5 Changes between 1.0.2a and 1.0.2b [xx XXX xxxx]
9 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
11 *) ClientHello sigalgs DoS fix
13 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
14 invalid signature algorithms extension a NULL pointer dereference will
15 occur. This can be exploited in a DoS attack against the server.
17 This issue was was reported to OpenSSL by David Ramos of Stanford
20 [Stephen Henson and Matt Caswell]
22 *) Multiblock corrupted pointer fix
24 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
25 feature only applies on 64 bit x86 architecture platforms that support AES
26 NI instructions. A defect in the implementation of "multiblock" can cause
27 OpenSSL's internal write buffer to become incorrectly set to NULL when
28 using non-blocking IO. Typically, when the user application is using a
29 socket BIO for writing, this will only result in a failed connection.
30 However if some other BIO is used then it is likely that a segmentation
31 fault will be triggered, thus enabling a potential DoS attack.
33 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
37 *) Segmentation fault in DTLSv1_listen fix
39 The DTLSv1_listen function is intended to be stateless and processes the
40 initial ClientHello from many peers. It is common for user code to loop
41 over the call to DTLSv1_listen until a valid ClientHello is received with
42 an associated cookie. A defect in the implementation of DTLSv1_listen means
43 that state is preserved in the SSL object from one invocation to the next
44 that can lead to a segmentation fault. Errors processing the initial
45 ClientHello can trigger this scenario. An example of such an error could be
46 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
49 This issue was reported to OpenSSL by Per Allansson.
53 *) Segmentation fault in ASN1_TYPE_cmp fix
55 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
56 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
57 certificate signature algorithm consistency this can be used to crash any
58 certificate verification operation and exploited in a DoS attack. Any
59 application which performs certificate verification is vulnerable including
60 OpenSSL clients and servers which enable client authentication.
64 *) Segmentation fault for invalid PSS parameters fix
66 The signature verification routines will crash with a NULL pointer
67 dereference if presented with an ASN.1 signature using the RSA PSS
68 algorithm and invalid parameters. Since these routines are used to verify
69 certificate signature algorithms this can be used to crash any
70 certificate verification operation and exploited in a DoS attack. Any
71 application which performs certificate verification is vulnerable including
72 OpenSSL clients and servers which enable client authentication.
74 This issue was was reported to OpenSSL by Brian Carpenter.
78 *) ASN.1 structure reuse memory corruption fix
80 Reusing a structure in ASN.1 parsing may allow an attacker to cause
81 memory corruption via an invalid write. Such reuse is and has been
82 strongly discouraged and is believed to be rare.
84 Applications that parse structures containing CHOICE or ANY DEFINED BY
85 components may be affected. Certificate parsing (d2i_X509 and related
86 functions) are however not affected. OpenSSL clients and servers are
91 *) PKCS7 NULL pointer dereferences fix
93 The PKCS#7 parsing code does not handle missing outer ContentInfo
94 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
95 missing content and trigger a NULL pointer dereference on parsing.
97 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
98 otherwise parse PKCS#7 structures from untrusted sources are
99 affected. OpenSSL clients and servers are not affected.
101 This issue was reported to OpenSSL by Michal Zalewski (Google).
105 *) DoS via reachable assert in SSLv2 servers fix
107 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
108 servers that both support SSLv2 and enable export cipher suites by sending
109 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
111 This issue was discovered by Sean Burford (Google) and Emilia Käsper
112 (OpenSSL development team).
116 *) Empty CKE with client auth and DHE fix
118 If client auth is used then a server can seg fault in the event of a DHE
119 ciphersuite being selected and a zero length ClientKeyExchange message
120 being sent by the client. This could be exploited in a DoS attack.
124 *) Handshake with unseeded PRNG fix
126 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
127 with an unseeded PRNG. The conditions are:
128 - The client is on a platform where the PRNG has not been seeded
129 automatically, and the user has not seeded manually
130 - A protocol specific client method version has been used (i.e. not
131 SSL_client_methodv23)
132 - A ciphersuite is used that does not require additional random data from
133 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
135 If the handshake succeeds then the client random that has been used will
136 have been generated from a PRNG with insufficient entropy and therefore the
137 output may be predictable.
139 For example using the following command with an unseeded openssl will
140 succeed on an unpatched platform:
142 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
146 *) Use After Free following d2i_ECPrivatekey error fix
148 A malformed EC private key file consumed via the d2i_ECPrivateKey function
149 could cause a use after free condition. This, in turn, could cause a double
150 free in several private key parsing functions (such as d2i_PrivateKey
151 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
152 for applications that receive EC private keys from untrusted
153 sources. This scenario is considered rare.
155 This issue was discovered by the BoringSSL project and fixed in their
160 *) X509_to_X509_REQ NULL pointer deref fix
162 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
163 the certificate key is invalid. This function is rarely used in practice.
165 This issue was discovered by Brian Carpenter.
169 *) Removed the export ciphers from the DEFAULT ciphers
172 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
174 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
175 ARMv5 through ARMv8, as opposite to "locking" it to single one.
176 So far those who have to target multiple plaforms would compromise
177 and argue that binary targeting say ARMv5 would still execute on
178 ARMv8. "Universal" build resolves this compromise by providing
179 near-optimal performance even on newer platforms.
182 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
183 (other platforms pending).
184 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
186 *) Add support for the SignedCertificateTimestampList certificate and
187 OCSP response extensions from RFC6962.
190 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
191 for corner cases. (Certain input points at infinity could lead to
192 bogus results, with non-infinity inputs mapped to infinity too.)
195 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
196 This covers AES, SHA256/512 and GHASH. "Initial" means that most
197 common cases are optimized and there still is room for further
198 improvements. Vector Permutation AES for Altivec is also added.
201 *) Add support for little-endian ppc64 Linux target.
202 [Marcelo Cerri (IBM)]
204 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
205 SHA1, SHA256 and GHASH. "Initial" means that most common cases
206 are optimized and there still is room for further improvements.
207 Both 32- and 64-bit modes are supported.
208 [Andy Polyakov, Ard Biesheuvel (Linaro)]
210 *) Improved ARMv7 NEON support.
213 *) Support for SPARC Architecture 2011 crypto extensions, first
214 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
215 SHA256/512, MD5, GHASH and modular exponentiation.
216 [Andy Polyakov, David Miller]
218 *) Accelerated modular exponentiation for Intel processors, a.k.a.
220 [Shay Gueron & Vlad Krasnov (Intel Corp)]
222 *) Support for new and upcoming Intel processors, including AVX2,
223 BMI and SHA ISA extensions. This includes additional "stitched"
224 implementations, AESNI-SHA256 and GCM, and multi-buffer support
227 This work was sponsored by Intel Corp.
230 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
231 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
232 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
235 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
236 this fixes a limiation in previous versions of OpenSSL.
239 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
240 MGF1 digest and OAEP label.
243 *) Add EVP support for key wrapping algorithms, to avoid problems with
244 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
245 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
246 algorithms and include tests cases.
249 *) Add functions to allocate and set the fields of an ECDSA_METHOD
251 [Douglas E. Engert, Steve Henson]
253 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
254 difference in days and seconds between two tm or ASN1_TIME structures.
257 *) Add -rev test option to s_server to just reverse order of characters
258 received by client and send back to server. Also prints an abbreviated
259 summary of the connection parameters.
262 *) New option -brief for s_client and s_server to print out a brief summary
263 of connection parameters.
266 *) Add callbacks for arbitrary TLS extensions.
267 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
269 *) New option -crl_download in several openssl utilities to download CRLs
270 from CRLDP extension in certificates.
273 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
276 *) New function X509_CRL_diff to generate a delta CRL from the difference
277 of two full CRLs. Add support to "crl" utility.
280 *) New functions to set lookup_crls function and to retrieve
281 X509_STORE from X509_STORE_CTX.
284 *) Print out deprecated issuer and subject unique ID fields in
288 *) Extend OCSP I/O functions so they can be used for simple general purpose
289 HTTP as well as OCSP. New wrapper function which can be used to download
290 CRLs using the OCSP API.
293 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
296 *) SSL_CONF* functions. These provide a common framework for application
297 configuration using configuration files or command lines.
