4302 some needed definitions.
4305 *) Undo Cygwin change.
4308 *) Added support for proxy certificates according to RFC 3820.
4309 Because they may be a security thread to unaware applications,
4310 they must be explicitely allowed in run-time. See
4311 docs/HOWTO/proxy_certificates.txt for further information.
4314 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
4316 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
4317 server and client random values. Previously
4318 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
4319 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
4321 This change has negligible security impact because:
4323 1. Server and client random values still have 24 bytes of pseudo random
4326 2. Server and client random values are sent in the clear in the initial
4329 3. The master secret is derived using the premaster secret (48 bytes in
4330 size for static RSA ciphersuites) as well as client server and random
4333 The OpenSSL team would like to thank the UK NISCC for bringing this issue
4336 [Stephen Henson, reported by UK NISCC]
4338 *) Use Windows randomness collection on Cygwin.
4341 *) Fix hang in EGD/PRNGD query when communication socket is closed
4342 prematurely by EGD/PRNGD.
4343 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
4345 *) Prompt for pass phrases when appropriate for PKCS12 input format.
4348 *) Back-port of selected performance improvements from development
4349 branch, as well as improved support for PowerPC platforms.
4352 *) Add lots of checks for memory allocation failure, error codes to indicate
4353 failure and freeing up memory if a failure occurs.
4354 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
4356 *) Add new -passin argument to dgst.
4359 *) Perform some character comparisons of different types in X509_NAME_cmp:
4360 this is needed for some certificates that reencode DNs into UTF8Strings
4361 (in violation of RFC3280) and can't or wont issue name rollover
4365 *) Make an explicit check during certificate validation to see that
4366 the CA setting in each certificate on the chain is correct. As a
4367 side effect always do the following basic checks on extensions,
4368 not just when there's an associated purpose to the check:
4370 - if there is an unhandled critical extension (unless the user
4371 has chosen to ignore this fault)
4372 - if the path length has been exceeded (if one is set at all)
4373 - that certain extensions fit the associated purpose (if one has
4377 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
4379 *) Avoid a race condition when CRLs are checked in a multi threaded
4380 environment. This would happen due to the reordering of the revoked
4381 entries during signature checking and serial number lookup. Now the
4382 encoding is cached and the serial number sort performed under a lock.
4383 Add new STACK function sk_is_sorted().
4386 *) Add Delta CRL to the extension code.
4389 *) Various fixes to s3_pkt.c so alerts are sent properly.
4390 [David Holmes <d.holmes@f5.com>]
4392 *) Reduce the chances of duplicate issuer name and serial numbers (in
4393 violation of RFC3280) using the OpenSSL certificate creation utilities.
4394 This is done by creating a random 64 bit value for the initial serial
4395 number when a serial number file is created or when a self signed
4396 certificate is created using 'openssl req -x509'. The initial serial
4397 number file is created using 'openssl x509 -next_serial' in CA.pl
4398 rather than being initialized to 1.
4401 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
4403 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4404 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4405 [Joe Orton, Steve Henson]
4407 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
4409 [Joe Orton, Steve Henson]
4411 *) Make it possible to have multiple active certificates with the same
4412 subject in the CA index file. This is done only if the keyword
4413 'unique_subject' is set to 'no' in the main CA section (default
4414 if 'CA_default') of the configuration file. The value is saved
4415 with the database itself in a separate index attribute file,
4416 named like the index file with '.attr' appended to the name.
4419 *) X509 verify fixes. Disable broken certificate workarounds when
4420 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
4421 keyUsage extension present. Don't accept CRLs with unhandled critical
4422 extensions: since verify currently doesn't process CRL extensions this
4423 rejects a CRL with *any* critical extensions. Add new verify error codes
4427 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
4428 A clarification of RFC2560 will require the use of OCTET STRINGs and
4429 some implementations cannot handle the current raw format. Since OpenSSL
4430 copies and compares OCSP nonces as opaque blobs without any attempt at
4431 parsing them this should not create any compatibility issues.
4434 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
4435 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
4436 this HMAC (and other) operations are several times slower than OpenSSL
4440 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
4441 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4443 *) Use the correct content when signing type "other".
4446 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
4448 *) Fix various bugs revealed by running the NISCC test suite:
4450 Stop out of bounds reads in the ASN1 code when presented with
4451 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4453 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
4455 If verify callback ignores invalid public key errors don't try to check
4456 certificate signature with the NULL public key.
4460 *) New -ignore_err option in ocsp application to stop the server
4461 exiting on the first error in a request.
4464 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4465 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4469 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4470 extra data after the compression methods not only for TLS 1.0
4471 but also for SSL 3.0 (as required by the specification).
4472 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4474 *) Change X509_certificate_type() to mark the key as exported/exportable
4475 when it's 512 *bits* long, not 512 bytes.
4478 *) Change AES_cbc_encrypt() so it outputs exact multiple of
4479 blocks during encryption.
4482 *) Various fixes to base64 BIO and non blocking I/O. On write
4483 flushes were not handled properly if the BIO retried. On read
4484 data was not being buffered properly and had various logic bugs.
4485 This also affects blocking I/O when the data being decoded is a
4489 *) Various S/MIME bugfixes and compatibility changes:
4490 output correct application/pkcs7 MIME type if
4491 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
4492 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
4493 of files as .eml work). Correctly handle very long lines in MIME
4497 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
4499 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4500 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4501 a protocol version number mismatch like a decryption error
4502 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4505 *) Turn on RSA blinding by default in the default implementation
4506 to avoid a timing attack. Applications that don't want it can call
4507 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4508 They would be ill-advised to do so in most cases.
4509 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4511 *) Change RSA blinding code so that it works when the PRNG is not
4512 seeded (in this case, the secret RSA exponent is abused as
4513 an unpredictable seed -- if it is not unpredictable, there
4514 is no point in blinding anyway). Make RSA blinding thread-safe
4515 by remembering the creator's thread ID in rsa->blinding and
4516 having all other threads use local one-time blinding factors
4517 (this requires more computation than sharing rsa->blinding, but
4518 avoids excessive locking; and if an RSA object is not shared
4519 between threads, blinding will still be very fast).
4522 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
4523 ENGINE as defaults for all supported algorithms irrespective of
4524 the 'flags' parameter. 'flags' is now honoured, so applications
4525 should make sure they are passing it correctly.
4528 *) Target "mingw" now allows native Windows code to be generated in
4529 the Cygwin environment as well as with the MinGW compiler.
4532 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
4534 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4535 via timing by performing a MAC computation even if incorrrect
4536 block cipher padding has been found. This is a countermeasure
4537 against active attacks where the attacker has to distinguish
4538 between bad padding and a MAC verification error. (CVE-2003-0078)
4540 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4541 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4542 Martin Vuagnoux (EPFL, Ilion)]
4544 *) Make the no-err option work as intended. The intention with no-err
4545 is not to have the whole error stack handling routines removed from
4546 libcrypto, it's only intended to remove all the function name and
4547 reason texts, thereby removing some of the footprint that may not
4548 be interesting if those errors aren't displayed anyway.
4550 NOTE: it's still possible for any application or module to have it's
4551 own set of error texts inserted. The routines are there, just not
4552 used by default when no-err is given.
4555 *) Add support for FreeBSD on IA64.
4556 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
4558 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
4559 Kerberos function mit_des_cbc_cksum(). Before this change,
4560 the value returned by DES_cbc_cksum() was like the one from
4561 mit_des_cbc_cksum(), except the bytes were swapped.
4562 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
4564 *) Allow an application to disable the automatic SSL chain building.
4565 Before this a rather primitive chain build was always performed in
4566 ssl3_output_cert_chain(): an application had no way to send the
4567 correct chain if the automatic operation produced an incorrect result.
4569 Now the chain builder is disabled if either:
4571 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
4573 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
4575 The reasoning behind this is that an application would not want the
4576 auto chain building to take place if extra chain certificates are
4577 present and it might also want a means of sending no additional
4578 certificates (for example the chain has two certificates and the
4582 *) Add the possibility to build without the ENGINE framework.
4583 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4585 *) Under Win32 gmtime() can return NULL: check return value in
4586 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
4589 *) DSA routines: under certain error conditions uninitialized BN objects
4590 could be freed. Solution: make sure initialization is performed early
4591 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
4592 Nils Larsch <nla@trustcenter.de> via PR#459)
4595 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
4596 checked on reconnect on the client side, therefore session resumption
4597 could still fail with a "ssl session id is different" error. This
4598 behaviour is masked when SSL_OP_ALL is used due to
4599 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
4600 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4601 followup to PR #377.
4604 *) IA-32 assembler support enhancements: unified ELF targets, support
4605 for SCO/Caldera platforms, fix for Cygwin shared build.
4608 *) Add support for FreeBSD on sparc64. As a consequence, support for
4609 FreeBSD on non-x86 processors is separate from x86 processors on
4610 the config script, much like the NetBSD support.
4611 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
4613 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
4615 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
4618 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
4619 code (06) was taken as the first octet of the session ID and the last
4620 octet was ignored consequently. As a result SSLv2 client side session
4621 caching could not have worked due to the session ID mismatch between
4623 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4627 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
4628 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
4632 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
4633 seems that in spite of existing for more than a year, many application
4634 author have done nothing to provide the necessary callbacks, which
4635 means that this particular engine will not work properly anywhere.
4636 This is a very unfortunate situation which forces us, in the name
4637 of usability, to give the hw_ncipher.c a static lock, which is part
4639 NOTE: This is for the 0.9.7 series ONLY. This hack will never
4640 appear in 0.9.8 or later. We EXPECT application authors to have
4641 dealt properly with this when 0.9.8 is released (unless we actually
4642 make such changes in the libcrypto locking code that changes will
4643 have to be made anyway).
4646 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
4647 octets have been read, EOF or an error occurs. Without this change
4648 some truncated ASN1 structures will not produce an error.
4651 *) Disable Heimdal support, since it hasn't been fully implemented.
4652 Still give the possibility to force the use of Heimdal, but with
4653 warnings and a request that patches get sent to openssl-dev.
4656 *) Add the VC-CE target, introduce the WINCE sysname, and add
4657 INSTALL.WCE and appropriate conditionals to make it build.
4658 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4660 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
4661 cygssl-x.y.z.dll, where x, y and z are the major, minor and
4662 edit numbers of the version.
4663 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4665 *) Introduce safe string copy and catenation functions
4666 (BUF_strlcpy() and BUF_strlcat()).
4667 [Ben Laurie (CHATS) and Richard Levitte]
4669 *) Avoid using fixed-size buffers for one-line DNs.
4670 [Ben Laurie (CHATS)]
4672 *) Add BUF_MEM_grow_clean() to avoid information leakage when
4673 resizing buffers containing secrets, and use where appropriate.
4674 [Ben Laurie (CHATS)]
4676 *) Avoid using fixed size buffers for configuration file location.
4677 [Ben Laurie (CHATS)]
4679 *) Avoid filename truncation for various CA files.
4680 [Ben Laurie (CHATS)]
4682 *) Use sizeof in preference to magic numbers.
4683 [Ben Laurie (CHATS)]
4685 *) Avoid filename truncation in cert requests.
4686 [Ben Laurie (CHATS)]
4688 *) Add assertions to check for (supposedly impossible) buffer
4690 [Ben Laurie (CHATS)]
4692 *) Don't cache truncated DNS entries in the local cache (this could
4693 potentially lead to a spoofing attack).
4694 [Ben Laurie (CHATS)]
4696 *) Fix various buffers to be large enough for hex/decimal
4697 representations in a platform independent manner.
4698 [Ben Laurie (CHATS)]
4700 *) Add CRYPTO_realloc_clean() to avoid information leakage when
4701 resizing buffers containing secrets, and use where appropriate.
4702 [Ben Laurie (CHATS)]
4704 *) Add BIO_indent() to avoid much slightly worrying code to do
4706 [Ben Laurie (CHATS)]
4708 *) Convert sprintf()/BIO_puts() to BIO_printf().
4709 [Ben Laurie (CHATS)]
4711 *) buffer_gets() could terminate with the buffer only half
4713 [Ben Laurie (CHATS)]
4715 *) Add assertions to prevent user-supplied crypto functions from
4716 overflowing internal buffers by having large block sizes, etc.
4717 [Ben Laurie (CHATS)]
4719 *) New OPENSSL_assert() macro (similar to assert(), but enabled
4721 [Ben Laurie (CHATS)]
4723 *) Eliminate unused copy of key in RC4.
4724 [Ben Laurie (CHATS)]
4726 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
4727 [Ben Laurie (CHATS)]
4729 *) Fix off-by-one error in EGD path.
4730 [Ben Laurie (CHATS)]
4732 *) If RANDFILE path is too long, ignore instead of truncating.
4733 [Ben Laurie (CHATS)]
4735 *) Eliminate unused and incorrectly sized X.509 structure
4737 [Ben Laurie (CHATS)]
4739 *) Eliminate unused and dangerous function knumber().
4740 [Ben Laurie (CHATS)]
4742 *) Eliminate unused and dangerous structure, KSSL_ERR.
4743 [Ben Laurie (CHATS)]
4745 *) Protect against overlong session ID context length in an encoded
4746 session object. Since these are local, this does not appear to be
4748 [Ben Laurie (CHATS)]
4750 *) Change from security patch (see 0.9.6e below) that did not affect
4751 the 0.9.6 release series:
4753 Remote buffer overflow in SSL3 protocol - an attacker could
4754 supply an oversized master key in Kerberos-enabled versions.
4756 [Ben Laurie (CHATS)]
4758 *) Change the SSL kerb5 codes to match RFC 2712.
4761 *) Make -nameopt work fully for req and add -reqopt switch.
4762 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
4764 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
4765 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
4767 *) Make sure tests can be performed even if the corresponding algorithms
4768 have been removed entirely. This was also the last step to make
4769 OpenSSL compilable with DJGPP under all reasonable conditions.
4770 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
4772 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
4773 to allow version independent disabling of normally unselected ciphers,
4774 which may be activated as a side-effect of selecting a single cipher.
4776 (E.g., cipher list string "RSA" enables ciphersuites that are left
4777 out of "ALL" because they do not provide symmetric encryption.
4778 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
4779 [Lutz Jaenicke, Bodo Moeller]
4781 *) Add appropriate support for separate platform-dependent build
4782 directories. The recommended way to make a platform-dependent
4783 build directory is the following (tested on Linux), maybe with
4786 # Place yourself outside of the OpenSSL source tree. In
4787 # this example, the environment variable OPENSSL_SOURCE
4788 # is assumed to contain the absolute OpenSSL source directory.
4789 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
4790 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
4791 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
4792 mkdir -p `dirname $F`
4793 ln -s $OPENSSL_SOURCE/$F $F
4796 To be absolutely sure not to disturb the source tree, a "make clean"
4797 is a good thing. If it isn't successfull, don't worry about it,
4798 it probably means the source directory is very clean.
4801 *) Make sure any ENGINE control commands make local copies of string
4802 pointers passed to them whenever necessary. Otherwise it is possible
4803 the caller may have overwritten (or deallocated) the original string
4804 data when a later ENGINE operation tries to use the stored values.
4805 [Götz Babin-Ebell <babinebell@trustcenter.de>]
4807 *) Improve diagnostics in file reading and command-line digests.
4808 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
4810 *) Add AES modes CFB and OFB to the object database. Correct an
4811 error in AES-CFB decryption.
4814 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
4815 allows existing EVP_CIPHER_CTX structures to be reused after
4816 calling EVP_*Final(). This behaviour is used by encryption
4817 BIOs and some applications. This has the side effect that
4818 applications must explicitly clean up cipher contexts with
4819 EVP_CIPHER_CTX_cleanup() or they will leak memory.
4822 *) Check the values of dna and dnb in bn_mul_recursive before calling
4823 bn_mul_comba (a non zero value means the a or b arrays do not contain
4824 n2 elements) and fallback to bn_mul_normal if either is not zero.
4827 *) Fix escaping of non-ASCII characters when using the -subj option
4828 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
4831 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
4832 form for "surname", serialNumber has no short form.
4833 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
4834 therefore remove "mail" short name for "internet 7".
4835 The OID for unique identifiers in X509 certificates is
4836 x500UniqueIdentifier, not uniqueIdentifier.
4837 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
4840 *) Add an "init" command to the ENGINE config module and auto initialize
4841 ENGINEs. Without any "init" command the ENGINE will be initialized
4842 after all ctrl commands have been executed on it. If init=1 the
4843 ENGINE is initailized at that point (ctrls before that point are run
4844 on the uninitialized ENGINE and after on the initialized one). If
4845 init=0 then the ENGINE will not be iniatialized at all.
4848 *) Fix the 'app_verify_callback' interface so that the user-defined
4849 argument is actually passed to the callback: In the
4850 SSL_CTX_set_cert_verify_callback() prototype, the callback
4851 declaration has been changed from
4854 int (*cb)(X509_STORE_CTX *,void *);
4855 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
4856 i=s->ctx->app_verify_callback(&ctx)
4857 has been changed into
4858 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
4860 To update applications using SSL_CTX_set_cert_verify_callback(),
4861 a dummy argument can be added to their callback functions.
4862 [D. K. Smetters <smetters@parc.xerox.com>]
4864 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
4865 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
4867 *) Add and OPENSSL_LOAD_CONF define which will cause
4868 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
4869 This allows older applications to transparently support certain
4870 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
4871 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
4872 load the config file and OPENSSL_add_all_algorithms_conf() which will
4873 always load it have also been added.
4876 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
4877 Adjust NIDs and EVP layer.
4878 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4880 *) Config modules support in openssl utility.
4882 Most commands now load modules from the config file,
4883 though in a few (such as version) this isn't done
4884 because it couldn't be used for anything.
4886 In the case of ca and req the config file used is
4887 the same as the utility itself: that is the -config
4888 command line option can be used to specify an
4892 *) Move default behaviour from OPENSSL_config(). If appname is NULL
4893 use "openssl_conf" if filename is NULL use default openssl config file.
4896 *) Add an argument to OPENSSL_config() to allow the use of an alternative
4897 config section name. Add a new flag to tolerate a missing config file
4898 and move code to CONF_modules_load_file().
4901 *) Support for crypto accelerator cards from Accelerated Encryption
4902 Processing, www.aep.ie. (Use engine 'aep')
4903 The support was copied from 0.9.6c [engine] and adapted/corrected
4904 to work with the new engine framework.
4905 [AEP Inc. and Richard Levitte]
4907 *) Support for SureWare crypto accelerator cards from Baltimore
4908 Technologies. (Use engine 'sureware')
4909 The support was copied from 0.9.6c [engine] and adapted
4910 to work with the new engine framework.
4913 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
4914 make the newer ENGINE framework commands for the CHIL engine work.
4915 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
4917 *) Make it possible to produce shared libraries on ReliantUNIX.
4918 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
4920 *) Add the configuration target debug-linux-ppro.
4921 Make 'openssl rsa' use the general key loading routines
4922 implemented in apps.c, and make those routines able to
4923 handle the key format FORMAT_NETSCAPE and the variant
4925 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4927 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4928 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4930 *) Add -keyform to rsautl, and document -engine.
4931 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
4933 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
4934 BIO_R_NO_SUCH_FILE error code rather than the generic
4935 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
4938 *) Add new functions
4940 ERR_peek_last_error_line
4941 ERR_peek_last_error_line_data.
4942 These are similar to
4945 ERR_peek_error_line_data,
4946 but report on the latest error recorded rather than the first one
4947 still in the error queue.
4948 [Ben Laurie, Bodo Moeller]
4950 *) default_algorithms option in ENGINE config module. This allows things
4952 default_algorithms = ALL
4953 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
4956 *) Prelminary ENGINE config module.
4959 *) New experimental application configuration code.
4962 *) Change the AES code to follow the same name structure as all other
4963 symmetric ciphers, and behave the same way. Move everything to
4964 the directory crypto/aes, thereby obsoleting crypto/rijndael.
4965 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4967 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
4968 [Ben Laurie and Theo de Raadt]
4970 *) Add option to output public keys in req command.
4971 [Massimiliano Pala madwolf@openca.org]
4973 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4974 (up to about 10% better than before for P-192 and P-224).
4977 *) New functions/macros
4979 SSL_CTX_set_msg_callback(ctx, cb)
4980 SSL_CTX_set_msg_callback_arg(ctx, arg)
4981 SSL_set_msg_callback(ssl, cb)
4982 SSL_set_msg_callback_arg(ssl, arg)
4984 to request calling a callback function
4986 void cb(int write_p, int version, int content_type,
4987 const void *buf, size_t len, SSL *ssl, void *arg)
4989 whenever a protocol message has been completely received
4990 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4991 protocol version according to which the SSL library interprets
4992 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4993 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4994 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4995 specification (change_cipher_spec(20), alert(21), handshake(22)).
4996 'buf' and 'len' point to the actual message, 'ssl' to the
4997 SSL object, and 'arg' is the application-defined value set by
4998 SSL[_CTX]_set_msg_callback_arg().
5000 'openssl s_client' and 'openssl s_server' have new '-msg' options
5001 to enable a callback that displays all protocol messages.
5004 *) Change the shared library support so shared libraries are built as
5005 soon as the corresponding static library is finished, and thereby get
5006 openssl and the test programs linked against the shared library.
5007 This still only happens when the keyword "shard" has been given to
5008 the configuration scripts.
5010 NOTE: shared library support is still an experimental thing, and
5011 backward binary compatibility is still not guaranteed.
5012 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
5014 *) Add support for Subject Information Access extension.
5015 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5017 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
5018 additional bytes when new memory had to be allocated, not just
5019 when reusing an existing buffer.
5022 *) New command line and configuration option 'utf8' for the req command.
5023 This allows field values to be specified as UTF8 strings.
5026 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
5027 runs for the former and machine-readable output for the latter.
5030 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
5031 of the e-mail address in the DN (i.e., it will go into a certificate
5032 extension only). The new configuration file option 'email_in_dn = no'
5033 has the same effect.
5034 [Massimiliano Pala madwolf@openca.org]
5036 *) Change all functions with names starting with des_ to be starting
5037 with DES_ instead. Add wrappers that are compatible with libdes,
5038 but are named _ossl_old_des_*. Finally, add macros that map the
5039 des_* symbols to the corresponding _ossl_old_des_* if libdes
5040 compatibility is desired. If OpenSSL 0.9.6c compatibility is
5041 desired, the des_* symbols will be mapped to DES_*, with one
5044 Since we provide two compatibility mappings, the user needs to
5045 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
5046 compatibility is desired. The default (i.e., when that macro
5047 isn't defined) is OpenSSL 0.9.6c compatibility.
5049 There are also macros that enable and disable the support of old
5050 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
5051 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
5052 are defined, the default will apply: to support the old des routines.
5054 In either case, one must include openssl/des.h to get the correct
5055 definitions. Do not try to just include openssl/des_old.h, that
5058 NOTE: This is a major break of an old API into a new one. Software
5059 authors are encouraged to switch to the DES_ style functions. Some
5060 time in the future, des_old.h and the libdes compatibility functions
5061 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
5062 default), and then completely removed.
5065 *) Test for certificates which contain unsupported critical extensions.
5066 If such a certificate is found during a verify operation it is
5067 rejected by default: this behaviour can be overridden by either
5068 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
5069 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
5070 X509_supported_extension() has also been added which returns 1 if a
5071 particular extension is supported.
5074 *) Modify the behaviour of EVP cipher functions in similar way to digests
5075 to retain compatibility with existing code.
5078 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
5079 compatibility with existing code. In particular the 'ctx' parameter does
5080 not have to be to be initialized before the call to EVP_DigestInit() and
5081 it is tidied up after a call to EVP_DigestFinal(). New function
5082 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
5083 EVP_MD_CTX_copy() changed to not require the destination to be
5084 initialized valid and new function EVP_MD_CTX_copy_ex() added which
5085 requires the destination to be valid.
5087 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
5088 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
5091 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
5092 so that complete 'Handshake' protocol structures are kept in memory
5093 instead of overwriting 'msg_type' and 'length' with 'body' data.
5096 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
5097 [Massimo Santin via Richard Levitte]
5099 *) Major restructuring to the underlying ENGINE code. This includes
5100 reduction of linker bloat, separation of pure "ENGINE" manipulation
5101 (initialisation, etc) from functionality dealing with implementations
5102 of specific crypto iterfaces. This change also introduces integrated
5103 support for symmetric ciphers and digest implementations - so ENGINEs
5104 can now accelerate these by providing EVP_CIPHER and EVP_MD
5105 implementations of their own. This is detailed in crypto/engine/README
5106 as it couldn't be adequately described here. However, there are a few
5107 API changes worth noting - some RSA, DSA, DH, and RAND functions that
5108 were changed in the original introduction of ENGINE code have now
5109 reverted back - the hooking from this code to ENGINE is now a good
5110 deal more passive and at run-time, operations deal directly with
5111 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
5112 dereferencing through an ENGINE pointer any more. Also, the ENGINE
5113 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
5114 they were not being used by the framework as there is no concept of a
5115 BIGNUM_METHOD and they could not be generalised to the new
5116 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
5117 ENGINE_cpy() has been removed as it cannot be consistently defined in
5121 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
5124 *) Change mkdef.pl to sort symbols that get the same entry number,
5125 and make sure the automatically generated functions ERR_load_*
5126 become part of libeay.num as well.