300 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
301 message callback and prints the results. Needs compile time option
302 "enable-ssl-trace". New options to s_client and s_server to enable
306 *) New ctrl and macro to retrieve supported points extensions.
307 Print out extension in s_server and s_client.
310 *) New functions to retrieve certificate signature and signature
314 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
318 *) New Suite B modes for TLS code. These use and enforce the requirements
319 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
320 only use Suite B curves. The Suite B modes can be set by using the
321 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
324 *) New chain verification flags for Suite B levels of security. Check
325 algorithms are acceptable when flags are set in X509_verify_cert.
328 *) Make tls1_check_chain return a set of flags indicating checks passed
329 by a certificate chain. Add additional tests to handle client
330 certificates: checks for matching certificate type and issuer name
334 *) If an attempt is made to use a signature algorithm not in the peer
335 preference list abort the handshake. If client has no suitable
336 signature algorithms in response to a certificate request do not
340 *) If server EC tmp key is not in client preference list abort handshake.
343 *) Add support for certificate stores in CERT structure. This makes it
344 possible to have different stores per SSL structure or one store in
345 the parent SSL_CTX. Include distint stores for certificate chain
346 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
347 to build and store a certificate chain in CERT structure: returing
348 an error if the chain cannot be built: this will allow applications
349 to test if a chain is correctly configured.
351 Note: if the CERT based stores are not set then the parent SSL_CTX
352 store is used to retain compatibility with existing behaviour.
356 *) New function ssl_set_client_disabled to set a ciphersuite disabled
357 mask based on the current session, check mask when sending client
358 hello and checking the requested ciphersuite.
361 *) New ctrls to retrieve and set certificate types in a certificate
362 request message. Print out received values in s_client. If certificate
363 types is not set with custom values set sensible values based on
364 supported signature algorithms.
367 *) Support for distinct client and server supported signature algorithms.
370 *) Add certificate callback. If set this is called whenever a certificate
371 is required by client or server. An application can decide which
372 certificate chain to present based on arbitrary criteria: for example
373 supported signature algorithms. Add very simple example to s_server.
374 This fixes many of the problems and restrictions of the existing client
375 certificate callback: for example you can now clear an existing
376 certificate and specify the whole chain.
379 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
380 the certificate can be used for (if anything). Set valid_flags field
381 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
382 to have similar checks in it.
384 Add new "cert_flags" field to CERT structure and include a "strict mode".
385 This enforces some TLS certificate requirements (such as only permitting
386 certificate signature algorithms contained in the supported algorithms
387 extension) which some implementations ignore: this option should be used
388 with caution as it could cause interoperability issues.
391 *) Update and tidy signature algorithm extension processing. Work out
392 shared signature algorithms based on preferences and peer algorithms
393 and print them out in s_client and s_server. Abort handshake if no
394 shared signature algorithms.
397 *) Add new functions to allow customised supported signature algorithms
398 for SSL and SSL_CTX structures. Add options to s_client and s_server
402 *) New function SSL_certs_clear() to delete all references to certificates
403 from an SSL structure. Before this once a certificate had been added
404 it couldn't be removed.
407 *) Integrate hostname, email address and IP address checking with certificate
408 verification. New verify options supporting checking in opensl utility.
411 *) Fixes and wildcard matching support to hostname and email checking
412 functions. Add manual page.
413 [Florian Weimer (Red Hat Product Security Team)]
415 *) New functions to check a hostname email or IP address against a
416 certificate. Add options x509 utility to print results of checks against
420 *) Fix OCSP checking.
421 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
423 *) Initial experimental support for explicitly trusted non-root CAs.
424 OpenSSL still tries to build a complete chain to a root but if an
425 intermediate CA has a trust setting included that is used. The first
426 setting is used: whether to trust (e.g., -addtrust option to the x509
430 *) Add -trusted_first option which attempts to find certificates in the
431 trusted store even if an untrusted chain is also supplied.
434 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
435 platform support for Linux and Android.
438 *) Support for linux-x32, ILP32 environment in x86_64 framework.
441 *) Experimental multi-implementation support for FIPS capable OpenSSL.
442 When in FIPS mode the approved implementations are used as normal,
443 when not in FIPS mode the internal unapproved versions are used instead.
444 This means that the FIPS capable OpenSSL isn't forced to use the
445 (often lower perfomance) FIPS implementations outside FIPS mode.
448 *) Transparently support X9.42 DH parameters when calling
449 PEM_read_bio_DHparameters. This means existing applications can handle
450 the new parameter format automatically.
453 *) Initial experimental support for X9.42 DH parameter format: mainly
454 to support use of 'q' parameter for RFC5114 parameters.
457 *) Add DH parameters from RFC5114 including test data to dhtest.
460 *) Support for automatic EC temporary key parameter selection. If enabled
461 the most preferred EC parameters are automatically used instead of
462 hardcoded fixed parameters. Now a server just has to call:
463 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
464 support ECDH and use the most appropriate parameters.
467 *) Enhance and tidy EC curve and point format TLS extension code. Use
468 static structures instead of allocation if default values are used.
469 New ctrls to set curves we wish to support and to retrieve shared curves.
470 Print out shared curves in s_server. New options to s_server and s_client
471 to set list of supported curves.
474 *) New ctrls to retrieve supported signature algorithms and
475 supported curve values as an array of NIDs. Extend openssl utility
476 to print out received values.
479 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
480 between NIDs and the more common NIST names such as "P-256". Enhance
481 ecparam utility and ECC method to recognise the NIST names for curves.
484 *) Enhance SSL/TLS certificate chain handling to support different
485 chains for each certificate instead of one chain in the parent SSL_CTX.
488 *) Support for fixed DH ciphersuite client authentication: where both
489 server and client use DH certificates with common parameters.
492 *) Support for fixed DH ciphersuites: those requiring DH server
496 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
498 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
499 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
500 X509_CINF_get_signature were reverted post internal team review.
502 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
504 *) Build fixes for the Windows and OpenVMS platforms
505 [Matt Caswell and Richard Levitte]
507 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
509 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
510 message can cause a segmentation fault in OpenSSL due to a NULL pointer
511 dereference. This could lead to a Denial Of Service attack. Thanks to
512 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
516 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
517 dtls1_buffer_record function under certain conditions. In particular this
518 could occur if an attacker sent repeated DTLS records with the same
519 sequence number but for the next epoch. The memory leak could be exploited
520 by an attacker in a Denial of Service attack through memory exhaustion.
521 Thanks to Chris Mueller for reporting this issue.
525 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
526 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
527 method would be set to NULL which could later result in a NULL pointer
528 dereference. Thanks to Frank Schmirler for reporting this issue.
532 *) Abort handshake if server key exchange message is omitted for ephemeral
535 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
536 reporting this issue.
540 *) Remove non-export ephemeral RSA code on client and server. This code
541 violated the TLS standard by allowing the use of temporary RSA keys in
542 non-export ciphersuites and could be used by a server to effectively
543 downgrade the RSA key length used to a value smaller than the server
544 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
545 INRIA or reporting this issue.
549 *) Fixed issue where DH client certificates are accepted without verification.
550 An OpenSSL server will accept a DH certificate for client authentication
551 without the certificate verify message. This effectively allows a client to
552 authenticate without the use of a private key. This only affects servers
553 which trust a client certificate authority which issues certificates
554 containing DH keys: these are extremely rare and hardly ever encountered.
555 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
560 *) Ensure that the session ID context of an SSL is updated when its
561 SSL_CTX is updated via SSL_set_SSL_CTX.
563 The session ID context is typically set from the parent SSL_CTX,
564 and can vary with the CTX.
567 *) Fix various certificate fingerprint issues.
569 By using non-DER or invalid encodings outside the signed portion of a
570 certificate the fingerprint can be changed without breaking the signature.
571 Although no details of the signed portion of the certificate can be changed
572 this can cause problems with some applications: e.g. those using the
573 certificate fingerprint for blacklists.
575 1. Reject signatures with non zero unused bits.
577 If the BIT STRING containing the signature has non zero unused bits reject
578 the signature. All current signature algorithms require zero unused bits.
580 2. Check certificate algorithm consistency.
582 Check the AlgorithmIdentifier inside TBS matches the one in the
583 certificate signature. NB: this will result in signature failure
584 errors for some broken certificates.
586 Thanks to Konrad Kraszewski from Google for reporting this issue.
588 3. Check DSA/ECDSA signatures use DER.
590 Reencode DSA/ECDSA signatures and compare with the original received
591 signature. Return an error if there is a mismatch.
593 This will reject various cases including garbage after signature
594 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
595 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
596 (negative or with leading zeroes).
598 Further analysis was conducted and fixes were developed by Stephen Henson
599 of the OpenSSL core team.
604 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
605 results on some platforms, including x86_64. This bug occurs at random
606 with a very low probability, and is not known to be exploitable in any
607 way, though its exact impact is difficult to determine. Thanks to Pieter
608 Wuille (Blockstream) who reported this issue and also suggested an initial
609 fix. Further analysis was conducted by the OpenSSL development team and
610 Adam Langley of Google. The final fix was developed by Andy Polyakov of
611 the OpenSSL core team.
615 *) Do not resume sessions on the server if the negotiated protocol
616 version does not match the session's version. Resuming with a different
617 version, while not strictly forbidden by the RFC, is of questionable
618 sanity and breaks all known clients.
619 [David Benjamin, Emilia Käsper]
621 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
622 early CCS messages during renegotiation. (Note that because
623 renegotiation is encrypted, this early CCS was not exploitable.)
626 *) Tighten client-side session ticket handling during renegotiation:
627 ensure that the client only accepts a session ticket if the server sends
628 the extension anew in the ServerHello. Previously, a TLS client would
629 reuse the old extension state and thus accept a session ticket if one was
630 announced in the initial ServerHello.
632 Similarly, ensure that the client requires a session ticket if one
633 was advertised in the ServerHello. Previously, a TLS client would
634 ignore a missing NewSessionTicket message.
637 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
641 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
642 sends a carefully crafted handshake message, to cause OpenSSL to fail
643 to free up to 64k of memory causing a memory leak. This could be
644 exploited in a Denial Of Service attack. This issue affects OpenSSL
645 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
646 whether SRTP is used or configured. Implementations of OpenSSL that
647 have been compiled with OPENSSL_NO_SRTP defined are not affected.
649 The fix was developed by the OpenSSL team.
653 *) Session Ticket Memory Leak.
655 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
656 integrity of that ticket is first verified. In the event of a session
657 ticket integrity check failing, OpenSSL will fail to free memory
658 causing a memory leak. By sending a large number of invalid session
659 tickets an attacker could exploit this issue in a Denial Of Service
664 *) Build option no-ssl3 is incomplete.
666 When OpenSSL is configured with "no-ssl3" as a build option, servers
667 could accept and complete a SSL 3.0 handshake, and clients could be
668 configured to send them.
670 [Akamai and the OpenSSL team]
672 *) Add support for TLS_FALLBACK_SCSV.
673 Client applications doing fallback retries should call
674 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
676 [Adam Langley, Bodo Moeller]
678 *) Add additional DigestInfo checks.
680 Reencode DigestInto in DER and check against the original when
681 verifying RSA signature: this will reject any improperly encoded
682 DigestInfo structures.
684 Note: this is a precautionary measure and no attacks are currently known.
688 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
690 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
691 SRP code can be overrun an internal buffer. Add sanity check that
692 g, A, B < N to SRP code.
694 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
695 Group for discovering this issue.
699 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
700 TLS 1.0 instead of higher protocol versions when the ClientHello message
701 is badly fragmented. This allows a man-in-the-middle attacker to force a
702 downgrade to TLS 1.0 even if both the server and the client support a
703 higher protocol version, by modifying the client's TLS records.
705 Thanks to David Benjamin and Adam Langley (Google) for discovering and
706 researching this issue.
710 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
711 to a denial of service attack. A malicious server can crash the client
712 with a null pointer dereference (read) by specifying an anonymous (EC)DH
713 ciphersuite and sending carefully crafted handshake messages.
715 Thanks to Felix Gröbert (Google) for discovering and researching this
720 *) By sending carefully crafted DTLS packets an attacker could cause openssl
721 to leak memory. This can be exploited through a Denial of Service attack.
722 Thanks to Adam Langley for discovering and researching this issue.
726 *) An attacker can force openssl to consume large amounts of memory whilst
727 processing DTLS handshake messages. This can be exploited through a
728 Denial of Service attack.
729 Thanks to Adam Langley for discovering and researching this issue.
733 *) An attacker can force an error condition which causes openssl to crash
734 whilst processing DTLS packets due to memory being freed twice. This
735 can be exploited through a Denial of Service attack.
736 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
741 *) If a multithreaded client connects to a malicious server using a resumed
742 session and the server sends an ec point format extension it could write
743 up to 255 bytes to freed memory.
745 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
750 *) A malicious server can crash an OpenSSL client with a null pointer
751 dereference (read) by specifying an SRP ciphersuite even though it was not
752 properly negotiated with the client. This can be exploited through a
753 Denial of Service attack.
755 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
756 discovering and researching this issue.
760 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
761 X509_name_oneline, X509_name_print_ex et al. to leak some information
762 from the stack. Applications may be affected if they echo pretty printing
763 output to the attacker.
765 Thanks to Ivan Fratric (Google) for discovering this issue.
767 [Emilia Käsper, and Steve Henson]
769 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
770 for corner cases. (Certain input points at infinity could lead to
771 bogus results, with non-infinity inputs mapped to infinity too.)
774 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
776 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
777 handshake can force the use of weak keying material in OpenSSL
778 SSL/TLS clients and servers.
780 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
781 researching this issue. (CVE-2014-0224)
782 [KIKUCHI Masashi, Steve Henson]
784 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
785 OpenSSL DTLS client the code can be made to recurse eventually crashing
788 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
790 [Imre Rad, Steve Henson]
792 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
793 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
794 client or server. This is potentially exploitable to run arbitrary
795 code on a vulnerable client or server.
797 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
798 [Jüri Aedla, Steve Henson]
800 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
801 are subject to a denial of service attack.
803 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
804 this issue. (CVE-2014-3470)
805 [Felix Gröbert, Ivan Fratric, Steve Henson]
807 *) Harmonize version and its documentation. -f flag is used to display
809 [mancha <mancha1@zoho.com>]
811 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
813 [mancha <mancha1@zoho.com>]
815 *) Fix some double frees. These are not thought to be exploitable.
816 [mancha <mancha1@zoho.com>]
818 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
820 *) A missing bounds check in the handling of the TLS heartbeat extension
821 can be used to reveal up to 64k of memory to a connected client or
824 Thanks for Neel Mehta of Google Security for discovering this bug and to
825 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
826 preparing the fix (CVE-2014-0160)
827 [Adam Langley, Bodo Moeller]
829 *) Fix for the attack described in the paper "Recovering OpenSSL
830 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
831 by Yuval Yarom and Naomi Benger. Details can be obtained from:
832 http://eprint.iacr.org/2014/140
834 Thanks to Yuval Yarom and Naomi Benger for discovering this
835 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
836 [Yuval Yarom and Naomi Benger]
838 *) TLS pad extension: draft-agl-tls-padding-03
840 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
841 TLS client Hello record length value would otherwise be > 255 and
842 less that 512 pad with a dummy extension containing zeroes so it
843 is at least 512 bytes long.
845 [Adam Langley, Steve Henson]
847 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
849 *) Fix for TLS record tampering bug. A carefully crafted invalid
850 handshake could crash OpenSSL with a NULL pointer exception.
851 Thanks to Anton Johansson for reporting this issues.
854 *) Keep original DTLS digest and encryption contexts in retransmission
855 structures so we can use the previous session parameters if they need
856 to be resent. (CVE-2013-6450)
859 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
860 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
861 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
862 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
863 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
864 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
865 [Rob Stradling, Adam Langley]
867 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
869 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
870 supporting platforms or when small records were transferred.
871 [Andy Polyakov, Steve Henson]
873 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
875 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
877 This addresses the flaw in CBC record processing discovered by
878 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
879 at: http://www.isg.rhul.ac.uk/tls/
881 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
882 Security Group at Royal Holloway, University of London
883 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
884 Emilia Käsper for the initial patch.
886 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
888 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
889 ciphersuites which can be exploited in a denial of service attack.