5129 *) New function SSL_renegotiate_pending(). This returns true once
5130 renegotiation has been requested (either SSL_renegotiate() call
5131 or HelloRequest/ClientHello receveived from the peer) and becomes
5132 false once a handshake has been completed.
5133 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
5134 sends a HelloRequest, but does not ensure that a handshake takes
5135 place. SSL_renegotiate_pending() is useful for checking if the
5136 client has followed the request.)
5139 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
5140 By default, clients may request session resumption even during
5141 renegotiation (if session ID contexts permit); with this option,
5142 session resumption is possible only in the first handshake.
5144 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
5145 more bits available for options that should not be part of
5146 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
5149 *) Add some demos for certificate and certificate request creation.
5152 *) Make maximum certificate chain size accepted from the peer application
5153 settable (SSL*_get/set_max_cert_list()), as proposed by
5154 "Douglas E. Engert" <deengert@anl.gov>.
5157 *) Add support for shared libraries for Unixware-7
5158 (Boyd Lynn Gerber <gerberb@zenez.com>).
5161 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
5162 be done prior to destruction. Use this to unload error strings from
5163 ENGINEs that load their own error strings. NB: This adds two new API
5164 functions to "get" and "set" this destroy handler in an ENGINE.
5167 *) Alter all existing ENGINE implementations (except "openssl" and
5168 "openbsd") to dynamically instantiate their own error strings. This
5169 makes them more flexible to be built both as statically-linked ENGINEs
5170 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
5171 Also, add stub code to each that makes building them as self-contained
5172 shared-libraries easier (see README.ENGINE).
5175 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
5176 implementations into applications that are completely implemented in
5177 self-contained shared-libraries. The "dynamic" ENGINE exposes control
5178 commands that can be used to configure what shared-library to load and
5179 to control aspects of the way it is handled. Also, made an update to
5180 the README.ENGINE file that brings its information up-to-date and
5181 provides some information and instructions on the "dynamic" ENGINE
5182 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
5185 *) Make it possible to unload ranges of ERR strings with a new
5186 "ERR_unload_strings" function.
5189 *) Add a copy() function to EVP_MD.
5192 *) Make EVP_MD routines take a context pointer instead of just the
5193 md_data void pointer.
5196 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
5197 that the digest can only process a single chunk of data
5198 (typically because it is provided by a piece of
5199 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
5200 is only going to provide a single chunk of data, and hence the
5201 framework needn't accumulate the data for oneshot drivers.
5204 *) As with "ERR", make it possible to replace the underlying "ex_data"
5205 functions. This change also alters the storage and management of global
5206 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
5207 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
5208 index counters. The API functions that use this state have been changed
5209 to take a "class_index" rather than pointers to the class's local STACK
5210 and counter, and there is now an API function to dynamically create new
5211 classes. This centralisation allows us to (a) plug a lot of the
5212 thread-safety problems that existed, and (b) makes it possible to clean
5213 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
5214 such data would previously have always leaked in application code and
5215 workarounds were in place to make the memory debugging turn a blind eye
5216 to it. Application code that doesn't use this new function will still
5217 leak as before, but their memory debugging output will announce it now
5218 rather than letting it slide.
5220 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
5221 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
5222 has a return value to indicate success or failure.
5225 *) Make it possible to replace the underlying "ERR" functions such that the
5226 global state (2 LHASH tables and 2 locks) is only used by the "default"
5227 implementation. This change also adds two functions to "get" and "set"
5228 the implementation prior to it being automatically set the first time
5229 any other ERR function takes place. Ie. an application can call "get",
5230 pass the return value to a module it has just loaded, and that module
5231 can call its own "set" function using that value. This means the
5232 module's "ERR" operations will use (and modify) the error state in the
5233 application and not in its own statically linked copy of OpenSSL code.
5236 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
5237 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
5238 the operation, and provides a more encapsulated way for external code
5239 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
5240 to use these functions rather than manually incrementing the counts.
5242 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
5245 *) Add EVP test program.
5248 *) Add symmetric cipher support to ENGINE. Expect the API to change!
5251 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
5252 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
5253 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
5254 These allow a CRL to be built without having to access X509_CRL fields
5255 directly. Modify 'ca' application to use new functions.
5258 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
5259 bug workarounds. Rollback attack detection is a security feature.
5260 The problem will only arise on OpenSSL servers when TLSv1 is not
5261 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
5262 Software authors not wanting to support TLSv1 will have special reasons
5263 for their choice and can explicitly enable this option.
5264 [Bodo Moeller, Lutz Jaenicke]
5266 *) Rationalise EVP so it can be extended: don't include a union of
5267 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
5268 (similar to those existing for EVP_CIPHER_CTX).
5273 EVP_MD_CTX_init(&md); /* new function call */
5274 EVP_DigestInit(&md, EVP_sha1());
5275 EVP_DigestUpdate(&md, in, len);
5276 EVP_DigestFinal(&md, out, NULL);
5277 EVP_MD_CTX_cleanup(&md); /* new function call */
5281 *) Make DES key schedule conform to the usual scheme, as well as
5282 correcting its structure. This means that calls to DES functions
5283 now have to pass a pointer to a des_key_schedule instead of a
5284 plain des_key_schedule (which was actually always a pointer
5287 des_key_schedule ks;
5289 des_set_key_checked(..., &ks);
5290 des_ncbc_encrypt(..., &ks, ...);
5292 (Note that a later change renames 'des_...' into 'DES_...'.)
5295 *) Initial reduction of linker bloat: the use of some functions, such as
5296 PEM causes large amounts of unused functions to be linked in due to
5297 poor organisation. For example pem_all.c contains every PEM function
5298 which has a knock on effect of linking in large amounts of (unused)
5299 ASN1 code. Grouping together similar functions and splitting unrelated
5300 functions prevents this.
5303 *) Cleanup of EVP macros.
5306 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
5307 correct _ecb suffix.
5310 *) Add initial OCSP responder support to ocsp application. The
5311 revocation information is handled using the text based index
5312 use by the ca application. The responder can either handle
5313 requests generated internally, supplied in files (for example
5314 via a CGI script) or using an internal minimal server.
5317 *) Add configuration choices to get zlib compression for TLS.
5320 *) Changes to Kerberos SSL for RFC 2712 compliance:
5321 1. Implemented real KerberosWrapper, instead of just using
5322 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
5323 2. Implemented optional authenticator field of KerberosWrapper.
5325 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
5326 and authenticator structs; see crypto/krb5/.
5328 Generalized Kerberos calls to support multiple Kerberos libraries.
5329 [Vern Staats <staatsvr@asc.hpc.mil>,
5330 Jeffrey Altman <jaltman@columbia.edu>
5331 via Richard Levitte]
5333 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
5334 already does with RSA. testdsa.h now has 'priv_key/pub_key'
5335 values for each of the key sizes rather than having just
5336 parameters (and 'speed' generating keys each time).
5339 *) Speed up EVP routines.
5342 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
5343 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
5344 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
5345 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
5347 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
5348 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
5349 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
5352 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
5354 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
5357 *) Added the OS2-EMX target.
5358 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
5360 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
5361 to support NCONF routines in extension code. New function CONF_set_nconf()
5362 to allow functions which take an NCONF to also handle the old LHASH
5363 structure: this means that the old CONF compatible routines can be
5364 retained (in particular wrt extensions) without having to duplicate the
5365 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
5368 *) Enhance the general user interface with mechanisms for inner control
5369 and with possibilities to have yes/no kind of prompts.
5372 *) Change all calls to low level digest routines in the library and
5373 applications to use EVP. Add missing calls to HMAC_cleanup() and
5374 don't assume HMAC_CTX can be copied using memcpy().
5375 [Verdon Walker <VWalker@novell.com>, Steve Henson]
5377 *) Add the possibility to control engines through control names but with
5378 arbitrary arguments instead of just a string.
5379 Change the key loaders to take a UI_METHOD instead of a callback
5380 function pointer. NOTE: this breaks binary compatibility with earlier
5381 versions of OpenSSL [engine].
5382 Adapt the nCipher code for these new conditions and add a card insertion
5386 *) Enhance the general user interface with mechanisms to better support
5387 dialog box interfaces, application-defined prompts, the possibility
5388 to use defaults (for example default passwords from somewhere else)
5389 and interrupts/cancellations.
5392 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
5393 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
5396 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
5397 tidy up some unnecessarily weird code in 'sk_new()').
5398 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
5400 *) Change the key loading routines for ENGINEs to use the same kind
5401 callback (pem_password_cb) as all other routines that need this
5405 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
5406 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
5407 than this minimum value is recommended.
5410 *) New random seeder for OpenVMS, using the system process statistics
5411 that are easily reachable.
5414 *) Windows apparently can't transparently handle global
5415 variables defined in DLLs. Initialisations such as:
5417 const ASN1_ITEM *it = &ASN1_INTEGER_it;
5419 wont compile. This is used by the any applications that need to
5420 declare their own ASN1 modules. This was fixed by adding the option
5421 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
5422 needed for static libraries under Win32.
5425 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
5426 setting of purpose and trust fields. New X509_STORE trust and
5427 purpose functions and tidy up setting in other SSL functions.
5430 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
5431 structure. These are inherited by X509_STORE_CTX when it is
5432 initialised. This allows various defaults to be set in the
5433 X509_STORE structure (such as flags for CRL checking and custom
5434 purpose or trust settings) for functions which only use X509_STORE_CTX
5435 internally such as S/MIME.
5437 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
5438 trust settings if they are not set in X509_STORE. This allows X509_STORE
5439 purposes and trust (in S/MIME for example) to override any set by default.
5441 Add command line options for CRL checking to smime, s_client and s_server
5445 *) Initial CRL based revocation checking. If the CRL checking flag(s)
5446 are set then the CRL is looked up in the X509_STORE structure and
5447 its validity and signature checked, then if the certificate is found
5448 in the CRL the verify fails with a revoked error.
5450 Various new CRL related callbacks added to X509_STORE_CTX structure.
5452 Command line options added to 'verify' application to support this.
5454 This needs some additional work, such as being able to handle multiple
5455 CRLs with different times, extension based lookup (rather than just
5456 by subject name) and ultimately more complete V2 CRL extension
5460 *) Add a general user interface API (crypto/ui/). This is designed
5461 to replace things like des_read_password and friends (backward
5462 compatibility functions using this new API are provided).
5463 The purpose is to remove prompting functions from the DES code
5464 section as well as provide for prompting through dialog boxes in
5465 a window system and the like.
5468 *) Add "ex_data" support to ENGINE so implementations can add state at a
5469 per-structure level rather than having to store it globally.
5472 *) Make it possible for ENGINE structures to be copied when retrieved by
5473 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
5474 This causes the "original" ENGINE structure to act like a template,
5475 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
5476 operational state can be localised to each ENGINE structure, despite the
5477 fact they all share the same "methods". New ENGINE structures returned in
5478 this case have no functional references and the return value is the single
5479 structural reference. This matches the single structural reference returned
5480 by ENGINE_by_id() normally, when it is incremented on the pre-existing
5484 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
5485 needs to match any other type at all we need to manually clear the
5489 *) Changes to the "openssl engine" utility to include;
5490 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
5491 about an ENGINE's available control commands.
5492 - executing control commands from command line arguments using the
5493 '-pre' and '-post' switches. '-post' is only used if '-t' is
5494 specified and the ENGINE is successfully initialised. The syntax for
5495 the individual commands are colon-separated, for example;
5496 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
5499 *) New dynamic control command support for ENGINEs. ENGINEs can now
5500 declare their own commands (numbers), names (strings), descriptions,
5501 and input types for run-time discovery by calling applications. A
5502 subset of these commands are implicitly classed as "executable"
5503 depending on their input type, and only these can be invoked through
5504 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
5505 can be based on user input, config files, etc). The distinction is
5506 that "executable" commands cannot return anything other than a boolean
5507 result and can only support numeric or string input, whereas some
5508 discoverable commands may only be for direct use through
5509 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
5510 pointers, or other custom uses. The "executable" commands are to
5511 support parameterisations of ENGINE behaviour that can be
5512 unambiguously defined by ENGINEs and used consistently across any
5513 OpenSSL-based application. Commands have been added to all the
5514 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
5515 control over shared-library paths without source code alterations.
5518 *) Changed all ENGINE implementations to dynamically allocate their
5519 ENGINEs rather than declaring them statically. Apart from this being
5520 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
5521 this also allows the implementations to compile without using the
5522 internal engine_int.h header.
5525 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
5526 'const' value. Any code that should be able to modify a RAND_METHOD
5527 should already have non-const pointers to it (ie. they should only
5528 modify their own ones).
5531 *) Made a variety of little tweaks to the ENGINE code.
5532 - "atalla" and "ubsec" string definitions were moved from header files
5533 to C code. "nuron" string definitions were placed in variables
5534 rather than hard-coded - allowing parameterisation of these values
5535 later on via ctrl() commands.
5536 - Removed unused "#if 0"'d code.
5537 - Fixed engine list iteration code so it uses ENGINE_free() to release
5538 structural references.
5539 - Constified the RAND_METHOD element of ENGINE structures.
5540 - Constified various get/set functions as appropriate and added
5541 missing functions (including a catch-all ENGINE_cpy that duplicates
5542 all ENGINE values onto a new ENGINE except reference counts/state).
5543 - Removed NULL parameter checks in get/set functions. Setting a method
5544 or function to NULL is a way of cancelling out a previously set
5545 value. Passing a NULL ENGINE parameter is just plain stupid anyway
5546 and doesn't justify the extra error symbols and code.
5547 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
5548 flags from engine_int.h to engine.h.
5549 - Changed prototypes for ENGINE handler functions (init(), finish(),
5550 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
5553 *) Implement binary inversion algorithm for BN_mod_inverse in addition
5554 to the algorithm using long division. The binary algorithm can be
5555 used only if the modulus is odd. On 32-bit systems, it is faster
5556 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
5557 roughly 5-15% for 256-bit moduli), so we use it only for moduli
5558 up to 450 bits. In 64-bit environments, the binary algorithm
5559 appears to be advantageous for much longer moduli; here we use it
5560 for moduli up to 2048 bits.
5563 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
5564 could not support the combine flag in choice fields.
5567 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
5568 extensions from a certificate request to the certificate.
5571 *) Allow multiple 'certopt' and 'nameopt' options to be separated
5572 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
5573 file: this allows the display of the certificate about to be
5574 signed to be customised, to allow certain fields to be included
5575 or excluded and extension details. The old system didn't display
5576 multicharacter strings properly, omitted fields not in the policy
5577 and couldn't display additional details such as extensions.
5580 *) Function EC_POINTs_mul for multiple scalar multiplication
5581 of an arbitrary number of elliptic curve points
5582 \sum scalars[i]*points[i],
5583 optionally including the generator defined for the EC_GROUP:
5584 scalar*generator + \sum scalars[i]*points[i].
5586 EC_POINT_mul is a simple wrapper function for the typical case
5587 that the point list has just one item (besides the optional
5591 *) First EC_METHODs for curves over GF(p):
5593 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
5594 operations and provides various method functions that can also
5595 operate with faster implementations of modular arithmetic.
5597 EC_GFp_mont_method() reuses most functions that are part of
5598 EC_GFp_simple_method, but uses Montgomery arithmetic.
5600 [Bodo Moeller; point addition and point doubling
5601 implementation directly derived from source code provided by
5602 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
5604 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
5605 crypto/ec/ec_lib.c):
5607 Curves are EC_GROUP objects (with an optional group generator)
5608 based on EC_METHODs that are built into the library.
5610 Points are EC_POINT objects based on EC_GROUP objects.
5612 Most of the framework would be able to handle curves over arbitrary
5613 finite fields, but as there are no obvious types for fields other
5614 than GF(p), some functions are limited to that for now.
5617 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
5618 that the file contains a complete HTTP response.
5621 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
5622 change the def and num file printf format specifier from "%-40sXXX"
5623 to "%-39s XXX". The latter will always guarantee a space after the
5624 field while the former will cause them to run together if the field
5625 is 40 of more characters long.
5628 *) Constify the cipher and digest 'method' functions and structures
5629 and modify related functions to take constant EVP_MD and EVP_CIPHER
5633 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
5634 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
5637 *) Modify EVP_Digest*() routines so they now return values. Although the
5638 internal software routines can never fail additional hardware versions
5642 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
5644 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
5645 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
5648 ERR_R_NESTED_ASN1_ERROR
5650 ERR_R_MISSING_ASN1_EOS
5651 were 4 .. 9, conflicting with
5652 ERR_LIB_RSA (= ERR_R_RSA_LIB)
5654 ERR_LIB_PEM (= ERR_R_PEM_LIB).
5655 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
5657 Add new error code 'ERR_R_INTERNAL_ERROR'.
5660 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
5664 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
5665 sets the subject name for a new request or supersedes the
5666 subject name in a given request. Formats that can be parsed are
5667 'CN=Some Name, OU=myOU, C=IT'
5669 'CN=Some Name/OU=myOU/C=IT'.
5671 Add options '-batch' and '-verbose' to 'openssl req'.
5672 [Massimiliano Pala <madwolf@hackmasters.net>]
5674 *) Introduce the possibility to access global variables through
5675 functions on platform were that's the best way to handle exporting
5676 global variables in shared libraries. To enable this functionality,
5677 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
5678 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
5679 is normally done by Configure or something similar).
5681 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
5682 in the source file (foo.c) like this:
5684 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
5685 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
5687 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
5688 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
5690 OPENSSL_DECLARE_GLOBAL(int,foo);
5691 #define foo OPENSSL_GLOBAL_REF(foo)
5692 OPENSSL_DECLARE_GLOBAL(double,bar);
5693 #define bar OPENSSL_GLOBAL_REF(bar)
5695 The #defines are very important, and therefore so is including the
5696 header file everywhere where the defined globals are used.
5698 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
5699 of ASN.1 items, but that structure is a bit different.
5701 The largest change is in util/mkdef.pl which has been enhanced with
5702 better and easier to understand logic to choose which symbols should
5703 go into the Windows .def files as well as a number of fixes and code
5704 cleanup (among others, algorithm keywords are now sorted
5705 lexicographically to avoid constant rewrites).
5708 *) In BN_div() keep a copy of the sign of 'num' before writing the
5709 result to 'rm' because if rm==num the value will be overwritten
5710 and produce the wrong result if 'num' is negative: this caused
5711 problems with BN_mod() and BN_nnmod().
5714 *) Function OCSP_request_verify(). This checks the signature on an
5715 OCSP request and verifies the signer certificate. The signer
5716 certificate is just checked for a generic purpose and OCSP request
5720 *) Add OCSP_check_validity() function to check the validity of OCSP
5721 responses. OCSP responses are prepared in real time and may only
5722 be a few seconds old. Simply checking that the current time lies
5723 between thisUpdate and nextUpdate max reject otherwise valid responses
5724 caused by either OCSP responder or client clock inaccuracy. Instead
5725 we allow thisUpdate and nextUpdate to fall within a certain period of
5726 the current time. The age of the response can also optionally be
5727 checked. Two new options -validity_period and -status_age added to
5731 *) If signature or public key algorithm is unrecognized print out its
5732 OID rather that just UNKNOWN.
5735 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
5736 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
5737 ID to be generated from the issuer certificate alone which can then be
5738 passed to OCSP_id_issuer_cmp().
5741 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
5742 ASN1 modules to export functions returning ASN1_ITEM pointers
5743 instead of the ASN1_ITEM structures themselves. This adds several
5744 new macros which allow the underlying ASN1 function/structure to
5745 be accessed transparently. As a result code should not use ASN1_ITEM
5746 references directly (such as &X509_it) but instead use the relevant
5747 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
5748 use of the new ASN1 code on platforms where exporting structures
5749 is problematical (for example in shared libraries) but exporting
5750 functions returning pointers to structures is not.
5753 *) Add support for overriding the generation of SSL/TLS session IDs.
5754 These callbacks can be registered either in an SSL_CTX or per SSL.
5755 The purpose of this is to allow applications to control, if they wish,
5756 the arbitrary values chosen for use as session IDs, particularly as it
5757 can be useful for session caching in multiple-server environments. A
5758 command-line switch for testing this (and any client code that wishes
5759 to use such a feature) has been added to "s_server".
5760 [Geoff Thorpe, Lutz Jaenicke]
5762 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
5763 of the form '#if defined(...) || defined(...) || ...' and
5764 '#if !defined(...) && !defined(...) && ...'. This also avoids
5765 the growing number of special cases it was previously handling.
5768 *) Make all configuration macros available for application by making
5769 sure they are available in opensslconf.h, by giving them names starting
5770 with "OPENSSL_" to avoid conflicts with other packages and by making
5771 sure e_os2.h will cover all platform-specific cases together with
5773 Additionally, it is now possible to define configuration/platform-
5774 specific names (called "system identities"). In the C code, these
5775 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
5776 macro with the name beginning with "OPENSSL_SYS_", which is determined
5777 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
5781 *) New option -set_serial to 'req' and 'x509' this allows the serial
5782 number to use to be specified on the command line. Previously self
5783 signed certificates were hard coded with serial number 0 and the
5784 CA options of 'x509' had to use a serial number in a file which was
5788 *) New options to 'ca' utility to support V2 CRL entry extensions.
5789 Currently CRL reason, invalidity date and hold instruction are
5790 supported. Add new CRL extensions to V3 code and some new objects.
5793 *) New function EVP_CIPHER_CTX_set_padding() this is used to
5794 disable standard block padding (aka PKCS#5 padding) in the EVP
5795 API, which was previously mandatory. This means that the data is
5796 not padded in any way and so the total length much be a multiple
5797 of the block size, otherwise an error occurs.
5800 *) Initial (incomplete) OCSP SSL support.
5803 *) New function OCSP_parse_url(). This splits up a URL into its host,
5804 port and path components: primarily to parse OCSP URLs. New -url
5805 option to ocsp utility.
5808 *) New nonce behavior. The return value of OCSP_check_nonce() now
5809 reflects the various checks performed. Applications can decide
5810 whether to tolerate certain situations such as an absent nonce
5811 in a response when one was present in a request: the ocsp application
5812 just prints out a warning. New function OCSP_add1_basic_nonce()
5813 this is to allow responders to include a nonce in a response even if
5814 the request is nonce-less.
5817 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
5818 skipped when using openssl x509 multiple times on a single input file,
5819 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
5822 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
5823 set string type: to handle setting ASN1_TIME structures. Fix ca
5824 utility to correctly initialize revocation date of CRLs.
5827 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
5828 the clients preferred ciphersuites and rather use its own preferences.
5829 Should help to work around M$ SGC (Server Gated Cryptography) bug in
5830 Internet Explorer by ensuring unchanged hash method during stepup.
5831 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
5834 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
5835 to aes and add a new 'exist' option to print out symbols that don't
5839 *) Additional options to ocsp utility to allow flags to be set and
5840 additional certificates supplied.
5843 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
5844 OCSP client a number of certificate to only verify the response
5848 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
5849 handle the new API. Currently only ECB, CBC modes supported. Add new
5852 Add TLS AES ciphersuites as described in RFC3268, "Advanced
5853 Encryption Standard (AES) Ciphersuites for Transport Layer
5854 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
5855 not enabled by default and were not part of the "ALL" ciphersuite
5856 alias because they were not yet official; they could be
5857 explicitly requested by specifying the "AESdraft" ciphersuite
5858 group alias. In the final release of OpenSSL 0.9.7, the group
5859 alias is called "AES" and is part of "ALL".)
5860 [Ben Laurie, Steve Henson, Bodo Moeller]
5862 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
5863 request to response.
5866 *) Functions for OCSP responders. OCSP_request_onereq_count(),
5867 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
5868 extract information from a certificate request. OCSP_response_create()
5869 creates a response and optionally adds a basic response structure.
5870 OCSP_basic_add1_status() adds a complete single response to a basic
5871 response and returns the OCSP_SINGLERESP structure just added (to allow
5872 extensions to be included for example). OCSP_basic_add1_cert() adds a
5873 certificate to a basic response and OCSP_basic_sign() signs a basic
5874 response with various flags. New helper functions ASN1_TIME_check()
5875 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
5876 (converts ASN1_TIME to GeneralizedTime).
5879 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
5880 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
5881 structure from a certificate. X509_pubkey_digest() digests the public_key
5882 contents: this is used in various key identifiers.
5885 *) Make sk_sort() tolerate a NULL argument.
5886 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
5888 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
5889 passed by the function are trusted implicitly. If any of them signed the
5890 response then it is assumed to be valid and is not verified.