890 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
891 and detecting this bug and to Wolfgang Ettlinger
892 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
896 *) Return an error when checking OCSP signatures when key is NULL.
897 This fixes a DoS attack. (CVE-2013-0166)
900 *) Make openssl verify return errors.
901 [Chris Palmer <palmer@google.com> and Ben Laurie]
903 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
904 the right response is stapled. Also change SSL_get_certificate()
905 so it returns the certificate actually sent.
906 See http://rt.openssl.org/Ticket/Display.html?id=2836.
907 [Rob Stradling <rob.stradling@comodo.com>]
909 *) Fix possible deadlock when decoding public keys.
912 *) Don't use TLS 1.0 record version number in initial client hello
916 Changes between 1.0.1b and 1.0.1c [10 May 2012]
918 *) Sanity check record length before skipping explicit IV in TLS
919 1.2, 1.1 and DTLS to fix DoS attack.
921 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
922 fuzzing as a service testing platform.
926 *) Initialise tkeylen properly when encrypting CMS messages.
927 Thanks to Solar Designer of Openwall for reporting this issue.
930 *) In FIPS mode don't try to use composite ciphers as they are not
934 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
936 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
937 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
938 mean any application compiled against OpenSSL 1.0.0 headers setting
939 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
940 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
941 0x10000000L Any application which was previously compiled against
942 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
943 will need to be recompiled as a result. Letting be results in
944 inability to disable specifically TLS 1.1 and in client context,
945 in unlike event, limit maximum offered version to TLS 1.0 [see below].
948 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
949 disable just protocol X, but all protocols above X *if* there are
950 protocols *below* X still enabled. In more practical terms it means
951 that if application wants to disable TLS1.0 in favor of TLS1.1 and
952 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
953 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
957 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
959 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
960 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
961 in CRYPTO_realloc_clean.
963 Thanks to Tavis Ormandy, Google Security Team, for discovering this
964 issue and to Adam Langley <agl@chromium.org> for fixing it.
966 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
968 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
971 *) Workarounds for some broken servers that "hang" if a client hello
972 record length exceeds 255 bytes.
974 1. Do not use record version number > TLS 1.0 in initial client
975 hello: some (but not all) hanging servers will now work.
976 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
977 the number of ciphers sent in the client hello. This should be
978 set to an even number, such as 50, for example by passing:
979 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
980 Most broken servers should now work.
981 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
982 TLS 1.2 client support entirely.
985 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
988 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
990 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
991 STRING form instead of a DigestInfo.
994 *) The format used for MDC2 RSA signatures is inconsistent between EVP
995 and the RSA_sign/RSA_verify functions. This was made more apparent when
996 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
997 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
998 the correct format in RSA_verify so both forms transparently work.
1001 *) Some servers which support TLS 1.0 can choke if we initially indicate
1002 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
1003 encrypted premaster secret. As a workaround use the maximum pemitted
1004 client version in client hello, this should keep such servers happy
1005 and still work with previous versions of OpenSSL.
1008 *) Add support for TLS/DTLS heartbeats.
1009 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1011 *) Add support for SCTP.
1012 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1014 *) Improved PRNG seeding for VOS.
1015 [Paul Green <Paul.Green@stratus.com>]
1017 *) Extensive assembler packs updates, most notably:
1019 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
1020 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
1021 - x86_64: bit-sliced AES implementation;
1022 - ARM: NEON support, contemporary platforms optimizations;
1023 - s390x: z196 support;
1024 - *: GHASH and GF(2^m) multiplication implementations;
1028 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
1029 (removal of unnecessary code)
1030 [Peter Sylvester <peter.sylvester@edelweb.fr>]
1032 *) Add TLS key material exporter from RFC 5705.
1035 *) Add DTLS-SRTP negotiation from RFC 5764.
1038 *) Add Next Protocol Negotiation,
1039 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
1040 disabled with a no-npn flag to config or Configure. Code donated
1042 [Adam Langley <agl@google.com> and Ben Laurie]
1044 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
1045 NIST-P256, NIST-P521, with constant-time single point multiplication on
1046 typical inputs. Compiler support for the nonstandard type __uint128_t is
1047 required to use this (present in gcc 4.4 and later, for 64-bit builds).
1048 Code made available under Apache License version 2.0.
1050 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
1051 line to include this in your build of OpenSSL, and run "make depend" (or
1052 "make update"). This enables the following EC_METHODs:
1054 EC_GFp_nistp224_method()
1055 EC_GFp_nistp256_method()
1056 EC_GFp_nistp521_method()
1058 EC_GROUP_new_by_curve_name() will automatically use these (while
1059 EC_GROUP_new_curve_GFp() currently prefers the more flexible
1061 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
1063 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
1064 all platforms. Move ssize_t definition from e_os.h to the public
1065 header file e_os2.h as it now appears in public header file cms.h
1068 *) New -sigopt option to the ca, req and x509 utilities. Additional
1069 signature parameters can be passed using this option and in
1073 *) Add RSA PSS signing function. This will generate and set the
1074 appropriate AlgorithmIdentifiers for PSS based on those in the
1075 corresponding EVP_MD_CTX structure. No application support yet.
1078 *) Support for companion algorithm specific ASN1 signing routines.
1079 New function ASN1_item_sign_ctx() signs a pre-initialised
1080 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
1081 the appropriate parameters.
1084 *) Add new algorithm specific ASN1 verification initialisation function
1085 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
1086 handling will be the same no matter what EVP_PKEY_METHOD is used.
1087 Add a PSS handler to support verification of PSS signatures: checked
1088 against a number of sample certificates.
1091 *) Add signature printing for PSS. Add PSS OIDs.
1092 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
1094 *) Add algorithm specific signature printing. An individual ASN1 method
1095 can now print out signatures instead of the standard hex dump.
1097 More complex signatures (e.g. PSS) can print out more meaningful
1098 information. Include DSA version that prints out the signature
1102 *) Password based recipient info support for CMS library: implementing
1106 *) Split password based encryption into PBES2 and PBKDF2 functions. This
1107 neatly separates the code into cipher and PBE sections and is required
1108 for some algorithms that split PBES2 into separate pieces (such as
1109 password based CMS).
1112 *) Session-handling fixes:
1113 - Fix handling of connections that are resuming with a session ID,
1114 but also support Session Tickets.
1115 - Fix a bug that suppressed issuing of a new ticket if the client
1116 presented a ticket with an expired session.
1117 - Try to set the ticket lifetime hint to something reasonable.
1118 - Make tickets shorter by excluding irrelevant information.
1119 - On the client side, don't ignore renewed tickets.
1120 [Adam Langley, Bodo Moeller (Google)]
1122 *) Fix PSK session representation.
1125 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
1127 This work was sponsored by Intel.
1130 *) Add GCM support to TLS library. Some custom code is needed to split
1131 the IV between the fixed (from PRF) and explicit (from TLS record)
1132 portions. This adds all GCM ciphersuites supported by RFC5288 and
1133 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
1134 add a special AESGCM string for GCM only.
1137 *) Expand range of ctrls for AES GCM. Permit setting invocation
1138 field on decrypt and retrieval of invocation field only on encrypt.
1141 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
1142 As required by RFC5289 these ciphersuites cannot be used if for
1143 versions of TLS earlier than 1.2.
1146 *) For FIPS capable OpenSSL interpret a NULL default public key method
1147 as unset and return the appopriate default but do *not* set the default.
1148 This means we can return the appopriate method in applications that
1149 swicth between FIPS and non-FIPS modes.
1152 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
1153 ENGINE is used then we cannot handle that in the FIPS module so we
1154 keep original code iff non-FIPS operations are allowed.
1157 *) Add -attime option to openssl utilities.
1158 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
1160 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
1163 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
1164 FIPS EC methods unconditionally for now.
1167 *) New build option no-ec2m to disable characteristic 2 code.
1170 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
1171 all cases can be covered as some introduce binary incompatibilities.
1174 *) Redirect RSA operations to FIPS module including keygen,
1175 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
1178 *) Add similar low level API blocking to ciphers.
1181 *) Low level digest APIs are not approved in FIPS mode: any attempt
1182 to use these will cause a fatal error. Applications that *really* want
1183 to use them can use the private_* version instead.
1186 *) Redirect cipher operations to FIPS module for FIPS builds.