5893 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
5894 to data. This was previously part of the PKCS7 ASN1 code. This
5895 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
5896 [Steve Henson, reported by Kenneth R. Robinette
5897 <support@securenetterm.com>]
5899 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
5900 routines: without these tracing memory leaks is very painful.
5901 Fix leaks in PKCS12 and PKCS7 routines.
5904 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
5905 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
5906 effectively meant GeneralizedTime would never be used. Now it
5907 is initialised to -1 but X509_time_adj() now has to check the value
5908 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
5909 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
5910 [Steve Henson, reported by Kenneth R. Robinette
5911 <support@securenetterm.com>]
5913 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
5914 result in a zero length in the ASN1_INTEGER structure which was
5915 not consistent with the structure when d2i_ASN1_INTEGER() was used
5916 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
5917 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
5918 where it did not print out a minus for negative ASN1_INTEGER.
5921 *) Add summary printout to ocsp utility. The various functions which
5922 convert status values to strings have been renamed to:
5923 OCSP_response_status_str(), OCSP_cert_status_str() and
5924 OCSP_crl_reason_str() and are no longer static. New options
5925 to verify nonce values and to disable verification. OCSP response
5926 printout format cleaned up.
5929 *) Add additional OCSP certificate checks. These are those specified
5930 in RFC2560. This consists of two separate checks: the CA of the
5931 certificate being checked must either be the OCSP signer certificate
5932 or the issuer of the OCSP signer certificate. In the latter case the
5933 OCSP signer certificate must contain the OCSP signing extended key
5934 usage. This check is performed by attempting to match the OCSP
5935 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
5936 in the OCSP_CERTID structures of the response.
5939 *) Initial OCSP certificate verification added to OCSP_basic_verify()
5940 and related routines. This uses the standard OpenSSL certificate
5941 verify routines to perform initial checks (just CA validity) and
5942 to obtain the certificate chain. Then additional checks will be
5943 performed on the chain. Currently the root CA is checked to see
5944 if it is explicitly trusted for OCSP signing. This is used to set
5945 a root CA as a global signing root: that is any certificate that
5946 chains to that CA is an acceptable OCSP signing certificate.
5949 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
5950 extensions from a separate configuration file.
5951 As when reading extensions from the main configuration file,
5952 the '-extensions ...' option may be used for specifying the
5954 [Massimiliano Pala <madwolf@comune.modena.it>]
5956 *) New OCSP utility. Allows OCSP requests to be generated or
5957 read. The request can be sent to a responder and the output
5958 parsed, outputed or printed in text form. Not complete yet:
5959 still needs to check the OCSP response validity.
5962 *) New subcommands for 'openssl ca':
5963 'openssl ca -status <serial>' prints the status of the cert with
5964 the given serial number (according to the index file).
5965 'openssl ca -updatedb' updates the expiry status of certificates
5967 [Massimiliano Pala <madwolf@comune.modena.it>]
5969 *) New '-newreq-nodes' command option to CA.pl. This is like
5970 '-newreq', but calls 'openssl req' with the '-nodes' option
5971 so that the resulting key is not encrypted.
5972 [Damien Miller <djm@mindrot.org>]
5974 *) New configuration for the GNU Hurd.
5975 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5977 *) Initial code to implement OCSP basic response verify. This
5978 is currently incomplete. Currently just finds the signer's
5979 certificate and verifies the signature on the response.
5982 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5983 value of OPENSSLDIR. This is available via the new '-d' option
5984 to 'openssl version', and is also included in 'openssl version -a'.
5987 *) Allowing defining memory allocation callbacks that will be given
5988 file name and line number information in additional arguments
5989 (a const char* and an int). The basic functionality remains, as
5990 well as the original possibility to just replace malloc(),
5991 realloc() and free() by functions that do not know about these
5992 additional arguments. To register and find out the current
5993 settings for extended allocation functions, the following
5994 functions are provided:
5996 CRYPTO_set_mem_ex_functions
5997 CRYPTO_set_locked_mem_ex_functions
5998 CRYPTO_get_mem_ex_functions
5999 CRYPTO_get_locked_mem_ex_functions
6001 These work the same way as CRYPTO_set_mem_functions and friends.
6002 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
6003 extended allocation function is enabled.
6004 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
6005 a conventional allocation function is enabled.
6006 [Richard Levitte, Bodo Moeller]
6008 *) Finish off removing the remaining LHASH function pointer casts.
6009 There should no longer be any prototype-casting required when using
6010 the LHASH abstraction, and any casts that remain are "bugs". See
6011 the callback types and macros at the head of lhash.h for details
6012 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
6015 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
6016 If /dev/[u]random devices are not available or do not return enough
6017 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
6019 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
6020 /etc/entropy will be queried once each in this sequence, quering stops
6021 when enough entropy was collected without querying more sockets.
6024 *) Change the Unix RAND_poll() variant to be able to poll several
6025 random devices, as specified by DEVRANDOM, until a sufficient amount
6026 of data has been collected. We spend at most 10 ms on each file
6027 (select timeout) and read in non-blocking mode. DEVRANDOM now
6028 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
6029 (previously it was just the string "/dev/urandom"), so on typical
6030 platforms the 10 ms delay will never occur.
6031 Also separate out the Unix variant to its own file, rand_unix.c.
6032 For VMS, there's a currently-empty rand_vms.c.
6035 *) Move OCSP client related routines to ocsp_cl.c. These
6036 provide utility functions which an application needing
6037 to issue a request to an OCSP responder and analyse the
6038 response will typically need: as opposed to those which an
6039 OCSP responder itself would need which will be added later.
6041 OCSP_request_sign() signs an OCSP request with an API similar
6042 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
6043 response. OCSP_response_get1_basic() extracts basic response
6044 from response. OCSP_resp_find_status(): finds and extracts status
6045 information from an OCSP_CERTID structure (which will be created
6046 when the request structure is built). These are built from lower
6047 level functions which work on OCSP_SINGLERESP structures but
6048 wont normally be used unless the application wishes to examine
6049 extensions in the OCSP response for example.
6051 Replace nonce routines with a pair of functions.
6052 OCSP_request_add1_nonce() adds a nonce value and optionally
6053 generates a random value. OCSP_check_nonce() checks the
6054 validity of the nonce in an OCSP response.
6057 *) Change function OCSP_request_add() to OCSP_request_add0_id().
6058 This doesn't copy the supplied OCSP_CERTID and avoids the
6059 need to free up the newly created id. Change return type
6060 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
6061 This can then be used to add extensions to the request.
6062 Deleted OCSP_request_new(), since most of its functionality
6063 is now in OCSP_REQUEST_new() (and the case insensitive name
6064 clash) apart from the ability to set the request name which
6065 will be added elsewhere.
6068 *) Update OCSP API. Remove obsolete extensions argument from
6069 various functions. Extensions are now handled using the new
6070 OCSP extension code. New simple OCSP HTTP function which
6071 can be used to send requests and parse the response.
6074 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
6075 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
6076 uses the special reorder version of SET OF to sort the attributes
6077 and reorder them to match the encoded order. This resolves a long
6078 standing problem: a verify on a PKCS7 structure just after signing
6079 it used to fail because the attribute order did not match the
6080 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
6081 it uses the received order. This is necessary to tolerate some broken
6082 software that does not order SET OF. This is handled by encoding
6083 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
6084 to produce the required SET OF.
6087 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
6088 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
6089 files to get correct declarations of the ASN.1 item variables.
6092 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
6093 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
6094 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
6095 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
6096 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
6097 ASN1_ITEM and no wrapper functions.
6100 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
6101 replace the old function pointer based I/O routines. Change most of
6102 the *_d2i_bio() and *_d2i_fp() functions to use these.
6105 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
6106 lines, recognice more "algorithms" that can be deselected, and make
6107 it complain about algorithm deselection that isn't recognised.
6110 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
6111 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
6112 to use new functions. Add NO_ASN1_OLD which can be set to remove
6113 some old style ASN1 functions: this can be used to determine if old
6114 code will still work when these eventually go away.
6117 *) New extension functions for OCSP structures, these follow the
6118 same conventions as certificates and CRLs.
6121 *) New function X509V3_add1_i2d(). This automatically encodes and
6122 adds an extension. Its behaviour can be customised with various
6123 flags to append, replace or delete. Various wrappers added for
6124 certifcates and CRLs.
6127 *) Fix to avoid calling the underlying ASN1 print routine when
6128 an extension cannot be parsed. Correct a typo in the
6129 OCSP_SERVICELOC extension. Tidy up print OCSP format.
6132 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
6133 entries for variables.
6136 *) Add functionality to apps/openssl.c for detecting locking
6137 problems: As the program is single-threaded, all we have
6138 to do is register a locking callback using an array for
6139 storing which locks are currently held by the program.
6142 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
6143 SSL_get_ex_data_X509_STORE_idx(), which is used in
6144 ssl_verify_cert_chain() and thus can be called at any time
6145 during TLS/SSL handshakes so that thread-safety is essential.
6146 Unfortunately, the ex_data design is not at all suited
6147 for multi-threaded use, so it probably should be abolished.
6150 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
6151 [Broadcom, tweaked and integrated by Geoff Thorpe]
6153 *) Move common extension printing code to new function
6154 X509V3_print_extensions(). Reorganise OCSP print routines and
6155 implement some needed OCSP ASN1 functions. Add OCSP extensions.
6158 *) New function X509_signature_print() to remove duplication in some
6162 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
6163 set (this was treated exactly the same as SET OF previously). This
6164 is used to reorder the STACK representing the structure to match the
6165 encoding. This will be used to get round a problem where a PKCS7
6166 structure which was signed could not be verified because the STACK
6167 order did not reflect the encoded order.
6170 *) Reimplement the OCSP ASN1 module using the new code.
6173 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
6174 for its ASN1 operations. The old style function pointers still exist
6175 for now but they will eventually go away.
6178 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
6179 completely replaces the old ASN1 functionality with a table driven
6180 encoder and decoder which interprets an ASN1_ITEM structure describing
6181 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
6182 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
6183 has also been converted to the new form.
6186 *) Change BN_mod_exp_recp so that negative moduli are tolerated
6187 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
6188 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
6189 for negative moduli.
6192 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
6193 of not touching the result's sign bit.
6196 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
6200 *) Changed the LHASH code to use prototypes for callbacks, and created
6201 macros to declare and implement thin (optionally static) functions
6202 that provide type-safety and avoid function pointer casting for the
6203 type-specific callbacks.
6206 *) Added Kerberos Cipher Suites to be used with TLS, as written in
6208 [Veers Staats <staatsvr@asc.hpc.mil>,
6209 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
6211 *) Reformat the FAQ so the different questions and answers can be divided
6212 in sections depending on the subject.
6215 *) Have the zlib compression code load ZLIB.DLL dynamically under
6219 *) New function BN_mod_sqrt for computing square roots modulo a prime
6220 (using the probabilistic Tonelli-Shanks algorithm unless
6221 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
6222 be handled deterministically).
6223 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6225 *) Make BN_mod_inverse faster by explicitly handling small quotients
6226 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
6227 512 bits], about 30% for larger ones [1024 or 2048 bits].)
6230 *) New function BN_kronecker.
6233 *) Fix BN_gcd so that it works on negative inputs; the result is
6234 positive unless both parameters are zero.
6235 Previously something reasonably close to an infinite loop was
6236 possible because numbers could be growing instead of shrinking
6237 in the implementation of Euclid's algorithm.
6240 *) Fix BN_is_word() and BN_is_one() macros to take into account the
6241 sign of the number in question.
6243 Fix BN_is_word(a,w) to work correctly for w == 0.
6245 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
6246 because its test if the absolute value of 'a' equals 'w'.
6247 Note that BN_abs_is_word does *not* handle w == 0 reliably;
6248 it exists mostly for use in the implementations of BN_is_zero(),
6249 BN_is_one(), and BN_is_word().
6252 *) New function BN_swap.
6255 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
6256 the exponentiation functions are more likely to produce reasonable
6257 results on negative inputs.
6260 *) Change BN_mod_mul so that the result is always non-negative.
6261 Previously, it could be negative if one of the factors was negative;
6262 I don't think anyone really wanted that behaviour.
6265 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
6266 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
6267 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
6268 and add new functions:
6277 BN_mod_lshift1_quick
6281 These functions always generate non-negative results.
6283 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
6284 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
6286 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
6287 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
6288 be reduced modulo m.
6289 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6292 The following entry accidentily appeared in the CHANGES file
6293 distributed with OpenSSL 0.9.7. The modifications described in
6294 it do *not* apply to OpenSSL 0.9.7.
6296 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6297 was actually never needed) and in BN_mul(). The removal in BN_mul()
6298 required a small change in bn_mul_part_recursive() and the addition
6299 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6300 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6301 bn_sub_words() and bn_add_words() except they take arrays with
6306 *) In 'openssl passwd', verify passwords read from the terminal
6307 unless the '-salt' option is used (which usually means that
6308 verification would just waste user's time since the resulting
6309 hash is going to be compared with some given password hash)
6310 or the new '-noverify' option is used.
6312 This is an incompatible change, but it does not affect
6313 non-interactive use of 'openssl passwd' (passwords on the command
6314 line, '-stdin' option, '-in ...' option) and thus should not
6318 *) Remove all references to RSAref, since there's no more need for it.
6321 *) Make DSO load along a path given through an environment variable
6322 (SHLIB_PATH) with shl_load().
6325 *) Constify the ENGINE code as a result of BIGNUM constification.
6326 Also constify the RSA code and most things related to it. In a
6327 few places, most notable in the depth of the ASN.1 code, ugly
6328 casts back to non-const were required (to be solved at a later
6332 *) Make it so the openssl application has all engines loaded by default.
6335 *) Constify the BIGNUM routines a little more.
6338 *) Add the following functions:
6340 ENGINE_load_cswift()
6342 ENGINE_load_atalla()
6344 ENGINE_load_builtin_engines()
6346 That way, an application can itself choose if external engines that
6347 are built-in in OpenSSL shall ever be used or not. The benefit is
6348 that applications won't have to be linked with libdl or other dso
6349 libraries unless it's really needed.
6351 Changed 'openssl engine' to load all engines on demand.
6352 Changed the engine header files to avoid the duplication of some
6353 declarations (they differed!).
6356 *) 'openssl engine' can now list capabilities.
6359 *) Better error reporting in 'openssl engine'.
6362 *) Never call load_dh_param(NULL) in s_server.
6365 *) Add engine application. It can currently list engines by name and
6366 identity, and test if they are actually available.
6369 *) Improve RPM specification file by forcing symbolic linking and making
6370 sure the installed documentation is also owned by root.root.
6371 [Damien Miller <djm@mindrot.org>]
6373 *) Give the OpenSSL applications more possibilities to make use of
6374 keys (public as well as private) handled by engines.
6377 *) Add OCSP code that comes from CertCo.
6380 *) Add VMS support for the Rijndael code.
6383 *) Added untested support for Nuron crypto accelerator.
6386 *) Add support for external cryptographic devices. This code was
6387 previously distributed separately as the "engine" branch.
6388 [Geoff Thorpe, Richard Levitte]
6390 *) Rework the filename-translation in the DSO code. It is now possible to
6391 have far greater control over how a "name" is turned into a filename
6392 depending on the operating environment and any oddities about the
6393 different shared library filenames on each system.
6396 *) Support threads on FreeBSD-elf in Configure.
6399 *) Fix for SHA1 assembly problem with MASM: it produces
6400 warnings about corrupt line number information when assembling
6401 with debugging information. This is caused by the overlapping
6403 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
6406 NCONF_get_number() has no error checking at all. As a replacement,
6407 NCONF_get_number_e() is defined (_e for "error checking") and is
6408 promoted strongly. The old NCONF_get_number is kept around for
6409 binary backward compatibility.
6410 Make it possible for methods to load from something other than a BIO,
6411 by providing a function pointer that is given a name instead of a BIO.
6412 For example, this could be used to load configuration data from an
6416 *) Fix for non blocking accept BIOs. Added new I/O special reason
6417 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
6418 with non blocking I/O was not possible because no retry code was
6419 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
6423 *) Added the beginnings of Rijndael support.
6426 *) Fix for bug in DirectoryString mask setting. Add support for
6427 X509_NAME_print_ex() in 'req' and X509_print_ex() function
6428 to allow certificate printing to more controllable, additional
6429 'certopt' option to 'x509' to allow new printing options to be
6433 *) Clean old EAY MD5 hack from e_os.h.
6436 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
6438 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6439 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6440 [Joe Orton, Steve Henson]
6442 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
6444 *) Fix additional bug revealed by the NISCC test suite:
6446 Stop bug triggering large recursion when presented with
6447 certain ASN.1 tags (CVE-2003-0851)
6450 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
6452 *) Fix various bugs revealed by running the NISCC test suite:
6454 Stop out of bounds reads in the ASN1 code when presented with
6455 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6457 If verify callback ignores invalid public key errors don't try to check
6458 certificate signature with the NULL public key.
6462 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6463 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6467 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6468 extra data after the compression methods not only for TLS 1.0
6469 but also for SSL 3.0 (as required by the specification).
6470 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6472 *) Change X509_certificate_type() to mark the key as exported/exportable
6473 when it's 512 *bits* long, not 512 bytes.
6476 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
6478 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6479 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6480 a protocol version number mismatch like a decryption error
6481 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6484 *) Turn on RSA blinding by default in the default implementation
6485 to avoid a timing attack. Applications that don't want it can call
6486 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6487 They would be ill-advised to do so in most cases.
6488 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6490 *) Change RSA blinding code so that it works when the PRNG is not
6491 seeded (in this case, the secret RSA exponent is abused as
6492 an unpredictable seed -- if it is not unpredictable, there
6493 is no point in blinding anyway). Make RSA blinding thread-safe
6494 by remembering the creator's thread ID in rsa->blinding and
6495 having all other threads use local one-time blinding factors
6496 (this requires more computation than sharing rsa->blinding, but
6497 avoids excessive locking; and if an RSA object is not shared
6498 between threads, blinding will still be very fast).
6501 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
6503 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6504 via timing by performing a MAC computation even if incorrrect
6505 block cipher padding has been found. This is a countermeasure
6506 against active attacks where the attacker has to distinguish
6507 between bad padding and a MAC verification error. (CVE-2003-0078)
6509 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6510 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6511 Martin Vuagnoux (EPFL, Ilion)]
6513 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
6515 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
6516 memory from it's contents. This is done with a counter that will
6517 place alternating values in each byte. This can be used to solve
6518 two issues: 1) the removal of calls to memset() by highly optimizing
6519 compilers, and 2) cleansing with other values than 0, since those can
6520 be read through on certain media, for example a swap space on disk.
6523 *) Bugfix: client side session caching did not work with external caching,
6524 because the session->cipher setting was not restored when reloading
6525 from the external cache. This problem was masked, when
6526 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
6527 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
6530 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
6531 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
6532 [Zeev Lieber <zeev-l@yahoo.com>]
6534 *) Undo an undocumented change introduced in 0.9.6e which caused
6535 repeated calls to OpenSSL_add_all_ciphers() and
6536 OpenSSL_add_all_digests() to be ignored, even after calling
6540 *) Change the default configuration reader to deal with last line not
6541 being properly terminated.
6544 *) Change X509_NAME_cmp() so it applies the special rules on handling
6545 DN values that are of type PrintableString, as well as RDNs of type
6546 emailAddress where the value has the type ia5String.
6547 [stefank@valicert.com via Richard Levitte]
6549 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
6550 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
6551 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
6552 the bitwise-OR of the two for use by the majority of applications
6553 wanting this behaviour, and update the docs. The documented
6554 behaviour and actual behaviour were inconsistent and had been
6555 changing anyway, so this is more a bug-fix than a behavioural
6557 [Geoff Thorpe, diagnosed by Nadav Har'El]
6559 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
6560 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
6563 *) Fix initialization code race conditions in
6564 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
6565 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
6566 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
6567 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
6568 ssl2_get_cipher_by_char(),
6569 ssl3_get_cipher_by_char().
6570 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
6572 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
6573 the cached sessions are flushed, as the remove_cb() might use ex_data
6574 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
6575 (see [openssl.org #212]).
6576 [Geoff Thorpe, Lutz Jaenicke]
6578 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
6579 length, instead of the encoding length to d2i_ASN1_OBJECT.
6582 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
6584 *) [In 0.9.6g-engine release:]
6585 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
6586 [Lynn Gazis <lgazis@rainbow.com>]
6588 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
6590 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
6591 and get fix the header length calculation.
6592 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
6593 Alon Kantor <alonk@checkpoint.com> (and others),
6596 *) Use proper error handling instead of 'assertions' in buffer
6597 overflow checks added in 0.9.6e. This prevents DoS (the
6598 assertions could call abort()).
6599 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
6601 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
6603 *) Add various sanity checks to asn1_get_length() to reject
6604 the ASN1 length bytes if they exceed sizeof(long), will appear
6605 negative or the content length exceeds the length of the
6607 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6609 *) Fix cipher selection routines: ciphers without encryption had no flags
6610 for the cipher strength set and where therefore not handled correctly
6611 by the selection routines (PR #130).
6614 *) Fix EVP_dsa_sha macro.
6618 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
6619 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
6620 that was added in OpenSSL 0.9.6d.
6622 As the countermeasure turned out to be incompatible with some
6623 broken SSL implementations, the new option is part of SSL_OP_ALL.
6624 SSL_OP_ALL is usually employed when compatibility with weird SSL
6625 implementations is desired (e.g. '-bugs' option to 's_client' and
6626 's_server'), so the new option is automatically set in many
6630 *) Changes in security patch:
6632 Changes marked "(CHATS)" were sponsored by the Defense Advanced
6633 Research Projects Agency (DARPA) and Air Force Research Laboratory,
6634 Air Force Materiel Command, USAF, under agreement number
6637 *) Add various sanity checks to asn1_get_length() to reject
6638 the ASN1 length bytes if they exceed sizeof(long), will appear
6639 negative or the content length exceeds the length of the
6640 supplied buffer. (CVE-2002-0659)
6641 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6643 *) Assertions for various potential buffer overflows, not known to
6645 [Ben Laurie (CHATS)]
6647 *) Various temporary buffers to hold ASCII versions of integers were
6648 too small for 64 bit platforms. (CVE-2002-0655)
6649 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
6651 *) Remote buffer overflow in SSL3 protocol - an attacker could
6652 supply an oversized session ID to a client. (CVE-2002-0656)
6653 [Ben Laurie (CHATS)]
6655 *) Remote buffer overflow in SSL2 protocol - an attacker could
6656 supply an oversized client master key. (CVE-2002-0656)
6657 [Ben Laurie (CHATS)]
6659 Changes between 0.9.6c and 0.9.6d [9 May 2002]
6661 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
6662 encoded as NULL) with id-dsa-with-sha1.
6663 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
6665 *) Check various X509_...() return values in apps/req.c.
6666 [Nils Larsch <nla@trustcenter.de>]
6668 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
6669 an end-of-file condition would erronously be flagged, when the CRLF
6670 was just at the end of a processed block. The bug was discovered when
6671 processing data through a buffering memory BIO handing the data to a
6672 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
6673 <ptsekov@syntrex.com> and Nedelcho Stanev.
6676 *) Implement a countermeasure against a vulnerability recently found
6677 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
6678 before application data chunks to avoid the use of known IVs
6679 with data potentially chosen by the attacker.
6682 *) Fix length checks in ssl3_get_client_hello().
6685 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
6686 to prevent ssl3_read_internal() from incorrectly assuming that
6687 ssl3_read_bytes() found application data while handshake
6688 processing was enabled when in fact s->s3->in_read_app_data was
6689 merely automatically cleared during the initial handshake.
6690 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
6692 *) Fix object definitions for Private and Enterprise: they were not
6693 recognized in their shortname (=lowercase) representation. Extend
6694 obj_dat.pl to issue an error when using undefined keywords instead
6695 of silently ignoring the problem (Svenning Sorensen
6696 <sss@sss.dnsalias.net>).
6699 *) Fix DH_generate_parameters() so that it works for 'non-standard'
6700 generators, i.e. generators other than 2 and 5. (Previously, the
6701 code did not properly initialise the 'add' and 'rem' values to
6702 BN_generate_prime().)
6704 In the new general case, we do not insist that 'generator' is
6705 actually a primitive root: This requirement is rather pointless;
6706 a generator of the order-q subgroup is just as good, if not
6710 *) Map new X509 verification errors to alerts. Discovered and submitted by
6711 Tom Wu <tom@arcot.com>.
6714 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
6715 returning non-zero before the data has been completely received
6716 when using non-blocking I/O.
6717 [Bodo Moeller; problem pointed out by John Hughes]
6719 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
6720 [Ben Laurie, Lutz Jaenicke]
6722 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
6723 Yoram Zahavi <YoramZ@gilian.com>).
6726 *) Add information about CygWin 1.3 and on, and preserve proper
6727 configuration for the versions before that.
6728 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6730 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
6731 check whether we deal with a copy of a session and do not delete from
6732 the cache in this case. Problem reported by "Izhar Shoshani Levi"
6733 <izhar@checkpoint.com>.