1189 *) Redirect digest operations to FIPS module for FIPS builds.
1192 *) Update build system to add "fips" flag which will link in fipscanister.o
1193 for static and shared library builds embedding a signature if needed.
1196 *) Output TLS supported curves in preference order instead of numerical
1197 order. This is currently hardcoded for the highest order curves first.
1198 This should be configurable so applications can judge speed vs strength.
1201 *) Add TLS v1.2 server support for client authentication.
1204 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
1208 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
1209 FIPS modules versions.
1212 *) Add TLS v1.2 client side support for client authentication. Keep cache
1213 of handshake records longer as we don't know the hash algorithm to use
1214 until after the certificate request message is received.
1217 *) Initial TLS v1.2 client support. Add a default signature algorithms
1218 extension including all the algorithms we support. Parse new signature
1219 format in client key exchange. Relax some ECC signing restrictions for
1220 TLS v1.2 as indicated in RFC5246.
1223 *) Add server support for TLS v1.2 signature algorithms extension. Switch
1224 to new signature format when needed using client digest preference.
1225 All server ciphersuites should now work correctly in TLS v1.2. No client
1226 support yet and no support for client certificates.
1229 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
1230 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
1231 ciphersuites. At present only RSA key exchange ciphersuites work with
1232 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
1233 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
1234 and version checking.
1237 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
1238 with this defined it will not be affected by any changes to ssl internal
1239 structures. Add several utility functions to allow openssl application
1240 to work with OPENSSL_NO_SSL_INTERN defined.
1244 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
1246 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
1249 *) Permit abbreviated handshakes when renegotiating using the function
1250 SSL_renegotiate_abbreviated().
1251 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1253 *) Add call to ENGINE_register_all_complete() to
1254 ENGINE_load_builtin_engines(), so some implementations get used
1255 automatically instead of needing explicit application support.
1258 *) Add support for TLS key exporter as described in RFC5705.
1259 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
1261 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
1262 a few changes are required:
1264 Add SSL_OP_NO_TLSv1_1 flag.
1265 Add TLSv1_1 methods.
1266 Update version checking logic to handle version 1.1.
1267 Add explicit IV handling (ported from DTLS code).
1268 Add command line options to s_client/s_server.
1271 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
1273 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1274 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1275 content decryption and always return the same error. Note: this attack
1276 needs on average 2^20 messages so it only affects automated senders. The
1277 old behaviour can be reenabled in the CMS code by setting the
1278 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1279 an MMA defence is not necessary.
1280 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1281 this issue. (CVE-2012-0884)
1284 *) Fix CVE-2011-4619: make sure we really are receiving a
1285 client hello before rejecting multiple SGC restarts. Thanks to
1286 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1289 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
1291 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1292 Thanks to Antonio Martin, Enterprise Secure Access Research and
1293 Development, Cisco Systems, Inc. for discovering this bug and
1294 preparing a fix. (CVE-2012-0050)
1297 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
1299 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1300 of the Vaudenay padding oracle attack on CBC mode encryption
1301 which enables an efficient plaintext recovery attack against
1302 the OpenSSL implementation of DTLS. Their attack exploits timing
1303 differences arising during decryption processing. A research
1304 paper describing this attack can be found at:
1305 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1306 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1307 Security Group at Royal Holloway, University of London
1308 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1309 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1310 for preparing the fix. (CVE-2011-4108)
1311 [Robin Seggelmann, Michael Tuexen]
1313 *) Clear bytes used for block padding of SSL 3.0 records.
1315 [Adam Langley (Google)]
1317 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1318 Kadianakis <desnacked@gmail.com> for discovering this issue and
1319 Adam Langley for preparing the fix. (CVE-2011-4619)
1320 [Adam Langley (Google)]
1322 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
1323 [Andrey Kulikov <amdeich@gmail.com>]
1325 *) Prevent malformed RFC3779 data triggering an assertion failure.
1326 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1327 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1328 [Rob Austein <sra@hactrn.net>]
1330 *) Improved PRNG seeding for VOS.
1331 [Paul Green <Paul.Green@stratus.com>]
1333 *) Fix ssl_ciph.c set-up race.
1334 [Adam Langley (Google)]
1336 *) Fix spurious failures in ecdsatest.c.
1337 [Emilia Käsper (Google)]
1339 *) Fix the BIO_f_buffer() implementation (which was mixing different
1340 interpretations of the '..._len' fields).
1341 [Adam Langley (Google)]
1343 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1344 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1345 threads won't reuse the same blinding coefficients.
1347 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1348 lock to call BN_BLINDING_invert_ex, and avoids one use of
1349 BN_BLINDING_update for each BN_BLINDING structure (previously,
1350 the last update always remained unused).
1351 [Emilia Käsper (Google)]
1353 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
1354 [Bob Buckholz (Google)]
1356 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
1358 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
1359 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
1360 [Kaspar Brand <ossl@velox.ch>]
1362 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1363 for multi-threaded use of ECDH. (CVE-2011-3210)
1364 [Adam Langley (Google)]
1366 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1369 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
1370 signature public key algorithm by using OID xref utilities instead.
1371 Before this you could only use some ECC ciphersuites with SHA1 only.
1374 *) Add protection against ECDSA timing attacks as mentioned in the paper
1375 by Billy Bob Brumley and Nicola Tuveri, see:
1377 http://eprint.iacr.org/2011/232.pdf
1379 [Billy Bob Brumley and Nicola Tuveri]
1381 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
1383 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1384 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1386 *) Fix bug in string printing code: if *any* escaping is enabled we must
1387 escape the escape character (backslash) or the resulting string is
1391 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
1393 *) Disable code workaround for ancient and obsolete Netscape browsers
1394 and servers: an attacker can use it in a ciphersuite downgrade attack.
1395 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1398 *) Fixed J-PAKE implementation error, originally discovered by
1399 Sebastien Martini, further info and confirmation from Stefan
1400 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1403 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
1405 *) Fix extension code to avoid race conditions which can result in a buffer
1406 overrun vulnerability: resumed sessions must not be modified as they can
1407 be shared by multiple threads. CVE-2010-3864
1410 *) Fix WIN32 build system to correctly link an ENGINE directory into
1414 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
1416 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
1418 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
1420 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
1422 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
1423 context. The operation can be customised via the ctrl mechanism in
1424 case ENGINEs want to include additional functionality.
1427 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
1430 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
1431 output hashes compatible with older versions of OpenSSL.
1432 [Willy Weisz <weisz@vcpc.univie.ac.at>]
1434 *) Fix compression algorithm handling: if resuming a session use the
1435 compression algorithm of the resumed session instead of determining
1436 it from client hello again. Don't allow server to change algorithm.
1439 *) Add load_crls() function to apps tidying load_certs() too. Add option
1440 to verify utility to allow additional CRLs to be included.
1443 *) Update OCSP request code to permit adding custom headers to the request:
1444 some responders need this.
1447 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
1449 [Julia Lawall <julia@diku.dk>]
1451 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
1452 needlessly dereferenced structures, used obsolete functions and
1453 didn't handle all updated verify codes correctly.
1456 *) Disable MD2 in the default configuration.
1459 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
1460 indicate the initial BIO being pushed or popped. This makes it possible
1461 to determine whether the BIO is the one explicitly called or as a result
1462 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
1463 it handles reference counts correctly and doesn't zero out the I/O bio
1464 when it is not being explicitly popped. WARNING: applications which
1465 included workarounds for the old buggy behaviour will need to be modified
1466 or they could free up already freed BIOs.
1469 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
1470 renaming to all platforms (within the 0.9.8 branch, this was
1471 done conditionally on Netware platforms to avoid a name clash).
1472 [Guenter <lists@gknw.net>]
1474 *) Add ECDHE and PSK support to DTLS.
1475 [Michael Tuexen <tuexen@fh-muenster.de>]
1477 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
1481 *) Add "missing" function EVP_MD_flags() (without this the only way to
1482 retrieve a digest flags is by accessing the structure directly. Update
1483 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
1484 or cipher is registered as in the "from" argument. Print out all
1485 registered digests in the dgst usage message instead of manually
1486 attempting to work them out.
1489 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
1490 this allows the use of compression and extensions. Change default cipher
1491 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
1492 by default unless an application cipher string requests it.