6736 *) Do not store session data into the internal session cache, if it
6737 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
6738 flag is set). Proposed by Aslam <aslam@funk.com>.
6741 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
6745 *) [In 0.9.6d-engine release:]
6746 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6747 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6749 *) Add the configuration target linux-s390x.
6750 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
6752 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
6753 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
6754 variable as an indication that a ClientHello message has been
6755 received. As the flag value will be lost between multiple
6756 invocations of ssl3_accept when using non-blocking I/O, the
6757 function may not be aware that a handshake has actually taken
6758 place, thus preventing a new session from being added to the
6761 To avoid this problem, we now set s->new_session to 2 instead of
6762 using a local variable.
6763 [Lutz Jaenicke, Bodo Moeller]
6765 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
6766 if the SSL_R_LENGTH_MISMATCH error is detected.
6767 [Geoff Thorpe, Bodo Moeller]
6769 *) New 'shared_ldflag' column in Configure platform table.
6772 *) Fix EVP_CIPHER_mode macro.
6773 ["Dan S. Camper" <dan@bti.net>]
6775 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
6776 type, we must throw them away by setting rr->length to 0.
6777 [D P Chang <dpc@qualys.com>]
6779 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
6781 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
6782 <Dominikus.Scherkl@biodata.com>. (The previous implementation
6783 worked incorrectly for those cases where range = 10..._2 and
6784 3*range is two bits longer than range.)
6787 *) Only add signing time to PKCS7 structures if it is not already
6791 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
6792 OBJ_ld_ce should be OBJ_id_ce.
6793 Also some ip-pda OIDs in crypto/objects/objects.txt were
6794 incorrect (cf. RFC 3039).
6795 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
6797 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
6798 returns early because it has nothing to do.
6799 [Andy Schneider <andy.schneider@bjss.co.uk>]
6801 *) [In 0.9.6c-engine release:]
6802 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
6803 [Andy Schneider <andy.schneider@bjss.co.uk>]
6805 *) [In 0.9.6c-engine release:]
6806 Add support for Cryptographic Appliance's keyserver technology.
6807 (Use engine 'keyclient')
6808 [Cryptographic Appliances and Geoff Thorpe]
6810 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
6811 is called via tools/c89.sh because arguments have to be
6812 rearranged (all '-L' options must appear before the first object
6814 [Richard Shapiro <rshapiro@abinitio.com>]
6816 *) [In 0.9.6c-engine release:]
6817 Add support for Broadcom crypto accelerator cards, backported
6819 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
6821 *) [In 0.9.6c-engine release:]
6822 Add support for SureWare crypto accelerator cards from
6823 Baltimore Technologies. (Use engine 'sureware')
6824 [Baltimore Technologies and Mark Cox]
6826 *) [In 0.9.6c-engine release:]
6827 Add support for crypto accelerator cards from Accelerated
6828 Encryption Processing, www.aep.ie. (Use engine 'aep')
6829 [AEP Inc. and Mark Cox]
6831 *) Add a configuration entry for gcc on UnixWare.
6832 [Gary Benson <gbenson@redhat.com>]
6834 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
6835 messages are stored in a single piece (fixed-length part and
6836 variable-length part combined) and fix various bugs found on the way.
6839 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
6840 instead. BIO_gethostbyname() does not know what timeouts are
6841 appropriate, so entries would stay in cache even when they have
6843 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
6845 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
6846 faced with a pathologically small ClientHello fragment that does
6847 not contain client_version: Instead of aborting with an error,
6848 simply choose the highest available protocol version (i.e.,
6849 TLS 1.0 unless it is disabled). In practice, ClientHello
6850 messages are never sent like this, but this change gives us
6851 strictly correct behaviour at least for TLS.
6854 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
6855 never resets s->method to s->ctx->method when called from within
6856 one of the SSL handshake functions.
6857 [Bodo Moeller; problem pointed out by Niko Baric]
6859 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
6860 (sent using the client's version number) if client_version is
6861 smaller than the protocol version in use. Also change
6862 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
6863 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
6864 the client will at least see that alert.
6867 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
6871 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
6872 client receives HelloRequest while in a handshake.
6873 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
6875 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
6876 should end in 'break', not 'goto end' which circuments various
6877 cleanups done in state SSL_ST_OK. But session related stuff
6878 must be disabled for SSL_ST_OK in the case that we just sent a
6881 Also avoid some overhead by not calling ssl_init_wbio_buffer()
6882 before just sending a HelloRequest.
6883 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
6885 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
6886 reveal whether illegal block cipher padding was found or a MAC
6887 verification error occured. (Neither SSLerr() codes nor alerts
6888 are directly visible to potential attackers, but the information
6889 may leak via logfiles.)
6891 Similar changes are not required for the SSL 2.0 implementation
6892 because the number of padding bytes is sent in clear for SSL 2.0,
6893 and the extra bytes are just ignored. However ssl/s2_pkt.c
6894 failed to verify that the purported number of padding bytes is in
6898 *) Add OpenUNIX-8 support including shared libraries
6899 (Boyd Lynn Gerber <gerberb@zenez.com>).
6902 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
6903 'wristwatch attack' using huge encoding parameters (cf.
6904 James H. Manger's CRYPTO 2001 paper). Note that the
6905 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
6906 encoding parameters and hence was not vulnerable.
6909 *) BN_sqr() bug fix.
6910 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
6912 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
6913 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
6914 followed by modular reduction.
6915 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
6917 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
6918 equivalent based on BN_pseudo_rand() instead of BN_rand().
6921 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
6922 This function was broken, as the check for a new client hello message
6923 to handle SGC did not allow these large messages.
6924 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
6927 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
6930 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
6931 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
6934 *) Rework the configuration and shared library support for Tru64 Unix.
6935 The configuration part makes use of modern compiler features and
6936 still retains old compiler behavior for those that run older versions
6937 of the OS. The shared library support part includes a variant that
6938 uses the RPATH feature, and is available through the special
6939 configuration target "alpha-cc-rpath", which will never be selected
6941 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
6943 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
6944 with the same message size as in ssl3_get_certificate_request().
6945 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
6946 messages might inadvertently be reject as too long.
6947 [Petr Lampa <lampa@fee.vutbr.cz>]
6949 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
6952 *) Modified SSL library such that the verify_callback that has been set
6953 specificly for an SSL object with SSL_set_verify() is actually being
6954 used. Before the change, a verify_callback set with this function was
6955 ignored and the verify_callback() set in the SSL_CTX at the time of
6956 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
6957 to allow the necessary settings.
6960 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
6961 explicitly to NULL, as at least on Solaris 8 this seems not always to be
6962 done automatically (in contradiction to the requirements of the C
6963 standard). This made problems when used from OpenSSH.
6966 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
6967 dh->length and always used
6969 BN_rand_range(priv_key, dh->p).
6971 BN_rand_range() is not necessary for Diffie-Hellman, and this
6972 specific range makes Diffie-Hellman unnecessarily inefficient if
6973 dh->length (recommended exponent length) is much smaller than the
6974 length of dh->p. We could use BN_rand_range() if the order of
6975 the subgroup was stored in the DH structure, but we only have
6980 BN_rand(priv_key, l, ...)
6982 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6988 RSA_eay_public_encrypt
6989 RSA_eay_private_decrypt
6990 RSA_eay_private_encrypt (signing)
6991 RSA_eay_public_decrypt (signature verification)
6993 (default implementations for RSA_public_encrypt,
6994 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6995 always reject numbers >= n.
6998 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6999 to synchronize access to 'locking_thread'. This is necessary on
7000 systems where access to 'locking_thread' (an 'unsigned long'
7001 variable) is not atomic.
7004 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
7005 *before* setting the 'crypto_lock_rand' flag. The previous code had
7006 a race condition if 0 is a valid thread ID.
7007 [Travis Vitek <vitek@roguewave.com>]
7009 *) Add support for shared libraries under Irix.
7010 [Albert Chin-A-Young <china@thewrittenword.com>]
7012 *) Add configuration option to build on Linux on both big-endian and
7014 [Ralf Baechle <ralf@uni-koblenz.de>]
7016 *) Add the possibility to create shared libraries on HP-UX.
7019 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
7021 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
7022 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
7023 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
7024 PRNG state recovery was possible based on the output of
7025 one PRNG request appropriately sized to gain knowledge on
7026 'md' followed by enough consecutive 1-byte PRNG requests
7027 to traverse all of 'state'.
7029 1. When updating 'md_local' (the current thread's copy of 'md')
7030 during PRNG output generation, hash all of the previous
7031 'md_local' value, not just the half used for PRNG output.
7033 2. Make the number of bytes from 'state' included into the hash
7034 independent from the number of PRNG bytes requested.
7036 The first measure alone would be sufficient to avoid
7037 Markku-Juhani's attack. (Actually it had never occurred
7038 to me that the half of 'md_local' used for chaining was the
7039 half from which PRNG output bytes were taken -- I had always
7040 assumed that the secret half would be used.) The second
7041 measure makes sure that additional data from 'state' is never
7042 mixed into 'md_local' in small portions; this heuristically
7043 further strengthens the PRNG.
7046 *) Fix crypto/bn/asm/mips3.s.
7049 *) When only the key is given to "enc", the IV is undefined. Print out
7050 an error message in this case.
7053 *) Handle special case when X509_NAME is empty in X509 printing routines.
7056 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
7057 positive and less than q.
7060 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
7061 used: it isn't thread safe and the add_lock_callback should handle
7063 [Paul Rose <Paul.Rose@bridge.com>]
7065 *) Verify that incoming data obeys the block size in
7066 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
7070 [Ulf Möller, Bodo Möller]
7072 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
7073 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
7074 when fixing the server behaviour for backwards-compatible 'client
7075 hello' messages. (Note that the attack is impractical against
7076 SSL 3.0 and TLS 1.0 anyway because length and version checking
7077 means that the probability of guessing a valid ciphertext is
7078 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
7081 Before 0.9.5, the countermeasure (hide the error by generating a
7082 random 'decryption result') did not work properly because
7083 ERR_clear_error() was missing, meaning that SSL_get_error() would
7084 detect the supposedly ignored error.
7086 Both problems are now fixed.
7089 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
7090 (previously it was 1024).
7093 *) Fix for compatibility mode trust settings: ignore trust settings
7094 unless some valid trust or reject settings are present.
7097 *) Fix for blowfish EVP: its a variable length cipher.
7100 *) Fix various bugs related to DSA S/MIME verification. Handle missing
7101 parameters in DSA public key structures and return an error in the
7102 DSA routines if parameters are absent.
7105 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
7106 in the current directory if neither $RANDFILE nor $HOME was set.
7107 RAND_file_name() in 0.9.6a returned NULL in this case. This has
7108 caused some confusion to Windows users who haven't defined $HOME.
7109 Thus RAND_file_name() is changed again: e_os.h can define a
7110 DEFAULT_HOME, which will be used if $HOME is not set.
7111 For Windows, we use "C:"; on other platforms, we still require
7112 environment variables.
7114 *) Move 'if (!initialized) RAND_poll()' into regions protected by
7115 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
7116 having multiple threads call RAND_poll() concurrently.
7119 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
7120 combination of a flag and a thread ID variable.
7121 Otherwise while one thread is in ssleay_rand_bytes (which sets the
7122 flag), *other* threads can enter ssleay_add_bytes without obeying
7123 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
7124 that they do not hold after the first thread unsets add_do_not_lock).
7127 *) Change bctest again: '-x' expressions are not available in all
7131 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
7133 *) Fix a couple of memory leaks in PKCS7_dataDecode()
7134 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
7136 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
7137 the default extension for executables, if any. Also, make the perl
7138 scripts that use symlink() to test if it really exists and use "cp"
7139 if it doesn't. All this made OpenSSL compilable and installable in
7143 *) Fix for asn1_GetSequence() for indefinite length constructed data.
7144 If SEQUENCE is length is indefinite just set c->slen to the total
7145 amount of data available.
7146 [Steve Henson, reported by shige@FreeBSD.org]
7147 [This change does not apply to 0.9.7.]
7149 *) Change bctest to avoid here-documents inside command substitution
7150 (workaround for FreeBSD /bin/sh bug).
7151 For compatibility with Ultrix, avoid shell functions (introduced
7152 in the bctest version that searches along $PATH).
7155 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
7156 with des_encrypt() defined on some operating systems, like Solaris
7160 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
7161 On the Importance of Eliminating Errors in Cryptographic
7162 Computations, J. Cryptology 14 (2001) 2, 101-119,
7163 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
7166 *) MIPS assembler BIGNUM division bug fix.
7169 *) Disabled incorrect Alpha assembler code.
7172 *) Fix PKCS#7 decode routines so they correctly update the length
7173 after reading an EOC for the EXPLICIT tag.
7175 [This change does not apply to 0.9.7.]
7177 *) Fix bug in PKCS#12 key generation routines. This was triggered
7178 if a 3DES key was generated with a 0 initial byte. Include
7179 PKCS12_BROKEN_KEYGEN compilation option to retain the old
7180 (but broken) behaviour.
7183 *) Enhance bctest to search for a working bc along $PATH and print
7185 [Tim Rice <tim@multitalents.net> via Richard Levitte]
7187 *) Fix memory leaks in err.c: free err_data string if necessary;
7188 don't write to the wrong index in ERR_set_error_data.
7191 *) Implement ssl23_peek (analogous to ssl23_read), which previously
7195 *) Replace rdtsc with _emit statements for VC++ version 5.
7196 [Jeremy Cooper <jeremy@baymoo.org>]
7198 *) Make it possible to reuse SSLv2 sessions.
7201 *) In copy_email() check for >= 0 as a return value for
7202 X509_NAME_get_index_by_NID() since 0 is a valid index.
7203 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
7205 *) Avoid coredump with unsupported or invalid public keys by checking if
7206 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
7207 PKCS7_verify() fails with non detached data.
7210 *) Don't use getenv in library functions when run as setuid/setgid.
7211 New function OPENSSL_issetugid().
7214 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7215 due to incorrect handling of multi-threading:
7217 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
7219 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
7221 3. Count how many times MemCheck_off() has been called so that
7222 nested use can be treated correctly. This also avoids
7223 inband-signalling in the previous code (which relied on the
7224 assumption that thread ID 0 is impossible).
7227 *) Add "-rand" option also to s_client and s_server.
7230 *) Fix CPU detection on Irix 6.x.
7231 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
7232 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7234 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
7237 [This change does not apply to 0.9.7.]
7239 *) Use the cached encoding of an X509_NAME structure rather than
7240 copying it. This is apparently the reason for the libsafe "errors"
7241 but the code is actually correct.
7244 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
7245 Bleichenbacher's DSA attack.
7246 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
7247 to be set and top=0 forces the highest bit to be set; top=-1 is new
7248 and leaves the highest bit random.
7249 [Ulf Moeller, Bodo Moeller]
7251 *) In the NCONF_...-based implementations for CONF_... queries
7252 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
7253 a temporary CONF structure with the data component set to NULL
7254 (which gives segmentation faults in lh_retrieve).
7255 Instead, use NULL for the CONF pointer in CONF_get_string and
7256 CONF_get_number (which may use environment variables) and directly
7257 return NULL from CONF_get_section.
7260 *) Fix potential buffer overrun for EBCDIC.
7263 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7264 keyUsage if basicConstraints absent for a CA.
7267 *) Make SMIME_write_PKCS7() write mail header values with a format that
7268 is more generally accepted (no spaces before the semicolon), since
7269 some programs can't parse those values properly otherwise. Also make
7270 sure BIO's that break lines after each write do not create invalid
7274 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
7275 macros previously used would not encode an empty SEQUENCE OF
7276 and break the signature.
7278 [This change does not apply to 0.9.7.]
7280 *) Zero the premaster secret after deriving the master secret in
7284 *) Add some EVP_add_digest_alias registrations (as found in
7285 OpenSSL_add_all_digests()) to SSL_library_init()
7286 aka OpenSSL_add_ssl_algorithms(). This provides improved
7287 compatibility with peers using X.509 certificates
7288 with unconventional AlgorithmIdentifier OIDs.
7291 *) Fix for Irix with NO_ASM.
7292 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7294 *) ./config script fixes.
7295 [Ulf Moeller, Richard Levitte]
7297 *) Fix 'openssl passwd -1'.
7300 *) Change PKCS12_key_gen_asc() so it can cope with non null
7301 terminated strings whose length is passed in the passlen
7302 parameter, for example from PEM callbacks. This was done
7303 by adding an extra length parameter to asc2uni().
7304 [Steve Henson, reported by <oddissey@samsung.co.kr>]
7306 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
7307 call failed, free the DSA structure.
7310 *) Fix to uni2asc() to cope with zero length Unicode strings.
7311 These are present in some PKCS#12 files.
7314 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
7315 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
7316 when writing a 32767 byte record.
7317 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
7319 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
7320 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
7322 (RSA objects have a reference count access to which is protected
7323 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
7324 so they are meant to be shared between threads.)
7325 [Bodo Moeller, Geoff Thorpe; original patch submitted by
7326 "Reddie, Steven" <Steven.Reddie@ca.com>]
7328 *) Fix a deadlock in CRYPTO_mem_leaks().
7331 *) Use better test patterns in bntest.
7334 *) rand_win.c fix for Borland C.
7337 *) BN_rshift bugfix for n == 0.
7340 *) Add a 'bctest' script that checks for some known 'bc' bugs
7341 so that 'make test' does not abort just because 'bc' is broken.
7344 *) Store verify_result within SSL_SESSION also for client side to
7345 avoid potential security hole. (Re-used sessions on the client side
7346 always resulted in verify_result==X509_V_OK, not using the original
7347 result of the server certificate verification.)
7350 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
7351 SSL3_RT_APPLICATION_DATA, return 0.
7352 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
7356 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
7357 releases, have been re-implemented by renaming the previous
7358 implementations of ssl2_read and ssl3_read to ssl2_read_internal
7359 and ssl3_read_internal, respectively, and adding 'peek' parameters
7360 to them. The new ssl[23]_{read,peek} functions are calls to
7361 ssl[23]_read_internal with the 'peek' flag set appropriately.
7362 A 'peek' parameter has also been added to ssl3_read_bytes, which
7363 does the actual work for ssl3_read_internal.
7366 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
7367 the method-specific "init()" handler. Also clean up ex_data after
7368 calling the method-specific "finish()" handler. Previously, this was
7369 happening the other way round.
7372 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
7373 The previous value, 12, was not always sufficient for BN_mod_exp().
7376 *) Make sure that shared libraries get the internal name engine with
7377 the full version number and not just 0. This should mark the
7378 shared libraries as not backward compatible. Of course, this should
7379 be changed again when we can guarantee backward binary compatibility.
7382 *) Fix typo in get_cert_by_subject() in by_dir.c
7383 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
7385 *) Rework the system to generate shared libraries:
7387 - Make note of the expected extension for the shared libraries and
7388 if there is a need for symbolic links from for example libcrypto.so.0
7389 to libcrypto.so.0.9.7. There is extended info in Configure for
7392 - Make as few rebuilds of the shared libraries as possible.
7394 - Still avoid linking the OpenSSL programs with the shared libraries.
7396 - When installing, install the shared libraries separately from the
7400 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
7402 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
7403 and not in SSL_clear because the latter is also used by the
7404 accept/connect functions; previously, the settings made by
7405 SSL_set_read_ahead would be lost during the handshake.
7406 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
7408 *) Correct util/mkdef.pl to be selective about disabled algorithms.
7409 Previously, it would create entries for disableed algorithms no
7413 *) Added several new manual pages for SSL_* function.
7416 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
7418 *) In ssl23_get_client_hello, generate an error message when faced
7419 with an initial SSL 3.0/TLS record that is too small to contain the
7420 first two bytes of the ClientHello message, i.e. client_version.
7421 (Note that this is a pathologic case that probably has never happened
7422 in real life.) The previous approach was to use the version number
7423 from the record header as a substitute; but our protocol choice
7424 should not depend on that one because it is not authenticated
7425 by the Finished messages.
7428 *) More robust randomness gathering functions for Windows.
7429 [Jeffrey Altman <jaltman@columbia.edu>]
7431 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
7432 not set then we don't setup the error code for issuer check errors
7433 to avoid possibly overwriting other errors which the callback does
7434 handle. If an application does set the flag then we assume it knows
7435 what it is doing and can handle the new informational codes
7439 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
7440 a general "ANY" type, as such it should be able to decode anything
7441 including tagged types. However it didn't check the class so it would
7442 wrongly interpret tagged types in the same way as their universal
7443 counterpart and unknown types were just rejected. Changed so that the
7444 tagged and unknown types are handled in the same way as a SEQUENCE:
7445 that is the encoding is stored intact. There is also a new type
7446 "V_ASN1_OTHER" which is used when the class is not universal, in this
7447 case we have no idea what the actual type is so we just lump them all
7451 *) On VMS, stdout may very well lead to a file that is written to
7452 in a record-oriented fashion. That means that every write() will
7453 write a separate record, which will be read separately by the
7454 programs trying to read from it. This can be very confusing.
7456 The solution is to put a BIO filter in the way that will buffer
7457 text until a linefeed is reached, and then write everything a
7458 line at a time, so every record written will be an actual line,
7459 not chunks of lines and not (usually doesn't happen, but I've
7460 seen it once) several lines in one record. BIO_f_linebuffer() is
7463 Currently, it's a VMS-only method, because that's where it has
7464 been tested well enough.
7467 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
7468 it can return incorrect results.
7469 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
7470 but it was in 0.9.6-beta[12].)
7473 *) Disable the check for content being present when verifying detached
7474 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
7475 include zero length content when signing messages.
7478 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
7479 BIO_ctrl (for BIO pairs).
7482 *) Add DSO method for VMS.
7485 *) Bug fix: Montgomery multiplication could produce results with the
7489 *) Add RPM specification openssl.spec and modify it to build three
7490 packages. The default package contains applications, application
7491 documentation and run-time libraries. The devel package contains
7492 include files, static libraries and function documentation. The
7493 doc package contains the contents of the doc directory. The original
7494 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
7497 *) Add a large number of documentation files for many SSL routines.
7498 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7500 *) Add a configuration entry for Sony News 4.
7501 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
7503 *) Don't set the two most significant bits to one when generating a
7504 random number < q in the DSA library.
7507 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
7508 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
7509 the underlying transport is blocking) if a handshake took place.
7510 (The default behaviour is needed by applications such as s_client
7511 and s_server that use select() to determine when to use SSL_read;
7512 but for applications that know in advance when to expect data, it
7513 just makes things more complicated.)
7516 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
7520 *) Add a few more EBCDIC conditionals that make `req' and `x509'
7521 work better on such systems.
7522 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7524 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
7525 Update PKCS12_parse() so it copies the friendlyName and the
7526 keyid to the certificates aux info.
7529 *) Fix bug in PKCS7_verify() which caused an infinite loop
7530 if there was more than one signature.
7531 [Sven Uszpelkat <su@celocom.de>]
7533 *) Major change in util/mkdef.pl to include extra information
7534 about each symbol, as well as presentig variables as well
7535 as functions. This change means that there's n more need
7536 to rebuild the .num files when some algorithms are excluded.
7539 *) Allow the verify time to be set by an application,
7540 rather than always using the current time.
7543 *) Phase 2 verify code reorganisation. The certificate
7544 verify code now looks up an issuer certificate by a
7545 number of criteria: subject name, authority key id
7546 and key usage. It also verifies self signed certificates
7547 by the same criteria. The main comparison function is
7548 X509_check_issued() which performs these checks.
7550 Lot of changes were necessary in order to support this
7551 without completely rewriting the lookup code.
7553 Authority and subject key identifier are now cached.
7555 The LHASH 'certs' is X509_STORE has now been replaced
7556 by a STACK_OF(X509_OBJECT). This is mainly because an
7557 LHASH can't store or retrieve multiple objects with
7558 the same hash value.
7560 As a result various functions (which were all internal
7561 use only) have changed to handle the new X509_STORE
7562 structure. This will break anything that messed round
7563 with X509_STORE internally.
7565 The functions X509_STORE_add_cert() now checks for an
7566 exact match, rather than just subject name.
7568 The X509_STORE API doesn't directly support the retrieval
7569 of multiple certificates matching a given criteria, however
7570 this can be worked round by performing a lookup first
7571 (which will fill the cache with candidate certificates)
7572 and then examining the cache for matches. This is probably
7573 the best we can do without throwing out X509_LOOKUP
7574 entirely (maybe later...).
7576 The X509_VERIFY_CTX structure has been enhanced considerably.
7578 All certificate lookup operations now go via a get_issuer()
7579 callback. Although this currently uses an X509_STORE it
7580 can be replaced by custom lookups. This is a simple way
7581 to bypass the X509_STORE hackery necessary to make this
7582 work and makes it possible to use more efficient techniques
7583 in future. A very simple version which uses a simple
7584 STACK for its trusted certificate store is also provided
7585 using X509_STORE_CTX_trusted_stack().