1495 *) Alter match criteria in PKCS12_parse(). It used to try to use local
1496 key ids to find matching certificates and keys but some PKCS#12 files
1497 don't follow the (somewhat unwritten) rules and this strategy fails.
1498 Now just gather all certificates together and the first private key
1499 then look for the first certificate that matches the key.
1502 *) Support use of registered digest and cipher names for dgst and cipher
1503 commands instead of having to add each one as a special case. So now
1510 openssl dgst -sha256 foo
1512 and this works for ENGINE based algorithms too.
1516 *) Update Gost ENGINE to support parameter files.
1517 [Victor B. Wagner <vitus@cryptocom.ru>]
1519 *) Support GeneralizedTime in ca utility.
1520 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
1522 *) Enhance the hash format used for certificate directory links. The new
1523 form uses the canonical encoding (meaning equivalent names will work
1524 even if they aren't identical) and uses SHA1 instead of MD5. This form
1525 is incompatible with the older format and as a result c_rehash should
1526 be used to rebuild symbolic links.
1529 *) Make PKCS#8 the default write format for private keys, replacing the
1530 traditional format. This form is standardised, more secure and doesn't
1531 include an implicit MD5 dependency.
1534 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
1535 committed to OpenSSL should pass this lot as a minimum.
1538 *) Add session ticket override functionality for use by EAP-FAST.
1539 [Jouni Malinen <j@w1.fi>]
1541 *) Modify HMAC functions to return a value. Since these can be implemented
1542 in an ENGINE errors can occur.
1545 *) Type-checked OBJ_bsearch_ex.
1548 *) Type-checked OBJ_bsearch. Also some constification necessitated
1549 by type-checking. Still to come: TXT_DB, bsearch(?),
1550 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
1554 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
1555 seconds to a tm structure directly, instead of going through OS
1556 specific date routines. This avoids any issues with OS routines such
1557 as the year 2038 bug. New *_adj() functions for ASN1 time structures
1558 and X509_time_adj_ex() to cover the extended range. The existing
1559 X509_time_adj() is still usable and will no longer have any date issues.
1562 *) Delta CRL support. New use deltas option which will attempt to locate
1563 and search any appropriate delta CRLs available.
1565 This work was sponsored by Google.
1568 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
1569 code and add additional score elements. Validate alternate CRL paths
1570 as part of the CRL checking and indicate a new error "CRL path validation
1571 error" in this case. Applications wanting additional details can use
1572 the verify callback and check the new "parent" field. If this is not
1573 NULL CRL path validation is taking place. Existing applications wont
1574 see this because it requires extended CRL support which is off by
1577 This work was sponsored by Google.
1580 *) Support for freshest CRL extension.
1582 This work was sponsored by Google.
1585 *) Initial indirect CRL support. Currently only supported in the CRLs
1586 passed directly and not via lookup. Process certificate issuer
1587 CRL entry extension and lookup CRL entries by bother issuer name
1588 and serial number. Check and process CRL issuer entry in IDP extension.
1590 This work was sponsored by Google.
1593 *) Add support for distinct certificate and CRL paths. The CRL issuer
1594 certificate is validated separately in this case. Only enabled if
1595 an extended CRL support flag is set: this flag will enable additional
1596 CRL functionality in future.
1598 This work was sponsored by Google.
1601 *) Add support for policy mappings extension.
1603 This work was sponsored by Google.
1606 *) Fixes to pathlength constraint, self issued certificate handling,
1607 policy processing to align with RFC3280 and PKITS tests.
1609 This work was sponsored by Google.
1612 *) Support for name constraints certificate extension. DN, email, DNS
1613 and URI types are currently supported.
1615 This work was sponsored by Google.
1618 *) To cater for systems that provide a pointer-based thread ID rather
1619 than numeric, deprecate the current numeric thread ID mechanism and
1620 replace it with a structure and associated callback type. This
1621 mechanism allows a numeric "hash" to be extracted from a thread ID in
1622 either case, and on platforms where pointers are larger than 'long',
1623 mixing is done to help ensure the numeric 'hash' is usable even if it
1624 can't be guaranteed unique. The default mechanism is to use "&errno"
1625 as a pointer-based thread ID to distinguish between threads.
1627 Applications that want to provide their own thread IDs should now use
1628 CRYPTO_THREADID_set_callback() to register a callback that will call
1629 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
1631 Note that ERR_remove_state() is now deprecated, because it is tied
1632 to the assumption that thread IDs are numeric. ERR_remove_state(0)
1633 to free the current thread's error state should be replaced by
1634 ERR_remove_thread_state(NULL).
1636 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
1637 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
1638 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
1639 application was previously providing a numeric thread callback that
1640 was inappropriate for distinguishing threads, then uniqueness might
1641 have been obtained with &errno that happened immediately in the
1642 intermediate development versions of OpenSSL; this is no longer the
1643 case, the numeric thread callback will now override the automatic use
1645 [Geoff Thorpe, with help from Bodo Moeller]
1647 *) Initial support for different CRL issuing certificates. This covers a
1648 simple case where the self issued certificates in the chain exist and
1649 the real CRL issuer is higher in the existing chain.
1651 This work was sponsored by Google.
1654 *) Removed effectively defunct crypto/store from the build.
1657 *) Revamp of STACK to provide stronger type-checking. Still to come:
1658 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
1659 ASN1_STRING, CONF_VALUE.
1662 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
1663 RAM on SSL connections. This option can save about 34k per idle SSL.
1666 *) Revamp of LHASH to provide stronger type-checking. Still to come:
1667 STACK, TXT_DB, bsearch, qsort.
1670 *) Initial support for Cryptographic Message Syntax (aka CMS) based
1671 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
1672 support for data, signedData, compressedData, digestedData and
1673 encryptedData, envelopedData types included. Scripts to check against
1674 RFC4134 examples draft and interop and consistency checks of many
1675 content types and variants.
1678 *) Add options to enc utility to support use of zlib compression BIO.
1681 *) Extend mk1mf to support importing of options and assembly language
1682 files from Configure script, currently only included in VC-WIN32.
1683 The assembly language rules can now optionally generate the source
1684 files from the associated perl scripts.
1687 *) Implement remaining functionality needed to support GOST ciphersuites.
1688 Interop testing has been performed using CryptoPro implementations.
1689 [Victor B. Wagner <vitus@cryptocom.ru>]
1691 *) s390x assembler pack.
1694 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
1698 *) Implement Opaque PRF Input TLS extension as specified in
1699 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
1700 official specification yet and no extension type assignment by
1701 IANA exists, this extension (for now) will have to be explicitly
1702 enabled when building OpenSSL by providing the extension number
1703 to use. For example, specify an option
1705 -DTLSEXT_TYPE_opaque_prf_input=0x9527
1707 to the "config" or "Configure" script to enable the extension,
1708 assuming extension number 0x9527 (which is a completely arbitrary
1709 and unofficial assignment based on the MD5 hash of the Internet
1710 Draft). Note that by doing so, you potentially lose
1711 interoperability with other TLS implementations since these might
1712 be using the same extension number for other purposes.
1714 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
1715 opaque PRF input value to use in the handshake. This will create
1716 an interal copy of the length-'len' string at 'src', and will
1717 return non-zero for success.
1719 To get more control and flexibility, provide a callback function
1722 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
1723 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
1727 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
1730 Callback function 'cb' will be called in handshakes, and is
1731 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
1732 Argument 'arg' is for application purposes (the value as given to
1733 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
1734 be provided to the callback function). The callback function
1735 has to return non-zero to report success: usually 1 to use opaque
1736 PRF input just if possible, or 2 to enforce use of the opaque PRF
1737 input. In the latter case, the library will abort the handshake
1738 if opaque PRF input is not successfully negotiated.
1740 Arguments 'peerinput' and 'len' given to the callback function
1741 will always be NULL and 0 in the case of a client. A server will
1742 see the client's opaque PRF input through these variables if
1743 available (NULL and 0 otherwise). Note that if the server
1744 provides an opaque PRF input, the length must be the same as the
1745 length of the client's opaque PRF input.
1747 Note that the callback function will only be called when creating
1748 a new session (session resumption can resume whatever was
1749 previously negotiated), and will not be called in SSL 2.0
1750 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
1751 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
1752 for applications that need to enforce opaque PRF input.