7587 The verify_cb() and verify() callbacks now have equivalents
7588 in the X509_STORE_CTX structure.
7590 X509_STORE_CTX also has a 'flags' field which can be used
7591 to customise the verify behaviour.
7594 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
7595 excludes S/MIME capabilities.
7598 *) When a certificate request is read in keep a copy of the
7599 original encoding of the signed data and use it when outputing
7600 again. Signatures then use the original encoding rather than
7601 a decoded, encoded version which may cause problems if the
7602 request is improperly encoded.
7605 *) For consistency with other BIO_puts implementations, call
7606 buffer_write(b, ...) directly in buffer_puts instead of calling
7609 In BIO_puts, increment b->num_write as in BIO_write.
7610 [Peter.Sylvester@EdelWeb.fr]
7612 *) Fix BN_mul_word for the case where the word is 0. (We have to use
7613 BN_zero, we may not return a BIGNUM with an array consisting of
7617 *) Avoid calling abort() from within the library when problems are
7618 detected, except if preprocessor symbols have been defined
7619 (such as REF_CHECK, BN_DEBUG etc.).
7622 *) New openssl application 'rsautl'. This utility can be
7623 used for low level RSA operations. DER public key
7624 BIO/fp routines also added.
7627 *) New Configure entry and patches for compiling on QNX 4.
7628 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
7630 *) A demo state-machine implementation was sponsored by
7631 Nuron (http://www.nuron.com/) and is now available in
7632 demos/state_machine.
7635 *) New options added to the 'dgst' utility for signature
7636 generation and verification.
7639 *) Unrecognized PKCS#7 content types are now handled via a
7640 catch all ASN1_TYPE structure. This allows unsupported
7641 types to be stored as a "blob" and an application can
7642 encode and decode it manually.
7645 *) Fix various signed/unsigned issues to make a_strex.c
7647 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
7649 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
7650 length if passed a buffer. ASN1_INTEGER_to_BN failed
7651 if passed a NULL BN and its argument was negative.
7652 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
7654 *) Modification to PKCS#7 encoding routines to output definite
7655 length encoding. Since currently the whole structures are in
7656 memory there's not real point in using indefinite length
7657 constructed encoding. However if OpenSSL is compiled with
7658 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
7661 *) Added BIO_vprintf() and BIO_vsnprintf().
7664 *) Added more prefixes to parse for in the the strings written
7665 through a logging bio, to cover all the levels that are available
7666 through syslog. The prefixes are now:
7668 PANIC, EMERG, EMR => LOG_EMERG
7669 ALERT, ALR => LOG_ALERT
7670 CRIT, CRI => LOG_CRIT
7671 ERROR, ERR => LOG_ERR
7672 WARNING, WARN, WAR => LOG_WARNING
7673 NOTICE, NOTE, NOT => LOG_NOTICE
7674 INFO, INF => LOG_INFO
7675 DEBUG, DBG => LOG_DEBUG
7677 and as before, if none of those prefixes are present at the
7678 beginning of the string, LOG_ERR is chosen.
7680 On Win32, the LOG_* levels are mapped according to this:
7682 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
7683 LOG_WARNING => EVENTLOG_WARNING_TYPE
7684 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
7688 *) Made it possible to reconfigure with just the configuration
7689 argument "reconf" or "reconfigure". The command line arguments
7690 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
7691 and are retrieved from there when reconfiguring.
7695 [Assar Westerlund <assar@sics.se>, Richard Levitte]
7697 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
7700 *) The obj_dat.pl script was messing up the sorting of object
7701 names. The reason was that it compared the quoted version
7702 of strings as a result "OCSP" > "OCSP Signing" because
7703 " > SPACE. Changed script to store unquoted versions of
7704 names and add quotes on output. It was also omitting some
7705 names from the lookup table if they were given a default
7706 value (that is if SN is missing it is given the same
7707 value as LN and vice versa), these are now added on the
7708 grounds that if an object has a name we should be able to
7709 look it up. Finally added warning output when duplicate
7710 short or long names are found.
7713 *) Changes needed for Tandem NSK.
7714 [Scott Uroff <scott@xypro.com>]
7716 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
7717 RSA_padding_check_SSLv23(), special padding was never detected
7718 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
7719 version rollback attacks was not effective.
7721 In s23_clnt.c, don't use special rollback-attack detection padding
7722 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
7723 client; similarly, in s23_srvr.c, don't do the rollback check if
7724 SSL 2.0 is the only protocol enabled in the server.
7727 *) Make it possible to get hexdumps of unprintable data with 'openssl
7728 asn1parse'. By implication, the functions ASN1_parse_dump() and
7729 BIO_dump_indent() are added.
7732 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
7733 these print out strings and name structures based on various
7734 flags including RFC2253 support and proper handling of
7735 multibyte characters. Added options to the 'x509' utility
7736 to allow the various flags to be set.
7739 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
7740 Also change the functions X509_cmp_current_time() and
7741 X509_gmtime_adj() work with an ASN1_TIME structure,
7742 this will enable certificates using GeneralizedTime in validity
7743 dates to be checked.
7746 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
7747 negative public key encodings) on by default,
7748 NO_NEG_PUBKEY_BUG can be set to disable it.
7751 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
7752 content octets. An i2c_ASN1_OBJECT is unnecessary because
7753 the encoding can be trivially obtained from the structure.
7756 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
7757 not read locks (CRYPTO_r_[un]lock).
7760 *) A first attempt at creating official support for shared
7761 libraries through configuration. I've kept it so the
7762 default is static libraries only, and the OpenSSL programs
7763 are always statically linked for now, but there are
7764 preparations for dynamic linking in place.
7765 This has been tested on Linux and Tru64.
7768 *) Randomness polling function for Win9x, as described in:
7769 Peter Gutmann, Software Generation of Practically Strong
7773 *) Fix so PRNG is seeded in req if using an already existing
7777 *) New options to smime application. -inform and -outform
7778 allow alternative formats for the S/MIME message including
7779 PEM and DER. The -content option allows the content to be
7780 specified separately. This should allow things like Netscape
7781 form signing output easier to verify.
7784 *) Fix the ASN1 encoding of tags using the 'long form'.
7787 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
7788 STRING types. These convert content octets to and from the
7789 underlying type. The actual tag and length octets are
7790 already assumed to have been read in and checked. These
7791 are needed because all other string types have virtually
7792 identical handling apart from the tag. By having versions
7793 of the ASN1 functions that just operate on content octets
7794 IMPLICIT tagging can be handled properly. It also allows
7795 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
7796 and ASN1_INTEGER are identical apart from the tag.
7799 *) Change the handling of OID objects as follows:
7801 - New object identifiers are inserted in objects.txt, following
7802 the syntax given in objects.README.
7803 - objects.pl is used to process obj_mac.num and create a new
7805 - obj_dat.pl is used to create a new obj_dat.h, using the data in
7808 This is currently kind of a hack, and the perl code in objects.pl
7809 isn't very elegant, but it works as I intended. The simplest way
7810 to check that it worked correctly is to look in obj_dat.h and
7811 check the array nid_objs and make sure the objects haven't moved
7812 around (this is important!). Additions are OK, as well as
7813 consistent name changes.
7816 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
7819 *) Addition of the command line parameter '-rand file' to 'openssl req'.
7820 The given file adds to whatever has already been seeded into the
7821 random pool through the RANDFILE configuration file option or
7822 environment variable, or the default random state file.
7825 *) mkstack.pl now sorts each macro group into lexical order.
7826 Previously the output order depended on the order the files
7827 appeared in the directory, resulting in needless rewriting
7831 *) Patches to make OpenSSL compile under Win32 again. Mostly
7832 work arounds for the VC++ problem that it treats func() as
7833 func(void). Also stripped out the parts of mkdef.pl that
7834 added extra typesafe functions: these no longer exist.
7837 *) Reorganisation of the stack code. The macros are now all
7838 collected in safestack.h . Each macro is defined in terms of
7839 a "stack macro" of the form SKM_<name>(type, a, b). The
7840 DEBUG_SAFESTACK is now handled in terms of function casts,
7841 this has the advantage of retaining type safety without the
7842 use of additional functions. If DEBUG_SAFESTACK is not defined
7843 then the non typesafe macros are used instead. Also modified the
7844 mkstack.pl script to handle the new form. Needs testing to see
7845 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
7846 the default if no major problems. Similar behaviour for ASN1_SET_OF
7847 and PKCS12_STACK_OF.
7850 *) When some versions of IIS use the 'NET' form of private key the
7851 key derivation algorithm is different. Normally MD5(password) is
7852 used as a 128 bit RC4 key. In the modified case
7853 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
7854 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
7855 as the old Netscape_RSA functions except they have an additional
7856 'sgckey' parameter which uses the modified algorithm. Also added
7857 an -sgckey command line option to the rsa utility. Thanks to
7858 Adrian Peck <bertie@ncipher.com> for posting details of the modified
7859 algorithm to openssl-dev.
7862 *) The evp_local.h macros were using 'c.##kname' which resulted in
7863 invalid expansion on some systems (SCO 5.0.5 for example).
7864 Corrected to 'c.kname'.
7865 [Phillip Porch <root@theporch.com>]
7867 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
7868 a STACK of email addresses from a certificate or request, these look
7869 in the subject name and the subject alternative name extensions and
7870 omit any duplicate addresses.
7873 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
7874 This makes DSA verification about 2 % faster.
7877 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
7878 (meaning that now 2^5 values will be precomputed, which is only 4 KB
7879 plus overhead for 1024 bit moduli).
7880 This makes exponentiations about 0.5 % faster for 1024 bit
7881 exponents (as measured by "openssl speed rsa2048").
7884 *) Rename memory handling macros to avoid conflicts with other
7886 Malloc => OPENSSL_malloc
7887 Malloc_locked => OPENSSL_malloc_locked
7888 Realloc => OPENSSL_realloc
7889 Free => OPENSSL_free
7892 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
7893 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
7896 *) CygWin32 support.
7897 [John Jarvie <jjarvie@newsguy.com>]
7899 *) The type-safe stack code has been rejigged. It is now only compiled
7900 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
7901 by default all type-specific stack functions are "#define"d back to
7902 standard stack functions. This results in more streamlined output
7903 but retains the type-safety checking possibilities of the original
7907 *) The STACK code has been cleaned up, and certain type declarations
7908 that didn't make a lot of sense have been brought in line. This has
7909 also involved a cleanup of sorts in safestack.h to more correctly
7910 map type-safe stack functions onto their plain stack counterparts.
7911 This work has also resulted in a variety of "const"ifications of
7912 lots of the code, especially "_cmp" operations which should normally
7913 be prototyped with "const" parameters anyway.
7916 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
7917 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
7918 (The PRNG state consists of two parts, the large pool 'state' and 'md',
7919 where all of 'md' is used each time the PRNG is used, but 'state'
7920 is used only indexed by a cyclic counter. As entropy may not be
7921 well distributed from the beginning, 'md' is important as a
7922 chaining variable. However, the output function chains only half
7923 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
7924 all of 'md', and seeding with STATE_SIZE dummy bytes will result
7925 in all of 'state' being rewritten, with the new values depending
7926 on virtually all of 'md'. This overcomes the 80 bit limitation.)
7929 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
7930 the handshake is continued after ssl_verify_cert_chain();
7931 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
7932 can lead to 'unexplainable' connection aborts later.
7933 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
7935 *) Major EVP API cipher revision.
7936 Add hooks for extra EVP features. This allows various cipher
7937 parameters to be set in the EVP interface. Support added for variable
7938 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
7939 setting of RC2 and RC5 parameters.
7941 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
7944 Remove lots of duplicated code from the EVP library. For example *every*
7945 cipher init() function handles the 'iv' in the same way according to the
7946 cipher mode. They also all do nothing if the 'key' parameter is NULL and
7947 for CFB and OFB modes they zero ctx->num.
7949 New functionality allows removal of S/MIME code RC2 hack.
7951 Most of the routines have the same form and so can be declared in terms
7954 By shifting this to the top level EVP_CipherInit() it can be removed from
7955 all individual ciphers. If the cipher wants to handle IVs or keys
7956 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
7959 Change lots of functions like EVP_EncryptUpdate() to now return a
7960 value: although software versions of the algorithms cannot fail
7961 any installed hardware versions can.
7964 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
7965 this option is set, tolerate broken clients that send the negotiated
7966 protocol version number instead of the requested protocol version
7970 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7971 i.e. non-zero for export ciphersuites, zero otherwise.
7972 Previous versions had this flag inverted, inconsistent with
7973 rsa_tmp_cb (..._TMP_RSA_CB).
7974 [Bodo Moeller; problem reported by Amit Chopra]
7976 *) Add missing DSA library text string. Work around for some IIS
7977 key files with invalid SEQUENCE encoding.
7980 *) Add a document (doc/standards.txt) that list all kinds of standards
7981 and so on that are implemented in OpenSSL.
7984 *) Enhance c_rehash script. Old version would mishandle certificates
7985 with the same subject name hash and wouldn't handle CRLs at all.
7986 Added -fingerprint option to crl utility, to support new c_rehash
7990 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7993 *) Fix for SSL server purpose checking. Server checking was
7994 rejecting certificates which had extended key usage present
7995 but no ssl client purpose.
7996 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7998 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7999 is a little unclear about how a blank password is handled.
8000 Since the password in encoded as a BMPString with terminating
8001 double NULL a zero length password would end up as just the
8002 double NULL. However no password at all is different and is
8003 handled differently in the PKCS#12 key generation code. NS
8004 treats a blank password as zero length. MSIE treats it as no
8005 password on export: but it will try both on import. We now do
8006 the same: PKCS12_parse() tries zero length and no password if
8007 the password is set to "" or NULL (NULL is now a valid password:
8008 it wasn't before) as does the pkcs12 application.
8011 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
8012 perror when PEM_read_bio_X509_REQ fails, the error message must
8013 be obtained from the error queue.
8016 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
8017 it in ERR_remove_state if appropriate, and change ERR_get_state
8018 accordingly to avoid race conditions (this is necessary because
8019 thread_hash is no longer constant once set).
8022 *) Bugfix for linux-elf makefile.one.
8025 *) RSA_get_default_method() will now cause a default
8026 RSA_METHOD to be chosen if one doesn't exist already.
8027 Previously this was only set during a call to RSA_new()
8028 or RSA_new_method(NULL) meaning it was possible for
8029 RSA_get_default_method() to return NULL.
8032 *) Added native name translation to the existing DSO code
8033 that will convert (if the flag to do so is set) filenames
8034 that are sufficiently small and have no path information
8035 into a canonical native form. Eg. "blah" converted to
8036 "libblah.so" or "blah.dll" etc.
8039 *) New function ERR_error_string_n(e, buf, len) which is like
8040 ERR_error_string(e, buf), but writes at most 'len' bytes
8041 including the 0 terminator. For ERR_error_string_n, 'buf'
8043 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
8045 *) CONF library reworked to become more general. A new CONF
8046 configuration file reader "class" is implemented as well as a
8047 new functions (NCONF_*, for "New CONF") to handle it. The now
8048 old CONF_* functions are still there, but are reimplemented to
8049 work in terms of the new functions. Also, a set of functions
8050 to handle the internal storage of the configuration data is
8051 provided to make it easier to write new configuration file
8052 reader "classes" (I can definitely see something reading a
8053 configuration file in XML format, for example), called _CONF_*,
8054 or "the configuration storage API"...
8056 The new configuration file reading functions are:
8058 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
8059 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
8061 NCONF_default, NCONF_WIN32
8063 NCONF_dump_fp, NCONF_dump_bio
8065 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
8066 NCONF_new creates a new CONF object. This works in the same way
8067 as other interfaces in OpenSSL, like the BIO interface.
8068 NCONF_dump_* dump the internal storage of the configuration file,
8069 which is useful for debugging. All other functions take the same
8070 arguments as the old CONF_* functions wth the exception of the
8071 first that must be a `CONF *' instead of a `LHASH *'.
8073 To make it easer to use the new classes with the old CONF_* functions,
8074 the function CONF_set_default_method is provided.
8077 *) Add '-tls1' option to 'openssl ciphers', which was already
8078 mentioned in the documentation but had not been implemented.
8079 (This option is not yet really useful because even the additional
8080 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
8083 *) Initial DSO code added into libcrypto for letting OpenSSL (and
8084 OpenSSL-based applications) load shared libraries and bind to
8085 them in a portable way.
8086 [Geoff Thorpe, with contributions from Richard Levitte]
8088 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
8090 *) Make sure _lrotl and _lrotr are only used with MSVC.
8092 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
8093 (the default implementation of RAND_status).
8095 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
8096 to '-clrext' (= clear extensions), as intended and documented.
8097 [Bodo Moeller; inconsistency pointed out by Michael Attili
8100 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
8101 was larger than the MD block size.
8102 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
8104 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
8105 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
8106 using the passed key: if the passed key was a private key the result
8107 of X509_print(), for example, would be to print out all the private key
8111 *) des_quad_cksum() byte order bug fix.
8112 [Ulf Möller, using the problem description in krb4-0.9.7, where
8113 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
8115 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
8117 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
8119 *) For easily testing in shell scripts whether some command
8120 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
8121 returns with exit code 0 iff no command of the given name is available.
8122 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
8123 the output goes to stdout and nothing is printed to stderr.
8124 Additional arguments are always ignored.
8126 Since for each cipher there is a command of the same name,
8127 the 'no-cipher' compilation switches can be tested this way.
8129 ('openssl no-XXX' is not able to detect pseudo-commands such
8130 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
8133 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
8136 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
8137 is set; it will be thrown away anyway because each handshake creates
8139 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
8140 to parameters -- in previous versions (since OpenSSL 0.9.3) the
8141 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
8142 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
8145 *) New s_client option -ign_eof: EOF at stdin is ignored, and
8146 'Q' and 'R' lose their special meanings (quit/renegotiate).
8147 This is part of what -quiet does; unlike -quiet, -ign_eof
8148 does not suppress any output.
8151 *) Add compatibility options to the purpose and trust code. The
8152 purpose X509_PURPOSE_ANY is "any purpose" which automatically
8153 accepts a certificate or CA, this was the previous behaviour,
8154 with all the associated security issues.
8156 X509_TRUST_COMPAT is the old trust behaviour: only and
8157 automatically trust self signed roots in certificate store. A
8158 new trust setting X509_TRUST_DEFAULT is used to specify that
8159 a purpose has no associated trust setting and it should instead
8160 use the value in the default purpose.
8163 *) Fix the PKCS#8 DSA private key code so it decodes keys again
8164 and fix a memory leak.
8167 *) In util/mkerr.pl (which implements 'make errors'), preserve
8168 reason strings from the previous version of the .c file, as
8169 the default to have only downcase letters (and digits) in
8170 automatically generated reasons codes is not always appropriate.
8173 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
8174 using strerror. Previously, ERR_reason_error_string() returned
8175 library names as reason strings for SYSerr; but SYSerr is a special
8176 case where small numbers are errno values, not library numbers.
8179 *) Add '-dsaparam' option to 'openssl dhparam' application. This
8180 converts DSA parameters into DH parameters. (When creating parameters,
8181 DSA_generate_parameters is used.)
8184 *) Include 'length' (recommended exponent length) in C code generated
8185 by 'openssl dhparam -C'.
8188 *) The second argument to set_label in perlasm was already being used
8189 so couldn't be used as a "file scope" flag. Moved to third argument
8193 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
8194 instead of RAND_bytes for encryption IVs and salts.
8197 *) Include RAND_status() into RAND_METHOD instead of implementing
8198 it only for md_rand.c Otherwise replacing the PRNG by calling
8199 RAND_set_rand_method would be impossible.
8202 *) Don't let DSA_generate_key() enter an infinite loop if the random
8203 number generation fails.
8206 *) New 'rand' application for creating pseudo-random output.
8209 *) Added configuration support for Linux/IA64
8210 [Rolf Haberrecker <rolf@suse.de>]
8212 *) Assembler module support for Mingw32.
8215 *) Shared library support for HPUX (in shlib/).
8216 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
8218 *) Shared library support for Solaris gcc.
8219 [Lutz Behnke <behnke@trustcenter.de>]
8221 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
8223 *) PKCS7_encrypt() was adding text MIME headers twice because they
8224 were added manually and by SMIME_crlf_copy().
8227 *) In bntest.c don't call BN_rand with zero bits argument.
8228 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
8230 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
8231 case was implemented. This caused BN_div_recp() to fail occasionally.
8234 *) Add an optional second argument to the set_label() in the perl
8235 assembly language builder. If this argument exists and is set
8236 to 1 it signals that the assembler should use a symbol whose
8237 scope is the entire file, not just the current function. This
8238 is needed with MASM which uses the format label:: for this scope.
8239 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
8241 *) Change the ASN1 types so they are typedefs by default. Before
8242 almost all types were #define'd to ASN1_STRING which was causing
8243 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
8247 *) Change names of new functions to the new get1/get0 naming
8248 convention: After 'get1', the caller owns a reference count
8249 and has to call ..._free; 'get0' returns a pointer to some
8250 data structure without incrementing reference counters.
8251 (Some of the existing 'get' functions increment a reference
8252 counter, some don't.)
8253 Similarly, 'set1' and 'add1' functions increase reference
8254 counters or duplicate objects.
8257 *) Allow for the possibility of temp RSA key generation failure:
8258 the code used to assume it always worked and crashed on failure.
8261 *) Fix potential buffer overrun problem in BIO_printf().
8262 [Ulf Möller, using public domain code by Patrick Powell; problem
8263 pointed out by David Sacerdote <das33@cornell.edu>]
8265 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
8266 RAND_egd() and RAND_status(). In the command line application,
8267 the EGD socket can be specified like a seed file using RANDFILE
8271 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
8272 Some CAs (e.g. Verisign) distribute certificates in this form.
8275 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
8276 list to exclude them. This means that no special compilation option
8277 is needed to use anonymous DH: it just needs to be included in the
8281 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
8282 EVP_MD_type. The old functionality is available in a new macro called
8283 EVP_MD_md(). Change code that uses it and update docs.
8286 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
8287 where the 'void *' argument is replaced by a function pointer argument.
8288 Previously 'void *' was abused to point to functions, which works on
8289 many platforms, but is not correct. As these functions are usually
8290 called by macros defined in OpenSSL header files, most source code
8291 should work without changes.
8294 *) <openssl/opensslconf.h> (which is created by Configure) now contains
8295 sections with information on -D... compiler switches used for
8296 compiling the library so that applications can see them. To enable
8297 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
8298 must be defined. E.g.,
8299 #define OPENSSL_ALGORITHM_DEFINES
8300 #include <openssl/opensslconf.h>
8301 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
8302 [Richard Levitte, Ulf and Bodo Möller]
8304 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
8308 *) Change the 'other' type in certificate aux info to a STACK_OF
8309 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
8310 the required ASN1 format: arbitrary types determined by an OID.
8313 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
8314 argument to 'req'. This is not because the function is newer or
8315 better than others it just uses the work 'NEW' in the certificate
8316 request header lines. Some software needs this.
8319 *) Reorganise password command line arguments: now passwords can be
8320 obtained from various sources. Delete the PEM_cb function and make
8321 it the default behaviour: i.e. if the callback is NULL and the
8322 usrdata argument is not NULL interpret it as a null terminated pass
8323 phrase. If usrdata and the callback are NULL then the pass phrase
8324 is prompted for as usual.
8327 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
8328 the support is automatically enabled. The resulting binaries will
8329 autodetect the card and use it if present.
8330 [Ben Laurie and Compaq Inc.]
8332 *) Work around for Netscape hang bug. This sends certificate request
8333 and server done in one record. Since this is perfectly legal in the
8334 SSL/TLS protocol it isn't a "bug" option and is on by default. See
8335 the bugs/SSLv3 entry for more info.
8338 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
8341 *) Add -rand argument to smime and pkcs12 applications and read/write
8345 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
8348 *) Add command line password options to the remaining applications.
8351 *) Bug fix for BN_div_recp() for numerators with an even number of
8355 *) More tests in bntest.c, and changed test_bn output.
8358 *) ./config recognizes MacOS X now.
8361 *) Bug fix for BN_div() when the first words of num and divsor are
8362 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
8365 *) Add support for various broken PKCS#8 formats, and command line
8366 options to produce them.
8369 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
8370 get temporary BIGNUMs from a BN_CTX.
8373 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
8377 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
8378 include a #define from the old name to the new. The original intent
8379 was that statically linked binaries could for example just call
8380 SSLeay_add_all_ciphers() to just add ciphers to the table and not
8381 link with digests. This never worked becayse SSLeay_add_all_digests()
8382 and SSLeay_add_all_ciphers() were in the same source file so calling
8383 one would link with the other. They are now in separate source files.
8386 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
8389 *) Use a less unusual form of the Miller-Rabin primality test (it used
8390 a binary algorithm for exponentiation integrated into the Miller-Rabin
8391 loop, our standard modexp algorithms are faster).