1756 *) Update ssl code to support digests other than SHA1+MD5 for handshake
1759 [Victor B. Wagner <vitus@cryptocom.ru>]
1761 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1762 RFC4507bis. The encrypted ticket format is an encrypted encoded
1763 SSL_SESSION structure, that way new session features are automatically
1766 If a client application caches session in an SSL_SESSION structure
1767 support is transparent because tickets are now stored in the encoded
1770 The SSL_CTX structure automatically generates keys for ticket
1771 protection in servers so again support should be possible
1772 with no application modification.
1774 If a client or server wishes to disable RFC4507 support then the option
1775 SSL_OP_NO_TICKET can be set.
1777 Add a TLS extension debugging callback to allow the contents of any client
1778 or server extensions to be examined.
1780 This work was sponsored by Google.
1783 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
1784 OpenSSL should now compile cleanly on gcc 4.2
1785 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
1787 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
1788 support including streaming MAC support: this is required for GOST
1789 ciphersuite support.
1790 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
1792 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1793 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1794 to output in BER and PEM format.
1797 *) Experimental support for use of HMAC via EVP_PKEY interface. This
1798 allows HMAC to be handled via the EVP_DigestSign*() interface. The
1799 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
1800 ENGINE support for HMAC keys which are unextractable. New -mac and
1801 -macopt options to dgst utility.
1804 *) New option -sigopt to dgst utility. Update dgst to use
1805 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
1806 alternative signing paramaters such as X9.31 or PSS in the dgst
1810 *) Change ssl_cipher_apply_rule(), the internal function that does
1811 the work each time a ciphersuite string requests enabling
1812 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
1813 removing ("!foo+bar") a class of ciphersuites: Now it maintains
1814 the order of disabled ciphersuites such that those ciphersuites
1815 that most recently went from enabled to disabled not only stay
1816 in order with respect to each other, but also have higher priority
1817 than other disabled ciphersuites the next time ciphersuites are
1820 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
1821 the same ciphersuites as with "HIGH" alone, but in a specific
1822 order where the PSK ciphersuites come first (since they are the
1823 most recently disabled ciphersuites when "HIGH" is parsed).
1825 Also, change ssl_create_cipher_list() (using this new
1826 funcionality) such that between otherwise identical
1827 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
1831 *) Change ssl_create_cipher_list() so that it automatically
1832 arranges the ciphersuites in reasonable order before starting
1833 to process the rule string. Thus, the definition for "DEFAULT"
1834 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
1835 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
1836 This makes it much easier to arrive at a reasonable default order
1837 in applications for which anonymous ciphers are OK (meaning
1838 that you can't actually use DEFAULT).
1839 [Bodo Moeller; suggested by Victor Duchovni]
1841 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
1842 processing) into multiple integers instead of setting
1843 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
1844 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
1845 (These masks as well as the individual bit definitions are hidden
1846 away into the non-exported interface ssl/ssl_locl.h, so this
1847 change to the definition of the SSL_CIPHER structure shouldn't
1848 affect applications.) This give us more bits for each of these
1849 categories, so there is no longer a need to coagulate AES128 and
1850 AES256 into a single algorithm bit, and to coagulate Camellia128
1851 and Camellia256 into a single algorithm bit, which has led to all
1854 Thus, among other things, the kludge introduced in 0.9.7m and
1855 0.9.8e for masking out AES256 independently of AES128 or masking
1856 out Camellia256 independently of AES256 is not needed here in 0.9.9.
1858 With the change, we also introduce new ciphersuite aliases that
1859 so far were missing: "AES128", "AES256", "CAMELLIA128", and
1863 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
1864 Use the leftmost N bytes of the signature input if the input is
1865 larger than the prime q (with N being the size in bytes of q).
1868 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
1869 it yet and it is largely untested.
1872 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
1875 *) Initial incomplete changes to avoid need for function casts in OpenSSL
1876 some compilers (gcc 4.2 and later) reject their use. Safestack is
1877 reimplemented. Update ASN1 to avoid use of legacy functions.
1880 *) Win32/64 targets are linked with Winsock2.
1883 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
1884 to external functions. This can be used to increase CRL handling
1885 efficiency especially when CRLs are very large by (for example) storing
1886 the CRL revoked certificates in a database.
1889 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
1890 new CRLs added to a directory can be used. New command line option
1891 -verify_return_error to s_client and s_server. This causes real errors
1892 to be returned by the verify callback instead of carrying on no matter
1893 what. This reflects the way a "real world" verify callback would behave.
1896 *) GOST engine, supporting several GOST algorithms and public key formats.
1897 Kindly donated by Cryptocom.
1900 *) Partial support for Issuing Distribution Point CRL extension. CRLs
1901 partitioned by DP are handled but no indirect CRL or reason partitioning
1902 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
1903 selected via a scoring technique which handles IDP and AKID in CRLs.
1906 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
1907 will ultimately be used for all verify operations: this will remove the
1908 X509_STORE dependency on certificate verification and allow alternative
1909 lookup methods. X509_STORE based implementations of these two callbacks.
1912 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
1913 Modify get_crl() to find a valid (unexpired) CRL if possible.
1916 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
1917 this would be called X509_CRL_cmp() but that name is already used by
1918 a function that just compares CRL issuer names. Cache several CRL
1919 extensions in X509_CRL structure and cache CRLDP in X509.
1922 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
1923 this maps equivalent X509_NAME structures into a consistent structure.
1924 Name comparison can then be performed rapidly using memcmp().
1927 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
1931 *) Allow digests to supply their own micalg string for S/MIME type using
1932 the ctrl EVP_MD_CTRL_MICALG.
1935 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
1936 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
1937 ctrl. It can then customise the structure before and/or after signing
1941 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1942 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1943 to free up any added signature OIDs.
1946 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1947 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1948 digest and cipher tables. New options added to openssl utility:
1949 list-message-digest-algorithms and list-cipher-algorithms.
1952 *) Change the array representation of binary polynomials: the list
1953 of degrees of non-zero coefficients is now terminated with -1.
1954 Previously it was terminated with 0, which was also part of the
1955 value; thus, the array representation was not applicable to
1956 polynomials where t^0 has coefficient zero. This change makes
1957 the array representation useful in a more general context.
1960 *) Various modifications and fixes to SSL/TLS cipher string
1961 handling. For ECC, the code now distinguishes between fixed ECDH
1962 with RSA certificates on the one hand and with ECDSA certificates
1963 on the other hand, since these are separate ciphersuites. The
1964 unused code for Fortezza ciphersuites has been removed.
1966 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1967 (not "ECDHE"). For consistency with the code for DH
1968 certificates, use of ECDH certificates is now considered ECDH
1969 authentication, not RSA or ECDSA authentication (the latter is
1970 merely the CA's signing algorithm and not actively used in the
1973 The temporary ciphersuite alias "ECCdraft" is no longer
1974 available, and ECC ciphersuites are no longer excluded from "ALL"
1975 and "DEFAULT". The following aliases now exist for RFC 4492
1976 ciphersuites, most of these by analogy with the DH case:
1978 kECDHr - ECDH cert, signed with RSA
1979 kECDHe - ECDH cert, signed with ECDSA
1980 kECDH - ECDH cert (signed with either RSA or ECDSA)
1981 kEECDH - ephemeral ECDH
1982 ECDH - ECDH cert or ephemeral ECDH
1988 AECDH - anonymous ECDH
1989 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1993 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1994 Use correct micalg parameters depending on digest(s) in signed message.
1997 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1998 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
2001 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
2002 an engine to register a method. Add ENGINE lookups for methods and
2003 functional reference processing.
2006 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
2007 EVP_{Sign,Verify}* which allow an application to customise the signature
2011 *) New -resign option to smime utility. This adds one or more signers
2012 to an existing PKCS#7 signedData structure. Also -md option to use an
2013 alternative message digest algorithm for signing.
2016 *) Tidy up PKCS#7 routines and add new functions to make it easier to
2017 create PKCS7 structures containing multiple signers. Update smime
2018 application to support multiple signers.