8394 *) Support for the EBCDIC character set completed.
8395 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
8397 *) Source code cleanups: use const where appropriate, eliminate casts,
8398 use void * instead of char * in lhash.
8401 *) Bugfix: ssl3_send_server_key_exchange was not restartable
8402 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
8403 this the server could overwrite ephemeral keys that the client
8407 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
8408 using 50 iterations of the Rabin-Miller test.
8410 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
8411 iterations of the Rabin-Miller test as required by the appendix
8412 to FIPS PUB 186[-1]) instead of DSA_is_prime.
8413 As BN_is_prime_fasttest includes trial division, DSA parameter
8414 generation becomes much faster.
8416 This implies a change for the callback functions in DSA_is_prime
8417 and DSA_generate_parameters: The callback function is called once
8418 for each positive witness in the Rabin-Miller test, not just
8419 occasionally in the inner loop; and the parameters to the
8420 callback function now provide an iteration count for the outer
8421 loop rather than for the current invocation of the inner loop.
8422 DSA_generate_parameters additionally can call the callback
8423 function with an 'iteration count' of -1, meaning that a
8424 candidate has passed the trial division test (when q is generated
8425 from an application-provided seed, trial division is skipped).
8428 *) New function BN_is_prime_fasttest that optionally does trial
8429 division before starting the Rabin-Miller test and has
8430 an additional BN_CTX * argument (whereas BN_is_prime always
8431 has to allocate at least one BN_CTX).
8432 'callback(1, -1, cb_arg)' is called when a number has passed the
8433 trial division stage.
8436 *) Fix for bug in CRL encoding. The validity dates weren't being handled
8440 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
8443 *) New function BN_pseudo_rand().
8446 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
8447 bignum version of BN_from_montgomery() with the working code from
8448 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
8452 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
8453 made it impossible to use the same SSL_SESSION data structure in
8454 SSL2 clients in multiple threads.
8457 *) The return value of RAND_load_file() no longer counts bytes obtained
8458 by stat(). RAND_load_file(..., -1) is new and uses the complete file
8459 to seed the PRNG (previously an explicit byte count was required).
8460 [Ulf Möller, Bodo Möller]
8462 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
8463 used (char *) instead of (void *) and had casts all over the place.
8466 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
8469 *) Retain source code compatibility for BN_prime_checks macro:
8470 BN_is_prime(..., BN_prime_checks, ...) now uses
8471 BN_prime_checks_for_size to determine the appropriate number of
8472 Rabin-Miller iterations.
8475 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
8476 DH_CHECK_P_NOT_SAFE_PRIME.
8477 (Check if this is true? OpenPGP calls them "strong".)
8480 *) Merge the functionality of "dh" and "gendh" programs into a new program
8481 "dhparam". The old programs are retained for now but will handle DH keys
8482 (instead of parameters) in future.
8485 *) Make the ciphers, s_server and s_client programs check the return values
8486 when a new cipher list is set.
8489 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
8490 ciphers. Before when the 56bit ciphers were enabled the sorting was
8493 The syntax for the cipher sorting has been extended to support sorting by
8494 cipher-strength (using the strength_bits hard coded in the tables).
8495 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
8497 Fix a bug in the cipher-command parser: when supplying a cipher command
8498 string with an "undefined" symbol (neither command nor alphanumeric
8499 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
8500 an error is flagged.
8502 Due to the strength-sorting extension, the code of the
8503 ssl_create_cipher_list() function was completely rearranged. I hope that
8504 the readability was also increased :-)
8505 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8507 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
8508 for the first serial number and places 2 in the serial number file. This
8509 avoids problems when the root CA is created with serial number zero and
8510 the first user certificate has the same issuer name and serial number
8514 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
8515 the new code. Add documentation for this stuff.
8518 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
8519 X509_*() to X509at_*() on the grounds that they don't handle X509
8520 structures and behave in an analagous way to the X509v3 functions:
8521 they shouldn't be called directly but wrapper functions should be used
8524 So we also now have some wrapper functions that call the X509at functions
8525 when passed certificate requests. (TO DO: similar things can be done with
8526 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
8527 things. Some of these need some d2i or i2d and print functionality
8528 because they handle more complex structures.)
8531 *) Add missing #ifndefs that caused missing symbols when building libssl
8532 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
8533 NO_RSA in ssl/s2*.c.
8534 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
8536 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
8537 has a return value which indicates the quality of the random data
8538 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
8539 error queue. New function RAND_pseudo_bytes() generates output that is
8540 guaranteed to be unique but not unpredictable. RAND_add is like
8541 RAND_seed, but takes an extra argument for an entropy estimate
8542 (RAND_seed always assumes full entropy).
8545 *) Do more iterations of Rabin-Miller probable prime test (specifically,
8546 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
8547 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
8548 in crypto/bn/bn_prime.c for the complete table). This guarantees a
8549 false-positive rate of at most 2^-80 for random input.
8552 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
8555 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
8556 in the 0.9.5 release), this returns the chain
8557 from an X509_CTX structure with a dup of the stack and all
8558 the X509 reference counts upped: so the stack will exist
8559 after X509_CTX_cleanup() has been called. Modify pkcs12.c
8562 Also make SSL_SESSION_print() print out the verify return
8566 *) Add manpage for the pkcs12 command. Also change the default
8567 behaviour so MAC iteration counts are used unless the new
8568 -nomaciter option is used. This improves file security and
8569 only older versions of MSIE (4.0 for example) need it.
8572 *) Honor the no-xxx Configure options when creating .DEF files.
8575 *) Add PKCS#10 attributes to field table: challengePassword,
8576 unstructuredName and unstructuredAddress. These are taken from
8577 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
8578 international characters are used.
8580 More changes to X509_ATTRIBUTE code: allow the setting of types
8581 based on strings. Remove the 'loc' parameter when adding
8582 attributes because these will be a SET OF encoding which is sorted
8586 *) Initial changes to the 'req' utility to allow request generation
8587 automation. This will allow an application to just generate a template
8588 file containing all the field values and have req construct the
8591 Initial support for X509_ATTRIBUTE handling. Stacks of these are
8592 used all over the place including certificate requests and PKCS#7
8593 structures. They are currently handled manually where necessary with
8594 some primitive wrappers for PKCS#7. The new functions behave in a
8595 manner analogous to the X509 extension functions: they allow
8596 attributes to be looked up by NID and added.
8598 Later something similar to the X509V3 code would be desirable to
8599 automatically handle the encoding, decoding and printing of the
8600 more complex types. The string types like challengePassword can
8601 be handled by the string table functions.
8603 Also modified the multi byte string table handling. Now there is
8604 a 'global mask' which masks out certain types. The table itself
8605 can use the flag STABLE_NO_MASK to ignore the mask setting: this
8606 is useful when for example there is only one permissible type
8607 (as in countryName) and using the mask might result in no valid
8611 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
8612 SSL_get_peer_finished to allow applications to obtain the latest
8613 Finished messages sent to the peer or expected from the peer,
8614 respectively. (SSL_get_peer_finished is usually the Finished message
8615 actually received from the peer, otherwise the protocol will be aborted.)
8617 As the Finished message are message digests of the complete handshake
8618 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
8619 be used for external authentication procedures when the authentication
8620 provided by SSL/TLS is not desired or is not enough.
8623 *) Enhanced support for Alpha Linux is added. Now ./config checks if
8624 the host supports BWX extension and if Compaq C is present on the
8625 $PATH. Just exploiting of the BWX extension results in 20-30%
8626 performance kick for some algorithms, e.g. DES and RC4 to mention
8627 a couple. Compaq C in turn generates ~20% faster code for MD5 and
8631 *) Add support for MS "fast SGC". This is arguably a violation of the
8632 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
8633 weak crypto and after checking the certificate is SGC a second one
8634 with strong crypto. MS SGC stops the first handshake after receiving
8635 the server certificate message and sends a second client hello. Since
8636 a server will typically do all the time consuming operations before
8637 expecting any further messages from the client (server key exchange
8638 is the most expensive) there is little difference between the two.
8640 To get OpenSSL to support MS SGC we have to permit a second client
8641 hello message after we have sent server done. In addition we have to
8642 reset the MAC if we do get this second client hello.
8645 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
8646 if a DER encoded private key is RSA or DSA traditional format. Changed
8647 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
8648 format DER encoded private key. Newer code should use PKCS#8 format which
8649 has the key type encoded in the ASN1 structure. Added DER private key
8650 support to pkcs8 application.
8653 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
8654 ciphersuites has been selected (as required by the SSL 3/TLS 1
8655 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
8656 is set, we interpret this as a request to violate the specification
8657 (the worst that can happen is a handshake failure, and 'correct'
8658 behaviour would result in a handshake failure anyway).
8661 *) In SSL_CTX_add_session, take into account that there might be multiple
8662 SSL_SESSION structures with the same session ID (e.g. when two threads
8663 concurrently obtain them from an external cache).
8664 The internal cache can handle only one SSL_SESSION with a given ID,
8665 so if there's a conflict, we now throw out the old one to achieve
8669 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
8670 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
8671 some routines that use cipher OIDs: some ciphers do not have OIDs
8672 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
8676 *) Simplify the trust setting structure and code. Now we just have
8677 two sequences of OIDs for trusted and rejected settings. These will
8678 typically have values the same as the extended key usage extension
8679 and any application specific purposes.
8681 The trust checking code now has a default behaviour: it will just
8682 check for an object with the same NID as the passed id. Functions can
8683 be provided to override either the default behaviour or the behaviour
8684 for a given id. SSL client, server and email already have functions
8685 in place for compatibility: they check the NID and also return "trusted"
8686 if the certificate is self signed.
8689 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
8690 traditional format into an EVP_PKEY structure.
8693 *) Add a password callback function PEM_cb() which either prompts for
8694 a password if usr_data is NULL or otherwise assumes it is a null
8695 terminated password. Allow passwords to be passed on command line
8696 environment or config files in a few more utilities.
8699 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
8700 keys. Add some short names for PKCS#8 PBE algorithms and allow them
8701 to be specified on the command line for the pkcs8 and pkcs12 utilities.
8702 Update documentation.
8705 *) Support for ASN1 "NULL" type. This could be handled before by using
8706 ASN1_TYPE but there wasn't any function that would try to read a NULL
8707 and produce an error if it couldn't. For compatibility we also have
8708 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
8709 don't allocate anything because they don't need to.
8712 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
8714 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
8716 *) Rebuild of the memory allocation routines used by OpenSSL code and
8717 possibly others as well. The purpose is to make an interface that
8718 provide hooks so anyone can build a separate set of allocation and
8719 deallocation routines to be used by OpenSSL, for example memory
8720 pool implementations, or something else, which was previously hard
8721 since Malloc(), Realloc() and Free() were defined as macros having
8722 the values malloc, realloc and free, respectively (except for Win32
8723 compilations). The same is provided for memory debugging code.
8724 OpenSSL already comes with functionality to find memory leaks, but
8725 this gives people a chance to debug other memory problems.
8727 With these changes, a new set of functions and macros have appeared:
8729 CRYPTO_set_mem_debug_functions() [F]
8730 CRYPTO_get_mem_debug_functions() [F]
8731 CRYPTO_dbg_set_options() [F]
8732 CRYPTO_dbg_get_options() [F]
8733 CRYPTO_malloc_debug_init() [M]
8735 The memory debug functions are NULL by default, unless the library
8736 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
8737 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
8738 gives the standard debugging functions that come with OpenSSL) or
8739 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
8740 provided by the library user) must be used. When the standard
8741 debugging functions are used, CRYPTO_dbg_set_options can be used to
8742 request additional information:
8743 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
8744 the CRYPTO_MDEBUG_xxx macro when compiling the library.
8746 Also, things like CRYPTO_set_mem_functions will always give the
8747 expected result (the new set of functions is used for allocation
8748 and deallocation) at all times, regardless of platform and compiler
8751 To finish it up, some functions that were never use in any other
8752 way than through macros have a new API and new semantic:
8755 CRYPTO_dbg_realloc()
8758 All macros of value have retained their old syntax.
8759 [Richard Levitte and Bodo Moeller]
8761 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
8762 ordering of SMIMECapabilities wasn't in "strength order" and there
8763 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
8767 *) Some ASN1 types with illegal zero length encoding (INTEGER,
8768 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
8769 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
8771 *) Merge in my S/MIME library for OpenSSL. This provides a simple
8772 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
8773 functionality to handle multipart/signed properly) and a utility
8774 called 'smime' to call all this stuff. This is based on code I
8775 originally wrote for Celo who have kindly allowed it to be
8776 included in OpenSSL.
8779 *) Add variants des_set_key_checked and des_set_key_unchecked of
8780 des_set_key (aka des_key_sched). Global variable des_check_key
8781 decides which of these is called by des_set_key; this way
8782 des_check_key behaves as it always did, but applications and
8783 the library itself, which was buggy for des_check_key == 1,
8784 have a cleaner way to pick the version they need.
8787 *) New function PKCS12_newpass() which changes the password of a
8791 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
8792 dynamic mix. In both cases the ids can be used as an index into the
8793 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
8794 functions so they accept a list of the field values and the
8795 application doesn't need to directly manipulate the X509_TRUST
8799 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
8803 *) Modify the way the V3 extension code looks up extensions. This now
8804 works in a similar way to the object code: we have some "standard"
8805 extensions in a static table which is searched with OBJ_bsearch()
8806 and the application can add dynamic ones if needed. The file
8807 crypto/x509v3/ext_dat.h now has the info: this file needs to be
8808 updated whenever a new extension is added to the core code and kept
8809 in ext_nid order. There is a simple program 'tabtest.c' which checks
8810 this. New extensions are not added too often so this file can readily
8811 be maintained manually.
8813 There are two big advantages in doing things this way. The extensions
8814 can be looked up immediately and no longer need to be "added" using
8815 X509V3_add_standard_extensions(): this function now does nothing.
8816 [Side note: I get *lots* of email saying the extension code doesn't
8817 work because people forget to call this function]
8818 Also no dynamic allocation is done unless new extensions are added:
8819 so if we don't add custom extensions there is no need to call
8820 X509V3_EXT_cleanup().
8823 *) Modify enc utility's salting as follows: make salting the default. Add a
8824 magic header, so unsalted files fail gracefully instead of just decrypting
8825 to garbage. This is because not salting is a big security hole, so people
8826 should be discouraged from doing it.
8829 *) Fixes and enhancements to the 'x509' utility. It allowed a message
8830 digest to be passed on the command line but it only used this
8831 parameter when signing a certificate. Modified so all relevant
8832 operations are affected by the digest parameter including the
8833 -fingerprint and -x509toreq options. Also -x509toreq choked if a
8834 DSA key was used because it didn't fix the digest.
8837 *) Initial certificate chain verify code. Currently tests the untrusted
8838 certificates for consistency with the verify purpose (which is set
8839 when the X509_STORE_CTX structure is set up) and checks the pathlength.
8841 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
8842 this is because it will reject chains with invalid extensions whereas
8843 every previous version of OpenSSL and SSLeay made no checks at all.
8845 Trust code: checks the root CA for the relevant trust settings. Trust
8846 settings have an initial value consistent with the verify purpose: e.g.
8847 if the verify purpose is for SSL client use it expects the CA to be
8848 trusted for SSL client use. However the default value can be changed to
8849 permit custom trust settings: one example of this would be to only trust
8850 certificates from a specific "secure" set of CAs.
8852 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
8853 which should be used for version portability: especially since the
8854 verify structure is likely to change more often now.
8856 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
8857 to set them. If not set then assume SSL clients will verify SSL servers
8860 Two new options to the verify program: -untrusted allows a set of
8861 untrusted certificates to be passed in and -purpose which sets the
8862 intended purpose of the certificate. If a purpose is set then the
8863 new chain verify code is used to check extension consistency.
8866 *) Support for the authority information access extension.
8869 *) Modify RSA and DSA PEM read routines to transparently handle
8870 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
8871 public keys in a format compatible with certificate
8872 SubjectPublicKeyInfo structures. Unfortunately there were already
8873 functions called *_PublicKey_* which used various odd formats so
8874 these are retained for compatibility: however the DSA variants were
8875 never in a public release so they have been deleted. Changed dsa/rsa
8876 utilities to handle the new format: note no releases ever handled public
8877 keys so we should be OK.
8879 The primary motivation for this change is to avoid the same fiasco
8880 that dogs private keys: there are several incompatible private key
8881 formats some of which are standard and some OpenSSL specific and
8882 require various evil hacks to allow partial transparent handling and
8883 even then it doesn't work with DER formats. Given the option anything
8884 other than PKCS#8 should be dumped: but the other formats have to
8885 stay in the name of compatibility.
8887 With public keys and the benefit of hindsight one standard format
8888 is used which works with EVP_PKEY, RSA or DSA structures: though
8889 it clearly returns an error if you try to read the wrong kind of key.
8891 Added a -pubkey option to the 'x509' utility to output the public key.
8892 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
8893 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
8894 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
8895 that do the same as the EVP_PKEY_assign_*() except they up the
8896 reference count of the added key (they don't "swallow" the
8900 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
8901 CRLs would fail if the file contained no certificates or no CRLs:
8902 added a new function to read in both types and return the number
8903 read: this means that if none are read it will be an error. The
8904 DER versions of the certificate and CRL reader would always fail
8905 because it isn't possible to mix certificates and CRLs in DER format
8906 without choking one or the other routine. Changed this to just read
8907 a certificate: this is the best we can do. Also modified the code
8908 in apps/verify.c to take notice of return codes: it was previously
8909 attempting to read in certificates from NULL pointers and ignoring
8910 any errors: this is one reason why the cert and CRL reader seemed
8911 to work. It doesn't check return codes from the default certificate
8912 routines: these may well fail if the certificates aren't installed.
8915 *) Code to support otherName option in GeneralName.
8918 *) First update to verify code. Change the verify utility
8919 so it warns if it is passed a self signed certificate:
8920 for consistency with the normal behaviour. X509_verify
8921 has been modified to it will now verify a self signed
8922 certificate if *exactly* the same certificate appears
8923 in the store: it was previously impossible to trust a
8924 single self signed certificate. This means that:
8925 openssl verify ss.pem
8926 now gives a warning about a self signed certificate but
8927 openssl verify -CAfile ss.pem ss.pem
8931 *) For servers, store verify_result in SSL_SESSION data structure
8932 (and add it to external session representation).
8933 This is needed when client certificate verifications fails,
8934 but an application-provided verification callback (set by
8935 SSL_CTX_set_cert_verify_callback) allows accepting the session
8936 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
8937 but returns 1): When the session is reused, we have to set
8938 ssl->verify_result to the appropriate error code to avoid
8940 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
8942 *) Fix a bug in the new PKCS#7 code: it didn't consider the
8943 case in PKCS7_dataInit() where the signed PKCS7 structure
8944 didn't contain any existing data because it was being created.
8945 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
8947 *) Add a salt to the key derivation routines in enc.c. This
8948 forms the first 8 bytes of the encrypted file. Also add a
8949 -S option to allow a salt to be input on the command line.
8952 *) New function X509_cmp(). Oddly enough there wasn't a function
8953 to compare two certificates. We do this by working out the SHA1
8954 hash and comparing that. X509_cmp() will be needed by the trust
8958 *) SSL_get1_session() is like SSL_get_session(), but increments
8959 the reference count in the SSL_SESSION returned.
8960 [Geoff Thorpe <geoff@eu.c2.net>]
8962 *) Fix for 'req': it was adding a null to request attributes.
8963 Also change the X509_LOOKUP and X509_INFO code to handle
8964 certificate auxiliary information.
8967 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8971 *) Add the possibility to add extra information to the memory leak
8972 detecting output, to form tracebacks, showing from where each
8973 allocation was originated: CRYPTO_push_info("constant string") adds
8974 the string plus current file name and line number to a per-thread
8975 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8976 is like calling CYRPTO_pop_info() until the stack is empty.
8977 Also updated memory leak detection code to be multi-thread-safe.
8980 *) Add options -text and -noout to pkcs7 utility and delete the
8981 encryption options which never did anything. Update docs.
8984 *) Add options to some of the utilities to allow the pass phrase
8985 to be included on either the command line (not recommended on
8986 OSes like Unix) or read from the environment. Update the
8987 manpages and fix a few bugs.
8990 *) Add a few manpages for some of the openssl commands.
8993 *) Fix the -revoke option in ca. It was freeing up memory twice,
8994 leaking and not finding already revoked certificates.
8997 *) Extensive changes to support certificate auxiliary information.
8998 This involves the use of X509_CERT_AUX structure and X509_AUX
8999 functions. An X509_AUX function such as PEM_read_X509_AUX()
9000 can still read in a certificate file in the usual way but it
9001 will also read in any additional "auxiliary information". By
9002 doing things this way a fair degree of compatibility can be
9003 retained: existing certificates can have this information added
9004 using the new 'x509' options.
9006 Current auxiliary information includes an "alias" and some trust
9007 settings. The trust settings will ultimately be used in enhanced
9008 certificate chain verification routines: currently a certificate
9009 can only be trusted if it is self signed and then it is trusted
9013 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
9014 The problem was that one of the replacement routines had not been working
9015 since SSLeay releases. For now the offending routine has been replaced
9016 with non-optimised assembler. Even so, this now gives around 95%
9017 performance improvement for 1024 bit RSA signs.
9020 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
9021 handling. Most clients have the effective key size in bits equal to
9022 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
9023 A few however don't do this and instead use the size of the decrypted key
9024 to determine the RC2 key length and the AlgorithmIdentifier to determine
9025 the effective key length. In this case the effective key length can still
9026 be 40 bits but the key length can be 168 bits for example. This is fixed
9027 by manually forcing an RC2 key into the EVP_PKEY structure because the
9028 EVP code can't currently handle unusual RC2 key sizes: it always assumes
9029 the key length and effective key length are equal.
9032 *) Add a bunch of functions that should simplify the creation of
9033 X509_NAME structures. Now you should be able to do:
9034 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
9035 and have it automatically work out the correct field type and fill in
9036 the structures. The more adventurous can try:
9037 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
9038 and it will (hopefully) work out the correct multibyte encoding.
9041 *) Change the 'req' utility to use the new field handling and multibyte
9042 copy routines. Before the DN field creation was handled in an ad hoc
9043 way in req, ca, and x509 which was rather broken and didn't support
9044 BMPStrings or UTF8Strings. Since some software doesn't implement
9045 BMPStrings or UTF8Strings yet, they can be enabled using the config file
9046 using the dirstring_type option. See the new comment in the default
9047 openssl.cnf for more info.
9050 *) Make crypto/rand/md_rand.c more robust:
9051 - Assure unique random numbers after fork().
9052 - Make sure that concurrent threads access the global counter and
9053 md serializably so that we never lose entropy in them
9054 or use exactly the same state in multiple threads.
9055 Access to the large state is not always serializable because
9056 the additional locking could be a performance killer, and
9057 md should be large enough anyway.
9060 *) New file apps/app_rand.c with commonly needed functionality
9061 for handling the random seed file.
9063 Use the random seed file in some applications that previously did not:
9065 dsaparam -genkey (which also ignored its '-rand' option),
9068 x509 (when signing).
9069 Except on systems with /dev/urandom, it is crucial to have a random
9070 seed file at least for key creation, DSA signing, and for DH exchanges;
9071 for RSA signatures we could do without one.
9073 gendh and gendsa (unlike genrsa) used to read only the first byte
9074 of each file listed in the '-rand' option. The function as previously
9075 found in genrsa is now in app_rand.c and is used by all programs
9076 that support '-rand'.
9079 *) In RAND_write_file, use mode 0600 for creating files;
9080 don't just chmod when it may be too late.
9083 *) Report an error from X509_STORE_load_locations
9084 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
9087 *) New function ASN1_mbstring_copy() this copies a string in either
9088 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
9089 into an ASN1_STRING type. A mask of permissible types is passed
9090 and it chooses the "minimal" type to use or an error if not type
9094 *) Add function equivalents to the various macros in asn1.h. The old
9095 macros are retained with an M_ prefix. Code inside the library can
9096 use the M_ macros. External code (including the openssl utility)
9097 should *NOT* in order to be "shared library friendly".
9100 *) Add various functions that can check a certificate's extensions
9101 to see if it usable for various purposes such as SSL client,
9102 server or S/MIME and CAs of these types. This is currently
9103 VERY EXPERIMENTAL but will ultimately be used for certificate chain
9104 verification. Also added a -purpose flag to x509 utility to
9105 print out all the purposes.
9108 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
9112 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
9113 for, obtain and decode and extension and obtain its critical flag.
9114 This allows all the necessary extension code to be handled in a
9115 single function call.
9118 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
9119 platforms. See crypto/rc4/rc4_enc.c for further details.
9122 *) New -noout option to asn1parse. This causes no output to be produced
9123 its main use is when combined with -strparse and -out to extract data
9124 from a file (which may not be in ASN.1 format).
9127 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
9128 when producing the local key id.
9129 [Richard Levitte <levitte@stacken.kth.se>]
9131 *) New option -dhparam in s_server. This allows a DH parameter file to be
9132 stated explicitly. If it is not stated then it tries the first server
9133 certificate file. The previous behaviour hard coded the filename
9137 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
9138 a public key to be input or output. For example:
9139 openssl rsa -in key.pem -pubout -out pubkey.pem
9140 Also added necessary DSA public key functions to handle this.
9143 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
9144 in the message. This was handled by allowing
9145 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
9146 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
9148 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
9149 to the end of the strings whereas this didn't. This would cause problems
9150 if strings read with d2i_ASN1_bytes() were later modified.
9151 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
9153 *) Fix for base64 decode bug. When a base64 bio reads only one line of
9154 data and it contains EOF it will end up returning an error. This is
9155 caused by input 46 bytes long. The cause is due to the way base64
9156 BIOs find the start of base64 encoded data. They do this by trying a
9157 trial decode on each line until they find one that works. When they
9158 do a flag is set and it starts again knowing it can pass all the
9159 data directly through the decoder. Unfortunately it doesn't reset
9160 the context it uses. This means that if EOF is reached an attempt
9161 is made to pass two EOFs through the context and this causes the
9162 resulting error. This can also cause other problems as well. As is
9163 usual with these problems it takes *ages* to find and the fix is
9164 trivial: move one line.
9165 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
9167 *) Ugly workaround to get s_client and s_server working under Windows. The
9168 old code wouldn't work because it needed to select() on sockets and the
9169 tty (for keypresses and to see if data could be written). Win32 only
9170 supports select() on sockets so we select() with a 1s timeout on the
9171 sockets and then see if any characters are waiting to be read, if none
9172 are present then we retry, we also assume we can always write data to
9173 the tty. This isn't nice because the code then blocks until we've
9174 received a complete line of data and it is effectively polling the
9175 keyboard at 1s intervals: however it's quite a bit better than not
9176 working at all :-) A dedicated Windows application might handle this
9177 with an event loop for example.
9180 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
9181 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
9182 will be called when RSA_sign() and RSA_verify() are used. This is useful
9183 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
9184 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
9185 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
9186 This necessitated the support of an extra signature type NID_md5_sha1
9187 for SSL signatures and modifications to the SSL library to use it instead
9188 of calling RSA_public_decrypt() and RSA_private_encrypt().
9191 *) Add new -verify -CAfile and -CApath options to the crl program, these
9192 will lookup a CRL issuers certificate and verify the signature in a
9193 similar way to the verify program. Tidy up the crl program so it
9194 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
9195 less strict. It will now permit CRL extensions even if it is not
9196 a V2 CRL: this will allow it to tolerate some broken CRLs.
9199 *) Initialize all non-automatic variables each time one of the openssl
9200 sub-programs is started (this is necessary as they may be started
9201 multiple times from the "OpenSSL>" prompt).
9202 [Lennart Bang, Bodo Moeller]
9204 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
9205 removing all other RSA functionality (this is what NO_RSA does). This
9206 is so (for example) those in the US can disable those operations covered
9207 by the RSA patent while allowing storage and parsing of RSA keys and RSA
9211 *) Non-copying interface to BIO pairs.
9212 (still largely untested)
9215 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
9216 ASCII string. This was handled independently in various places before.
9219 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
9220 UTF8 strings a character at a time.
9223 *) Use client_version from client hello to select the protocol
9224 (s23_srvr.c) and for RSA client key exchange verification
9225 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
9228 *) Add various utility functions to handle SPKACs, these were previously
9229 handled by poking round in the structure internals. Added new function
9230 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
9231 print, verify and generate SPKACs. Based on an original idea from
9232 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
9235 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
9238 *) Allow the config file extension section to be overwritten on the
9239 command line. Based on an original idea from Massimiliano Pala
9240 <madwolf@comune.modena.it>. The new option is called -extensions
9241 and can be applied to ca, req and x509. Also -reqexts to override
9242 the request extensions in req and -crlexts to override the crl extensions
9246 *) Add new feature to the SPKAC handling in ca. Now you can include
9247 the same field multiple times by preceding it by "XXXX." for example:
9250 this is the same syntax as used in the req config file.
9253 *) Allow certificate extensions to be added to certificate requests. These
9254 are specified in a 'req_extensions' option of the req section of the
9255 config file. They can be printed out with the -text option to req but
9256 are otherwise ignored at present.
9259 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
9260 data read consists of only the final block it would not decrypted because
9261 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
9262 A misplaced 'break' also meant the decrypted final block might not be
9263 copied until the next read.
9266 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
9267 a few extra parameters to the DH structure: these will be useful if
9268 for example we want the value of 'q' or implement X9.42 DH.
9271 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
9272 provides hooks that allow the default DSA functions or functions on a
9273 "per key" basis to be replaced. This allows hardware acceleration and
9274 hardware key storage to be handled without major modification to the
9275 library. Also added low level modexp hooks and CRYPTO_EX structure and
9276 associated functions.
9279 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
9280 as "read only": it can't be written to and the buffer it points to will
9281 not be freed. Reading from a read only BIO is much more efficient than
9282 a normal memory BIO. This was added because there are several times when
9283 an area of memory needs to be read from a BIO. The previous method was
9284 to create a memory BIO and write the data to it, this results in two
9285 copies of the data and an O(n^2) reading algorithm. There is a new
9286 function BIO_new_mem_buf() which creates a read only memory BIO from
9287 an area of memory. Also modified the PKCS#7 routines to use read only
9291 *) Bugfix: ssl23_get_client_hello did not work properly when called in
9292 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
9293 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
9294 but a retry condition occured while trying to read the rest.
9297 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
9298 NID_pkcs7_encrypted by default: this was wrong since this should almost
9299 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
9300 the encrypted data type: this is a more sensible place to put it and it
9301 allows the PKCS#12 code to be tidied up that duplicated this
9305 *) Changed obj_dat.pl script so it takes its input and output files on
9306 the command line. This should avoid shell escape redirection problems
9310 *) Initial support for certificate extension requests, these are included
9311 in things like Xenroll certificate requests. Included functions to allow
9312 extensions to be obtained and added.
9315 *) -crlf option to s_client and s_server for sending newlines as
9316 CRLF (as required by many protocols).
9319 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
9321 *) Install libRSAglue.a when OpenSSL is built with RSAref.
9322 [Ralf S. Engelschall]
9324 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
9325 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
9327 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
9331 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
9332 DH parameters/keys (q is lost during that conversion, but the resulting
9333 DH parameters contain its length).
9335 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
9336 much faster than DH_generate_parameters (which creates parameters
9337 where p = 2*q + 1), and also the smaller q makes DH computations
9338 much more efficient (160-bit exponentiation instead of 1024-bit
9339 exponentiation); so this provides a convenient way to support DHE
9340 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
9341 utter importance to use
9342 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9344 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9345 when such DH parameters are used, because otherwise small subgroup
9346 attacks may become possible!
9349 *) Avoid memory leak in i2d_DHparams.
9352 *) Allow the -k option to be used more than once in the enc program:
9353 this allows the same encrypted message to be read by multiple recipients.
9356 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
9357 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
9358 it will always use the numerical form of the OID, even if it has a short
9362 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
9363 method only got called if p,q,dmp1,dmq1,iqmp components were present,
9364 otherwise bn_mod_exp was called. In the case of hardware keys for example
9365 no private key components need be present and it might store extra data
9366 in the RSA structure, which cannot be accessed from bn_mod_exp.
9367 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
9368 private key operations.
9371 *) Added support for SPARC Linux.
9374 *) pem_password_cb function type incompatibly changed from
9375 typedef int pem_password_cb(char *buf, int size, int rwflag);
9377 ....(char *buf, int size, int rwflag, void *userdata);
9378 so that applications can pass data to their callbacks:
9379 The PEM[_ASN1]_{read,write}... functions and macros now take an
9380 additional void * argument, which is just handed through whenever
9381 the password callback is called.
9382 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
9384 New function SSL_CTX_set_default_passwd_cb_userdata.
9386 Compatibility note: As many C implementations push function arguments
9387 onto the stack in reverse order, the new library version is likely to
9388 interoperate with programs that have been compiled with the old
9389 pem_password_cb definition (PEM_whatever takes some data that
9390 happens to be on the stack as its last argument, and the callback
9391 just ignores this garbage); but there is no guarantee whatsoever that
9394 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
9395 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
9396 problems not only on Windows, but also on some Unix platforms.
9397 To avoid problematic command lines, these definitions are now in an
9398 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
9399 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
9402 *) MIPS III/IV assembler module is reimplemented.
9405 *) More DES library cleanups: remove references to srand/rand and
9406 delete an unused file.
9409 *) Add support for the the free Netwide assembler (NASM) under Win32,
9410 since not many people have MASM (ml) and it can be hard to obtain.
9411 This is currently experimental but it seems to work OK and pass all
9412 the tests. Check out INSTALL.W32 for info.
9415 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
9416 without temporary keys kept an extra copy of the server key,
9417 and connections with temporary keys did not free everything in case
9421 *) New function RSA_check_key and new openssl rsa option -check
9422 for verifying the consistency of RSA keys.
9423 [Ulf Moeller, Bodo Moeller]
9425 *) Various changes to make Win32 compile work:
9426 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
9427 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
9428 comparison" warnings.
9429 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
9432 *) Add a debugging option to PKCS#5 v2 key generation function: when
9433 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
9434 derived keys are printed to stderr.
9437 *) Copy the flags in ASN1_STRING_dup().
9438 [Roman E. Pavlov <pre@mo.msk.ru>]
9440 *) The x509 application mishandled signing requests containing DSA
9441 keys when the signing key was also DSA and the parameters didn't match.
9443 It was supposed to omit the parameters when they matched the signing key:
9444 the verifying software was then supposed to automatically use the CA's
9445 parameters if they were absent from the end user certificate.
9447 Omitting parameters is no longer recommended. The test was also
9448 the wrong way round! This was probably due to unusual behaviour in
9449 EVP_cmp_parameters() which returns 1 if the parameters match.
9450 This meant that parameters were omitted when they *didn't* match and
9451 the certificate was useless. Certificates signed with 'ca' didn't have
9453 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
9455 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
9456 The interface is as follows:
9457 Applications can use
9458 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
9459 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
9460 "off" is now the default.
9461 The library internally uses
9462 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
9463 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
9464 to disable memory-checking temporarily.
9466 Some inconsistent states that previously were possible (and were
9467 even the default) are now avoided.
9469 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
9470 with each memory chunk allocated; this is occasionally more helpful
9471 than just having a counter.
9473 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
9475 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
9479 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
9480 which largely parallels "options", but is for changing API behaviour,
9481 whereas "options" are about protocol behaviour.
9482 Initial "mode" flags are:
9484 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
9485 a single record has been written.
9486 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
9487 retries use the same buffer location.
9488 (But all of the contents must be
9492 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
9495 *) Fix problems with no-hmac etc.
9496 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
9498 *) New functions RSA_get_default_method(), RSA_set_method() and
9499 RSA_get_method(). These allows replacement of RSA_METHODs without having
9500 to mess around with the internals of an RSA structure.
9503 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
9504 Also really enable memory leak checks in openssl.c and in some
9506 [Chad C. Mulligan, Bodo Moeller]
9508 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
9509 up the length of negative integers. This has now been simplified to just
9510 store the length when it is first determined and use it later, rather
9511 than trying to keep track of where data is copied and updating it to
9513 [Steve Henson, reported by Brien Wheeler
9514 <bwheeler@authentica-security.com>]
9516 *) Add a new function PKCS7_signatureVerify. This allows the verification
9517 of a PKCS#7 signature but with the signing certificate passed to the
9518 function itself. This contrasts with PKCS7_dataVerify which assumes the
9519 certificate is present in the PKCS#7 structure. This isn't always the
9520 case: certificates can be omitted from a PKCS#7 structure and be
9521 distributed by "out of band" means (such as a certificate database).
9524 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
9525 function prototypes in pem.h, also change util/mkdef.pl to add the
9526 necessary function names.
9529 *) mk1mf.pl (used by Windows builds) did not properly read the
9530 options set by Configure in the top level Makefile, and Configure
9531 was not even able to write more than one option correctly.
9532 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
9535 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
9536 file to be loaded from a BIO or FILE pointer. The BIO version will
9537 for example allow memory BIOs to contain config info.
9540 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
9541 Whoever hopes to achieve shared-library compatibility across versions
9542 must use this, not the compile-time macro.
9543 (Exercise 0.9.4: Which is the minimum library version required by
9545 Note: All this applies only to multi-threaded programs, others don't
9549 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
9550 through a BIO pair triggered the default case, i.e.
9551 SSLerr(...,SSL_R_UNKNOWN_STATE).
9554 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
9555 can use the SSL library even if none of the specific BIOs is
9559 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
9560 for the encoded length.
9561 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
9563 *) Add initial documentation of the X509V3 functions.
9566 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
9567 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
9568 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
9569 secure PKCS#8 private key format with a high iteration count.
9572 *) Fix determination of Perl interpreter: A perl or perl5
9573 _directory_ in $PATH was also accepted as the interpreter.
9574 [Ralf S. Engelschall]
9576 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
9577 wrong with it but it was very old and did things like calling
9578 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
9582 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
9583 to use the new extension code.
9586 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
9587 with macros. This should make it easier to change their form, add extra
9588 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
9592 *) Add to configuration table a new entry that can specify an alternative
9593 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
9594 according to Mark Crispin <MRC@Panda.COM>.
9598 *) DES CBC did not update the IV. Weird.
9601 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
9602 Changing the behaviour of the former might break existing programs --
9603 where IV updating is needed, des_ncbc_encrypt can be used.
9606 *) When bntest is run from "make test" it drives bc to check its
9607 calculations, as well as internally checking them. If an internal check
9608 fails, it needs to cause bc to give a non-zero result or make test carries
9609 on without noticing the failure. Fixed.
9612 *) DES library cleanups.
9615 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
9616 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
9617 ciphers. NOTE: although the key derivation function has been verified
9618 against some published test vectors it has not been extensively tested
9619 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
9623 *) Instead of "mkdir -p", which is not fully portable, use new
9624 Perl script "util/mkdir-p.pl".
9627 *) Rewrite the way password based encryption (PBE) is handled. It used to
9628 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
9629 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
9630 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
9631 the 'parameter' field of the AlgorithmIdentifier is passed to the
9632 underlying key generation function so it must do its own ASN1 parsing.
9633 This has also changed the EVP_PBE_CipherInit() function which now has a
9634 'parameter' argument instead of literal salt and iteration count values
9635 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
9638 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
9639 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
9640 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
9641 KEY" because this clashed with PKCS#8 unencrypted string. Since this
9642 value was just used as a "magic string" and not used directly its
9643 value doesn't matter.
9646 *) Introduce some semblance of const correctness to BN. Shame C doesn't
9650 *) "linux-sparc64" configuration (ultrapenguin).
9651 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
9652 "linux-sparc" configuration.
9653 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
9655 *) config now generates no-xxx options for missing ciphers.
9658 *) Support the EBCDIC character set (work in progress).
9659 File ebcdic.c not yet included because it has a different license.
9660 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9662 *) Support BS2000/OSD-POSIX.
9663 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9665 *) Make callbacks for key generation use void * instead of char *.
9668 *) Make S/MIME samples compile (not yet tested).
9671 *) Additional typesafe stacks.
9674 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
9678 Changes between 0.9.3 and 0.9.3a [29 May 1999]
9680 *) New configuration variant "sco5-gcc".
9682 *) Updated some demos.
9683 [Sean O Riordain, Wade Scholine]
9685 *) Add missing BIO_free at exit of pkcs12 application.
9688 *) Fix memory leak in conf.c.
9691 *) Updates for Win32 to assembler version of MD5.
9694 *) Set #! path to perl in apps/der_chop to where we found it
9695 instead of using a fixed path.
9698 *) SHA library changes for irix64-mips4-cc.
9701 *) Improvements for VMS support.
9705 Changes between 0.9.2b and 0.9.3 [24 May 1999]
9707 *) Bignum library bug fix. IRIX 6 passes "make test" now!
9708 This also avoids the problems with SC4.2 and unpatched SC5.
9709 [Andy Polyakov <appro@fy.chalmers.se>]
9711 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
9712 These are required because of the typesafe stack would otherwise break
9713 existing code. If old code used a structure member which used to be STACK
9714 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
9715 sk_num or sk_value it would produce an error because the num, data members
9716 are not present in STACK_OF. Now it just produces a warning. sk_set
9717 replaces the old method of assigning a value to sk_value
9718 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
9719 that does this will no longer work (and should use sk_set instead) but
9720 this could be regarded as a "questionable" behaviour anyway.
9723 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
9724 correctly handle encrypted S/MIME data.
9727 *) Change type of various DES function arguments from des_cblock
9728 (which means, in function argument declarations, pointer to char)
9729 to des_cblock * (meaning pointer to array with 8 char elements),
9730 which allows the compiler to do more typechecking; it was like
9731 that back in SSLeay, but with lots of ugly casts.
9733 Introduce new type const_des_cblock.
9736 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
9737 problems: find RecipientInfo structure that matches recipient certificate
9738 and initialise the ASN1 structures properly based on passed cipher.
9741 *) Belatedly make the BN tests actually check the results.
9744 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
9745 to and from BNs: it was completely broken. New compilation option
9746 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
9747 key elements as negative integers.
9750 *) Reorganize and speed up MD5.
9751 [Andy Polyakov <appro@fy.chalmers.se>]
9754 [Richard Levitte <richard@levitte.org>]
9756 *) New option -out to asn1parse to allow the parsed structure to be
9757 output to a file. This is most useful when combined with the -strparse
9758 option to examine the output of things like OCTET STRINGS.
9761 *) Make SSL library a little more fool-proof by not requiring any longer
9762 that SSL_set_{accept,connect}_state be called before
9763 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
9764 in many applications because usually everything *appeared* to work as
9765 intended anyway -- now it really works as intended).
9768 *) Move openssl.cnf out of lib/.
9771 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
9772 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
9773 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
9774 [Ralf S. Engelschall]
9776 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
9777 handle PKCS#7 enveloped data properly.
9778 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
9780 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
9781 copying pointers. The cert_st handling is changed by this in
9782 various ways (and thus what used to be known as ctx->default_cert
9783 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
9784 any longer when s->cert does not give us what we need).
9785 ssl_cert_instantiate becomes obsolete by this change.
9786 As soon as we've got the new code right (possibly it already is?),
9787 we have solved a couple of bugs of the earlier code where s->cert
9788 was used as if it could not have been shared with other SSL structures.
9790 Note that using the SSL API in certain dirty ways now will result
9791 in different behaviour than observed with earlier library versions:
9792 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
9793 does not influence s as it used to.
9795 In order to clean up things more thoroughly, inside SSL_SESSION
9796 we don't use CERT any longer, but a new structure SESS_CERT
9797 that holds per-session data (if available); currently, this is
9798 the peer's certificate chain and, for clients, the server's certificate
9799 and temporary key. CERT holds only those values that can have
9800 meaningful defaults in an SSL_CTX.
9803 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
9804 from the internal representation. Various PKCS#7 fixes: remove some
9805 evil casts and set the enc_dig_alg field properly based on the signing
9809 *) Allow PKCS#12 password to be set from the command line or the
9810 environment. Let 'ca' get its config file name from the environment
9811 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
9815 *) Allow certificate policies extension to use an IA5STRING for the
9816 organization field. This is contrary to the PKIX definition but
9817 VeriSign uses it and IE5 only recognises this form. Document 'x509'
9821 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
9822 without disallowing inline assembler and the like for non-pedantic builds.
9825 *) Support Borland C++ builder.
9826 [Janez Jere <jj@void.si>, modified by Ulf Möller]
9831 *) SHA-1 cleanups and performance enhancements.
9832 [Andy Polyakov <appro@fy.chalmers.se>]
9834 *) Sparc v8plus assembler for the bignum library.
9835 [Andy Polyakov <appro@fy.chalmers.se>]
9837 *) Accept any -xxx and +xxx compiler options in Configure.
9840 *) Update HPUX configuration.
9843 *) Add missing sk_<type>_unshift() function to safestack.h
9844 [Ralf S. Engelschall]
9846 *) New function SSL_CTX_use_certificate_chain_file that sets the
9847 "extra_cert"s in addition to the certificate. (This makes sense
9848 only for "PEM" format files, as chains as a whole are not
9852 *) Support verify_depth from the SSL API.
9853 x509_vfy.c had what can be considered an off-by-one-error:
9854 Its depth (which was not part of the external interface)
9855 was actually counting the number of certificates in a chain;
9856 now it really counts the depth.
9859 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
9860 instead of X509err, which often resulted in confusing error
9861 messages since the error codes are not globally unique
9862 (e.g. an alleged error in ssl3_accept when a certificate
9863 didn't match the private key).
9865 *) New function SSL_CTX_set_session_id_context that allows to set a default
9866 value (so that you don't need SSL_set_session_id_context for each
9867 connection using the SSL_CTX).
9870 *) OAEP decoding bug fix.
9873 *) Support INSTALL_PREFIX for package builders, as proposed by
9877 *) New Configure options "threads" and "no-threads". For systems
9878 where the proper compiler options are known (currently Solaris
9879 and Linux), "threads" is the default.
9882 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
9885 *) Install various scripts to $(OPENSSLDIR)/misc, not to
9886 $(INSTALLTOP)/bin -- they shouldn't clutter directories
9887 such as /usr/local/bin.
9890 *) "make linux-shared" to build shared libraries.
9891 [Niels Poppe <niels@netbox.org>]
9893 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
9896 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
9897 extension adding in x509 utility.
9900 *) Remove NOPROTO sections and error code comments.
9903 *) Partial rewrite of the DEF file generator to now parse the ANSI
9907 *) New Configure options --prefix=DIR and --openssldir=DIR.
9910 *) Complete rewrite of the error code script(s). It is all now handled
9911 by one script at the top level which handles error code gathering,
9912 header rewriting and C source file generation. It should be much better
9913 than the old method: it now uses a modified version of Ulf's parser to
9914 read the ANSI prototypes in all header files (thus the old K&R definitions
9915 aren't needed for error creation any more) and do a better job of
9916 translating function codes into names. The old 'ASN1 error code imbedded
9917 in a comment' is no longer necessary and it doesn't use .err files which
9918 have now been deleted. Also the error code call doesn't have to appear all
9919 on one line (which resulted in some large lines...).
9922 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
9925 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
9926 0 (which usually indicates a closed connection), but continue reading.
9929 *) Fix some race conditions.
9932 *) Add support for CRL distribution points extension. Add Certificate
9933 Policies and CRL distribution points documentation.
9936 *) Move the autogenerated header file parts to crypto/opensslconf.h.
9939 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
9940 8 of keying material. Merlin has also confirmed interop with this fix
9941 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
9942 [Merlin Hughes <merlin@baltimore.ie>]
9944 *) Fix lots of warnings.
9945 [Richard Levitte <levitte@stacken.kth.se>]
9947 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
9948 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
9949 [Richard Levitte <levitte@stacken.kth.se>]
9951 *) Fix problems with sizeof(long) == 8.
9952 [Andy Polyakov <appro@fy.chalmers.se>]
9954 *) Change functions to ANSI C.
9957 *) Fix typos in error codes.
9958 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
9960 *) Remove defunct assembler files from Configure.
9963 *) SPARC v8 assembler BIGNUM implementation.
9964 [Andy Polyakov <appro@fy.chalmers.se>]
9966 *) Support for Certificate Policies extension: both print and set.
9967 Various additions to support the r2i method this uses.
9970 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9971 return a const string when you are expecting an allocated buffer.
9974 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9975 types DirectoryString and DisplayText.
9978 *) Add code to allow r2i extensions to access the configuration database,
9979 add an LHASH database driver and add several ctx helper functions.
9982 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9983 fail when they extended the size of a BIGNUM.
9986 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9987 support typesafe stack.
9990 *) Fix typo in SSL_[gs]et_options().
9991 [Nils Frostberg <nils@medcom.se>]
9993 *) Delete various functions and files that belonged to the (now obsolete)
9994 old X509V3 handling code.
9997 *) New Configure option "rsaref".
10000 *) Don't auto-generate pem.h.
10003 *) Introduce type-safe ASN.1 SETs.
10006 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
10007 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
10009 *) Introduce type-safe STACKs. This will almost certainly break lots of code
10010 that links with OpenSSL (well at least cause lots of warnings), but fear
10011 not: the conversion is trivial, and it eliminates loads of evil casts. A
10012 few STACKed things have been converted already. Feel free to convert more.
10013 In the fullness of time, I'll do away with the STACK type altogether.
10016 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
10017 specified in <certfile> by updating the entry in the index.txt file.
10018 This way one no longer has to edit the index.txt file manually for
10019 revoking a certificate. The -revoke option does the gory details now.
10020 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
10022 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
10023 `-text' option at all and this way the `-noout -text' combination was
10024 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
10025 [Ralf S. Engelschall]
10027 *) Make sure a corresponding plain text error message exists for the
10028 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
10029 verify callback function determined that a certificate was revoked.
10030 [Ralf S. Engelschall]
10032 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
10033 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
10034 all available cipers including rc5, which was forgotten until now.
10035 In order to let the testing shell script know which algorithms
10036 are available, a new (up to now undocumented) command
10037 "openssl list-cipher-commands" is used.
10040 *) Bugfix: s_client occasionally would sleep in select() when
10041 it should have checked SSL_pending() first.
10044 *) New functions DSA_do_sign and DSA_do_verify to provide access to
10045 the raw DSA values prior to ASN.1 encoding.
10048 *) Tweaks to Configure
10049 [Niels Poppe <niels@netbox.org>]
10051 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
10055 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
10058 *) New config option to avoid instructions that are illegal on the 80386.
10059 The default code is faster, but requires at least a 486.
10062 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
10063 SSL2_SERVER_VERSION (not used at all) macros, which are now the
10064 same as SSL2_VERSION anyway.
10067 *) New "-showcerts" option for s_client.
10070 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
10071 application. Various cleanups and fixes.
10074 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
10075 modify error routines to work internally. Add error codes and PBE init
10076 to library startup routines.
10079 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
10080 packing functions to asn1 and evp. Changed function names and error
10081 codes along the way.
10084 *) PKCS12 integration: and so it begins... First of several patches to
10085 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
10086 objects to objects.h
10089 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
10090 and display support for Thawte strong extranet extension.
10093 *) Add LinuxPPC support.
10094 [Jeff Dubrule <igor@pobox.org>]
10096 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
10097 bn_div_words in alpha.s.
10098 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
10100 *) Make sure the RSA OAEP test is skipped under -DRSAref because
10101 OAEP isn't supported when OpenSSL is built with RSAref.
10102 [Ulf Moeller <ulf@fitug.de>]
10104 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
10105 so they no longer are missing under -DNOPROTO.
10106 [Soren S. Jorvang <soren@t.dk>]
10109 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
10111 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
10112 doesn't work when the session is reused. Coming soon!
10115 *) Fix a security hole, that allows sessions to be reused in the wrong
10116 context thus bypassing client cert protection! All software that uses
10117 client certs and session caches in multiple contexts NEEDS PATCHING to
10118 allow session reuse! A fuller solution is in the works.
10119 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
10121 *) Some more source tree cleanups (removed obsolete files
10122 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
10123 permission on "config" script to be executable) and a fix for the INSTALL
10125 [Ulf Moeller <ulf@fitug.de>]
10127 *) Remove some legacy and erroneous uses of malloc, free instead of
10129 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
10131 *) Make rsa_oaep_test return non-zero on error.
10132 [Ulf Moeller <ulf@fitug.de>]
10134 *) Add support for native Solaris shared libraries. Configure
10135 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
10136 if someone would make that last step automatic.
10137 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
10139 *) ctx_size was not built with the right compiler during "make links". Fixed.
10142 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
10143 except NULL ciphers". This means the default cipher list will no longer
10144 enable NULL ciphers. They need to be specifically enabled e.g. with
10145 the string "DEFAULT:eNULL".
10148 *) Fix to RSA private encryption routines: if p < q then it would
10149 occasionally produce an invalid result. This will only happen with
10150 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
10153 *) Be less restrictive and allow also `perl util/perlpath.pl
10154 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
10155 because this way one can also use an interpreter named `perl5' (which is
10156 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
10157 installed as `perl').
10158 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10160 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
10161 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10163 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
10164 advapi32.lib to Win32 build and change the pem test comparision
10165 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
10166 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
10167 and crypto/des/ede_cbcm_enc.c.
10170 *) DES quad checksum was broken on big-endian architectures. Fixed.
10173 *) Comment out two functions in bio.h that aren't implemented. Fix up the
10174 Win32 test batch file so it (might) work again. The Win32 test batch file
10175 is horrible: I feel ill....
10178 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
10179 in e_os.h. Audit of header files to check ANSI and non ANSI
10180 sections: 10 functions were absent from non ANSI section and not exported
10181 from Windows DLLs. Fixed up libeay.num for new functions.
10184 *) Make `openssl version' output lines consistent.
10185 [Ralf S. Engelschall]
10187 *) Fix Win32 symbol export lists for BIO functions: Added
10188 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
10189 to ms/libeay{16,32}.def.
10190 [Ralf S. Engelschall]
10192 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
10193 fine under Unix and passes some trivial tests I've now added. But the
10194 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
10195 added to make sure no one expects that this stuff really works in the
10196 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
10197 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
10199 [Ralf S. Engelschall]
10201 *) Fix the generation of two part addresses in perl.
10202 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
10204 *) Add config entry for Linux on MIPS.
10205 [John Tobey <jtobey@channel1.com>]
10207 *) Make links whenever Configure is run, unless we are on Windoze.
10210 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
10211 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
10215 *) Add a useful kludge to allow package maintainers to specify compiler and
10216 other platforms details on the command line without having to patch the
10217 Configure script everytime: One now can use ``perl Configure
10218 <id>:<details>'', i.e. platform ids are allowed to have details appended
10219 to them (seperated by colons). This is treated as there would be a static
10220 pre-configured entry in Configure's %table under key <id> with value
10221 <details> and ``perl Configure <id>'' is called. So, when you want to
10222 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
10223 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
10224 now, which overrides the FreeBSD-elf entry on-the-fly.
10225 [Ralf S. Engelschall]
10227 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
10230 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
10231 on the `perl Configure ...' command line. This way one can compile
10232 OpenSSL libraries with Position Independent Code (PIC) which is needed
10233 for linking it into DSOs.
10234 [Ralf S. Engelschall]
10236 *) Remarkably, export ciphers were totally broken and no-one had noticed!
10240 *) Cleaned up the LICENSE document: The official contact for any license
10241 questions now is the OpenSSL core team under openssl-core@openssl.org.
10242 And add a paragraph about the dual-license situation to make sure people
10243 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
10244 to the OpenSSL toolkit.
10245 [Ralf S. Engelschall]
10247 *) General source tree makefile cleanups: Made `making xxx in yyy...'
10248 display consistent in the source tree and replaced `/bin/rm' by `rm'.
10249 Additonally cleaned up the `make links' target: Remove unnecessary
10250 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
10251 to speed processing and no longer clutter the display with confusing
10252 stuff. Instead only the actually done links are displayed.
10253 [Ralf S. Engelschall]
10255 *) Permit null encryption ciphersuites, used for authentication only. It used
10256 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
10257 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
10261 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
10262 signed attributes when verifying signatures (this would break them),
10263 the detached data encoding was wrong and public keys obtained using
10264 X509_get_pubkey() weren't freed.
10267 *) Add text documentation for the BUFFER functions. Also added a work around
10268 to a Win95 console bug. This was triggered by the password read stuff: the
10269 last character typed gets carried over to the next fread(). If you were
10270 generating a new cert request using 'req' for example then the last
10271 character of the passphrase would be CR which would then enter the first
10275 *) Added the new `Includes OpenSSL Cryptography Software' button as
10276 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
10277 button and can be used by applications based on OpenSSL to show the
10278 relationship to the OpenSSL project.
10279 [Ralf S. Engelschall]
10281 *) Remove confusing variables in function signatures in files
10282 ssl/ssl_lib.c and ssl/ssl.h.
10283 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10285 *) Don't install bss_file.c under PREFIX/include/
10286 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10288 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
10289 functions that return function pointers and has support for NT specific
10290 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
10291 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
10292 unsigned to signed types: this was killing the Win32 compile.
10295 *) Add new certificate file to stack functions,
10296 SSL_add_dir_cert_subjects_to_stack() and
10297 SSL_add_file_cert_subjects_to_stack(). These largely supplant
10298 SSL_load_client_CA_file(), and can be used to add multiple certs easily
10299 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
10300 This means that Apache-SSL and similar packages don't have to mess around
10301 to add as many CAs as they want to the preferred list.
10304 *) Experiment with doxygen documentation. Currently only partially applied to
10306 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
10307 openssl.doxy as the configuration file.
10310 *) Get rid of remaining C++-style comments which strict C compilers hate.
10311 [Ralf S. Engelschall, pointed out by Carlos Amengual]
10313 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
10314 compiled in by default: it has problems with large keys.
10317 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
10318 DH private keys and/or callback functions which directly correspond to
10319 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
10320 is needed for applications which have to configure certificates on a
10321 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
10323 For the RSA certificate situation is makes no difference, but
10324 for the DSA certificate situation this fixes the "no shared cipher"
10325 problem where the OpenSSL cipher selection procedure failed because the
10326 temporary keys were not overtaken from the context and the API provided
10327 no way to reconfigure them.
10328 The new functions now let applications reconfigure the stuff and they
10329 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
10330 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
10331 non-public-API function ssl_cert_instantiate() is used as a helper
10332 function and also to reduce code redundancy inside ssl_rsa.c.
10333 [Ralf S. Engelschall]
10335 *) Move s_server -dcert and -dkey options out of the undocumented feature
10336 area because they are useful for the DSA situation and should be
10337 recognized by the users.
10338 [Ralf S. Engelschall]
10340 *) Fix the cipher decision scheme for export ciphers: the export bits are
10341 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
10342 SSL_EXP_MASK. So, the original variable has to be used instead of the
10343 already masked variable.
10344 [Richard Levitte <levitte@stacken.kth.se>]
10346 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
10347 [Richard Levitte <levitte@stacken.kth.se>]
10349 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
10350 from `int' to `unsigned int' because it's a length and initialized by
10351 EVP_DigestFinal() which expects an `unsigned int *'.
10352 [Richard Levitte <levitte@stacken.kth.se>]
10354 *) Don't hard-code path to Perl interpreter on shebang line of Configure
10355 script. Instead use the usual Shell->Perl transition trick.
10356 [Ralf S. Engelschall]
10358 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
10359 (in addition to RSA certificates) to match the behaviour of `openssl dsa
10360 -noout -modulus' as it's already the case for `openssl rsa -noout
10361 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
10362 currently the public key is printed (a decision which was already done by
10363 `openssl dsa -modulus' in the past) which serves a similar purpose.
10364 Additionally the NO_RSA no longer completely removes the whole -modulus
10365 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
10367 [Ralf S. Engelschall]
10369 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
10370 BIO. See the source (crypto/evp/bio_ok.c) for more info.
10371 [Arne Ansper <arne@ats.cyber.ee>]
10373 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
10374 to be added. Now both 'req' and 'ca' can use new objects defined in the
10378 *) Add cool BIO that does syslog (or event log on NT).
10379 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
10381 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
10382 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
10383 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
10384 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
10387 *) Add preliminary config info for new extension code.
10390 *) Make RSA_NO_PADDING really use no padding.
10391 [Ulf Moeller <ulf@fitug.de>]
10393 *) Generate errors when private/public key check is done.
10396 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
10397 for some CRL extensions and new objects added.
10400 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
10401 key usage extension and fuller support for authority key id.
10404 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
10405 padding method for RSA, which is recommended for new applications in PKCS
10406 #1 v2.0 (RFC 2437, October 1998).
10407 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
10408 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
10409 against Bleichbacher's attack on RSA.
10410 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
10413 *) Updates to the new SSL compression code
10414 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10416 *) Fix so that the version number in the master secret, when passed
10417 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
10418 (because the server will not accept higher), that the version number
10419 is 0x03,0x01, not 0x03,0x00
10420 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10422 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
10423 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
10424 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
10427 *) Support for RAW extensions where an arbitrary extension can be
10428 created by including its DER encoding. See apps/openssl.cnf for
10432 *) Make sure latest Perl versions don't interpret some generated C array
10433 code as Perl array code in the crypto/err/err_genc.pl script.
10434 [Lars Weber <3weber@informatik.uni-hamburg.de>]
10436 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
10437 not many people have the assembler. Various Win32 compilation fixes and
10438 update to the INSTALL.W32 file with (hopefully) more accurate Win32
10439 build instructions.
10442 *) Modify configure script 'Configure' to automatically create crypto/date.h
10443 file under Win32 and also build pem.h from pem.org. New script
10444 util/mkfiles.pl to create the MINFO file on environments that can't do a
10445 'make files': perl util/mkfiles.pl >MINFO should work.
10448 *) Major rework of DES function declarations, in the pursuit of correctness
10449 and purity. As a result, many evil casts evaporated, and some weirdness,
10450 too. You may find this causes warnings in your code. Zapping your evil
10451 casts will probably fix them. Mostly.
10454 *) Fix for a typo in asn1.h. Bug fix to object creation script
10455 obj_dat.pl. It considered a zero in an object definition to mean
10456 "end of object": none of the objects in objects.h have any zeros
10457 so it wasn't spotted.
10458 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
10460 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
10461 Masking (CBCM). In the absence of test vectors, the best I have been able
10462 to do is check that the decrypt undoes the encrypt, so far. Send me test
10463 vectors if you have them.
10466 *) Correct calculation of key length for export ciphers (too much space was
10467 allocated for null ciphers). This has not been tested!
10470 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
10471 message is now correct (it understands "crypto" and "ssl" on its
10472 command line). There is also now an "update" option. This will update
10473 the util/ssleay.num and util/libeay.num files with any new functions.
10475 perl util/mkdef.pl crypto ssl update
10476 it will update them.
10479 *) Overhauled the Perl interface (perl/*):
10480 - ported BN stuff to OpenSSL's different BN library
10481 - made the perl/ source tree CVS-aware
10482 - renamed the package from SSLeay to OpenSSL (the files still contain
10483 their history because I've copied them in the repository)
10484 - removed obsolete files (the test scripts will be replaced
10485 by better Test::Harness variants in the future)
10486 [Ralf S. Engelschall]
10488 *) First cut for a very conservative source tree cleanup:
10489 1. merge various obsolete readme texts into doc/ssleay.txt
10490 where we collect the old documents and readme texts.
10491 2. remove the first part of files where I'm already sure that we no
10492 longer need them because of three reasons: either they are just temporary
10493 files which were left by Eric or they are preserved original files where
10494 I've verified that the diff is also available in the CVS via "cvs diff
10495 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
10496 the crypto/md/ stuff).
10497 [Ralf S. Engelschall]
10499 *) More extension code. Incomplete support for subject and issuer alt
10500 name, issuer and authority key id. Change the i2v function parameters
10501 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
10502 what that's for :-) Fix to ASN1 macro which messed up
10503 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
10506 *) Preliminary support for ENUMERATED type. This is largely copied from the
10510 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
10511 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10513 *) Make sure `make rehash' target really finds the `openssl' program.
10514 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10516 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
10517 like to hear about it if this slows down other processors.
10520 *) Add CygWin32 platform information to Configure script.
10521 [Alan Batie <batie@aahz.jf.intel.com>]
10523 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
10524 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
10526 *) New program nseq to manipulate netscape certificate sequences
10529 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
10533 *) Fixes to BN code. Previously the default was to define BN_RECURSION
10534 but the BN code had some problems that would cause failures when
10535 doing certificate verification and some other functions.
10536 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10538 *) Add ASN1 and PEM code to support netscape certificate sequences.
10541 *) Add ASN1 and PEM code to support netscape certificate sequences.
10544 *) Add several PKIX and private extended key usage OIDs.
10547 *) Modify the 'ca' program to handle the new extension code. Modify
10548 openssl.cnf for new extension format, add comments.
10551 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
10552 and add a sample to openssl.cnf so req -x509 now adds appropriate
10556 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
10557 error code, add initial support to X509_print() and x509 application.
10560 *) Takes a deep breath and start addding X509 V3 extension support code. Add
10561 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
10562 stuff is currently isolated and isn't even compiled yet.
10565 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
10566 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
10567 Removed the versions check from X509 routines when loading extensions:
10568 this allows certain broken certificates that don't set the version
10569 properly to be processed.
10572 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
10573 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
10574 can still be regenerated with "make depend".
10577 *) Spelling mistake in C version of CAST-128.
10578 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
10580 *) Changes to the error generation code. The perl script err-code.pl
10581 now reads in the old error codes and retains the old numbers, only
10582 adding new ones if necessary. It also only changes the .err files if new
10583 codes are added. The makefiles have been modified to only insert errors
10584 when needed (to avoid needlessly modifying header files). This is done
10585 by only inserting errors if the .err file is newer than the auto generated
10586 C file. To rebuild all the error codes from scratch (the old behaviour)
10587 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
10588 or delete all the .err files.
10591 *) CAST-128 was incorrectly implemented for short keys. The C version has
10592 been fixed, but is untested. The assembler versions are also fixed, but
10593 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
10594 to regenerate it if needed.
10595 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
10596 Hagino <itojun@kame.net>]
10598 *) File was opened incorrectly in randfile.c.
10599 [Ulf Möller <ulf@fitug.de>]
10601 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
10602 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
10603 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
10604 al: it's just almost always a UTCTime. Note this patch adds new error
10605 codes so do a "make errors" if there are problems.
10608 *) Correct Linux 1 recognition in config.
10609 [Ulf Möller <ulf@fitug.de>]
10611 *) Remove pointless MD5 hash when using DSA keys in ca.
10612 [Anonymous <nobody@replay.com>]
10614 *) Generate an error if given an empty string as a cert directory. Also
10615 generate an error if handed NULL (previously returned 0 to indicate an
10616 error, but didn't set one).
10617 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
10619 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
10622 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
10623 parameters. This was causing a warning which killed off the Win32 compile.
10626 *) Remove C++ style comments from crypto/bn/bn_local.h.
10627 [Neil Costigan <neil.costigan@celocom.com>]
10629 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
10630 based on a text string, looking up short and long names and finally
10631 "dot" format. The "dot" format stuff didn't work. Added new function
10632 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
10633 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
10634 OID is not part of the table.
10637 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
10638 X509_LOOKUP_by_alias().
10641 *) Sort openssl functions by name.
10644 *) Get the gendsa program working (hopefully) and add it to app list. Remove
10645 encryption from sample DSA keys (in case anyone is interested the password
10649 *) Make _all_ *_free functions accept a NULL pointer.
10650 [Frans Heymans <fheymans@isaserver.be>]
10652 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
10654 [Anonymous <nobody@replay.com>]
10656 *) s_server should send the CAfile as acceptable CAs, not its own cert.
10657 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10659 *) Don't blow it for numeric -newkey arguments to apps/req.
10660 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10662 *) Temp key "for export" tests were wrong in s3_srvr.c.
10663 [Anonymous <nobody@replay.com>]
10665 *) Add prototype for temp key callback functions
10666 SSL_CTX_set_tmp_{rsa,dh}_callback().
10669 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
10670 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
10673 *) X509_name_add_entry() freed the wrong thing after an error.
10674 [Arne Ansper <arne@ats.cyber.ee>]
10676 *) rsa_eay.c would attempt to free a NULL context.
10677 [Arne Ansper <arne@ats.cyber.ee>]
10679 *) BIO_s_socket() had a broken should_retry() on Windoze.
10680 [Arne Ansper <arne@ats.cyber.ee>]
10682 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
10683 [Arne Ansper <arne@ats.cyber.ee>]
10685 *) Make sure the already existing X509_STORE->depth variable is initialized
10686 in X509_STORE_new(), but document the fact that this variable is still
10687 unused in the certificate verification process.
10688 [Ralf S. Engelschall]
10690 *) Fix the various library and apps files to free up pkeys obtained from
10691 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
10694 *) Fix reference counting in X509_PUBKEY_get(). This makes
10695 demos/maurice/example2.c work, amongst others, probably.
10696 [Steve Henson and Ben Laurie]
10698 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
10699 `openssl' and second, the shortcut symlinks for the `openssl <command>'
10700 are no longer created. This way we have a single and consistent command
10701 line interface `openssl <command>', similar to `cvs <command>'.
10702 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
10704 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
10705 BIT STRING wrapper always have zero unused bits.
10708 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
10711 *) Make the top-level INSTALL documentation easier to understand.
10714 *) Makefiles updated to exit if an error occurs in a sub-directory
10715 make (including if user presses ^C) [Paul Sutton]
10717 *) Make Montgomery context stuff explicit in RSA data structure.
10720 *) Fix build order of pem and err to allow for generated pem.h.
10723 *) Fix renumbering bug in X509_NAME_delete_entry().
10726 *) Enhanced the err-ins.pl script so it makes the error library number
10727 global and can add a library name. This is needed for external ASN1 and
10728 other error libraries.
10731 *) Fixed sk_insert which never worked properly.
10734 *) Fix ASN1 macros so they can handle indefinite length construted
10735 EXPLICIT tags. Some non standard certificates use these: they can now
10739 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
10740 into a single doc/ssleay.txt bundle. This way the information is still
10741 preserved but no longer messes up this directory. Now it's new room for
10742 the new set of documenation files.
10743 [Ralf S. Engelschall]
10745 *) SETs were incorrectly DER encoded. This was a major pain, because they
10746 shared code with SEQUENCEs, which aren't coded the same. This means that
10747 almost everything to do with SETs or SEQUENCEs has either changed name or
10748 number of arguments.
10749 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
10751 *) Fix test data to work with the above.
10754 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
10755 was already fixed by Eric for 0.9.1 it seems.
10756 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
10758 *) Autodetect FreeBSD3.
10761 *) Fix various bugs in Configure. This affects the following platforms:
10765 unixware-2.0-pentium
10769 *) Eliminate generated files from CVS. Reorder tests to regenerate files
10770 before they are needed.
10773 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
10777 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
10779 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
10780 changed SSLeay to OpenSSL in version strings.
10781 [Ralf S. Engelschall]
10783 *) Some fixups to the top-level documents.
10786 *) Fixed the nasty bug where rsaref.h was not found under compile-time
10787 because the symlink to include/ was missing.
10788 [Ralf S. Engelschall]
10790 *) Incorporated the popular no-RSA/DSA-only patches
10791 which allow to compile a RSA-free SSLeay.
10792 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
10794 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
10795 when "ssleay" is still not found.
10796 [Ralf S. Engelschall]
10798 *) Added more platforms to Configure: Cray T3E, HPUX 11,
10799 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
10801 *) Updated the README file.
10802 [Ralf S. Engelschall]
10804 *) Added various .cvsignore files in the CVS repository subdirs
10805 to make a "cvs update" really silent.
10806 [Ralf S. Engelschall]
10808 *) Recompiled the error-definition header files and added
10809 missing symbols to the Win32 linker tables.
10810 [Ralf S. Engelschall]
10812 *) Cleaned up the top-level documents;
10813 o new files: CHANGES and LICENSE
10814 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
10815 o merged COPYRIGHT into LICENSE
10816 o removed obsolete TODO file
10817 o renamed MICROSOFT to INSTALL.W32
10818 [Ralf S. Engelschall]
10820 *) Removed dummy files from the 0.9.1b source tree:
10821 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
10822 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
10823 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
10824 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
10825 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
10826 [Ralf S. Engelschall]
10828 *) Added various platform portability fixes.
10831 *) The Genesis of the OpenSSL rpject:
10832 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
10833 Young and Tim J. Hudson created while they were working for C2Net until
10835 [The OpenSSL Project]
10838 Changes between 0.9.0b and 0.9.1b [not released]
10840 *) Updated a few CA certificates under certs/
10843 *) Changed some BIGNUM api stuff.
10846 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
10847 DGUX x86, Linux Alpha, etc.
10850 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
10851 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
10855 *) Add -strparse option to asn1pars program which parses nested
10857 [Dr Stephen Henson <shenson@bigfoot.com>]
10859 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
10862 *) DSA fix for "ca" program.
10865 *) Added "-genkey" option to "dsaparam" program.
10868 *) Added RIPE MD160 (rmd160) message digest.
10871 *) Added -a (all) option to "ssleay version" command.
10874 *) Added PLATFORM define which is the id given to Configure.
10877 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
10880 *) Extended the ASN.1 parser routines.
10883 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
10886 *) Added a BN_CTX to the BN library.
10889 *) Fixed the weak key values in DES library
10892 *) Changed API in EVP library for cipher aliases.
10895 *) Added support for RC2/64bit cipher.
10898 *) Converted the lhash library to the crypto/mem.c functions.
10901 *) Added more recognized ASN.1 object ids.
10904 *) Added more RSA padding checks for SSL/TLS.
10907 *) Added BIO proxy/filter functionality.
10910 *) Added extra_certs to SSL_CTX which can be used
10911 send extra CA certificates to the client in the CA cert chain sending
10912 process. It can be configured with SSL_CTX_add_extra_chain_cert().
10915 *) Now Fortezza is denied in the authentication phase because
10916 this is key exchange mechanism is not supported by SSLeay at all.
10919 *) Additional PKCS1 checks.
10922 *) Support the string "TLSv1" for all TLS v1 ciphers.
10925 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
10926 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
10929 *) Fixed a few memory leaks.
10932 *) Fixed various code and comment typos.
10935 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
10936 bytes sent in the client random.
10937 [Edward Bishop <ebishop@spyglass.com>]