2021 *) New -macalg option to pkcs12 utility to allow setting of an alternative
2025 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
2026 Reorganize PBE internals to lookup from a static table using NIDs,
2027 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
2028 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
2029 PRF which will be automatically used with PBES2.
2032 *) Replace the algorithm specific calls to generate keys in "req" with the
2036 *) Update PKCS#7 enveloped data routines to use new API. This is now
2037 supported by any public key method supporting the encrypt operation. A
2038 ctrl is added to allow the public key algorithm to examine or modify
2039 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
2043 *) Add a ctrl to asn1 method to allow a public key algorithm to express
2044 a default digest type to use. In most cases this will be SHA1 but some
2045 algorithms (such as GOST) need to specify an alternative digest. The
2046 return value indicates how strong the prefernce is 1 means optional and
2047 2 is mandatory (that is it is the only supported type). Modify
2048 ASN1_item_sign() to accept a NULL digest argument to indicate it should
2049 use the default md. Update openssl utilities to use the default digest
2050 type for signing if it is not explicitly indicated.
2053 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
2054 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
2055 signing method from the key type. This effectively removes the link
2056 between digests and public key types.
2059 *) Add an OID cross reference table and utility functions. Its purpose is to
2060 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
2061 rsaEncryption. This will allow some of the algorithm specific hackery
2062 needed to use the correct OID to be removed.
2065 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
2066 structures for PKCS7_sign(). They are now set up by the relevant public
2070 *) Add provisional EC pkey method with support for ECDSA and ECDH.
2073 *) Add support for key derivation (agreement) in the API, DH method and
2077 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
2078 public and private key formats. As a side effect these add additional
2079 command line functionality not previously available: DSA signatures can be
2080 generated and verified using pkeyutl and DH key support and generation in
2085 [Oliver Tappe <zooey@hirschkaefer.de>]
2087 *) New make target "install_html_docs" installs HTML renditions of the
2089 [Oliver Tappe <zooey@hirschkaefer.de>]
2091 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
2092 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
2093 support key and parameter generation and add initial key generation
2094 functionality for RSA.
2097 *) Add functions for main EVP_PKEY_method operations. The undocumented
2098 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
2099 EVP_PKEY_{encrypt,decrypt}_old.
2102 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
2103 key API, doesn't do much yet.
2106 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
2107 public key algorithms. New option to openssl utility:
2108 "list-public-key-algorithms" to print out info.
2111 *) Implement the Supported Elliptic Curves Extension for
2112 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2115 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
2116 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
2119 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
2120 utilities such as rsa, dsa, dsaparam etc except they process any key
2124 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
2125 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
2126 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
2130 *) Initial support for pluggable public key ASN1.
2131 De-spaghettify the public key ASN1 handling. Move public and private
2132 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
2133 algorithm specific handling to a single module within the relevant
2134 algorithm directory. Add functions to allow (near) opaque processing
2135 of public and private key structures.
2138 *) Implement the Supported Point Formats Extension for
2139 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2142 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
2143 for the psk identity [hint] and the psk callback functions to the
2144 SSL_SESSION, SSL and SSL_CTX structure.
2147 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
2151 SSL_CTX_use_psk_identity_hint
2152 SSL_get_psk_identity_hint
2153 SSL_get_psk_identity
2154 SSL_use_psk_identity_hint
2156 [Mika Kousa and Pasi Eronen of Nokia Corporation]
2158 *) Add RFC 3161 compliant time stamp request creation, response generation
2159 and response verification functionality.
2160 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
2162 *) Add initial support for TLS extensions, specifically for the server_name
2163 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2164 have new members for a host name. The SSL data structure has an
2165 additional member SSL_CTX *initial_ctx so that new sessions can be
2166 stored in that context to allow for session resumption, even after the
2167 SSL has been switched to a new SSL_CTX in reaction to a client's
2168 server_name extension.
2170 New functions (subject to change):
2172 SSL_get_servername()
2173 SSL_get_servername_type()
2176 New CTRL codes and macros (subject to change):
2178 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2179 - SSL_CTX_set_tlsext_servername_callback()
2180 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2181 - SSL_CTX_set_tlsext_servername_arg()
2182 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2184 openssl s_client has a new '-servername ...' option.
2186 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2187 '-key2 ...', '-servername_fatal' (subject to change). This allows
2188 testing the HostName extension for a specific single host name ('-cert'
2189 and '-key' remain fallbacks for handshakes without HostName
2190 negotiation). If the unrecogninzed_name alert has to be sent, this by
2191 default is a warning; it becomes fatal with the '-servername_fatal'
2194 [Peter Sylvester, Remy Allais, Christophe Renou]
2196 *) Whirlpool hash implementation is added.
2199 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
2200 bn(64,32). Because of instruction set limitations it doesn't have
2201 any negative impact on performance. This was done mostly in order
2202 to make it possible to share assembler modules, such as bn_mul_mont
2203 implementations, between 32- and 64-bit builds without hassle.
2206 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
2207 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
2211 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
2212 dedicated Montgomery multiplication procedure, is introduced.
2213 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
2214 "64-bit" performance on certain 32-bit targets.
2217 *) New option SSL_OP_NO_COMP to disable use of compression selectively
2218 in SSL structures. New SSL ctrl to set maximum send fragment size.
2219 Save memory by seeting the I/O buffer sizes dynamically instead of
2220 using the maximum available value.
2223 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
2224 in addition to the text details.
2227 *) Very, very preliminary EXPERIMENTAL support for printing of general
2228 ASN1 structures. This currently produces rather ugly output and doesn't
2229 handle several customised structures at all.
2232 *) Integrated support for PVK file format and some related formats such
2233 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
2234 these in the 'rsa' and 'dsa' utilities.
2237 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
2240 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
2241 place for the (very old) "NETSCAPE" format certificates which are now
2242 handled using new ASN1 code equivalents.
2245 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
2246 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
2247 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
2250 *) Modify CRL distribution points extension code to print out previously
2251 unsupported fields. Enhance extension setting code to allow setting of
2255 *) Add print and set support for Issuing Distribution Point CRL extension.
2258 *) Change 'Configure' script to enable Camellia by default.
2261 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
2263 *) When rejecting SSL/TLS records due to an incorrect version number, never
2264 update s->server with a new major version number. As of
2265 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
2266 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
2267 the previous behavior could result in a read attempt at NULL when
2268 receiving specific incorrect SSL/TLS records once record payload
2269 protection is active. (CVE-2010-0740)
2270 [Bodo Moeller, Adam Langley <agl@chromium.org>]
2272 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
2273 could be crashed if the relevant tables were not present (e.g. chrooted).
2274 [Tomas Hoger <thoger@redhat.com>]
2276 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
2278 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
2279 [Martin Olsson, Neel Mehta]
2281 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
2282 accommodate for stack sorting, always a write lock!).
2285 *) On some versions of WIN32 Heap32Next is very slow. This can cause
2286 excessive delays in the RAND_poll(): over a minute. As a workaround
2287 include a time check in the inner Heap32Next loop too.
2290 *) The code that handled flushing of data in SSL/TLS originally used the
2291 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
2292 the problem outlined in PR#1949. The fix suggested there however can
2293 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
2294 of Apache). So instead simplify the code to flush unconditionally.
2295 This should be fine since flushing with no data to flush is a no op.
2298 *) Handle TLS versions 2.0 and later properly and correctly use the
2299 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
2300 off ancient servers have a habit of sticking around for a while...
2303 *) Modify compression code so it frees up structures without using the
2304 ex_data callbacks. This works around a problem where some applications
2305 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
2306 restarting) then use compression (e.g. SSL with compression) later.
2307 This results in significant per-connection memory leaks and
2308 has caused some security issues including CVE-2008-1678 and
2312 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
2313 change when encrypting or decrypting.
2316 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
2317 connect and renegotiate with servers which do not support RI.
2318 Until RI is more widely deployed this option is enabled by default.
2321 *) Add "missing" ssl ctrls to clear options and mode.
2324 *) If client attempts to renegotiate and doesn't support RI respond with
2325 a no_renegotiation alert as required by RFC5746. Some renegotiating
2326 TLS clients will continue a connection gracefully when they receive
2327 the alert. Unfortunately OpenSSL mishandled this alert and would hang
2328 waiting for a server hello which it will never receive. Now we treat a
2329 received no_renegotiation alert as a fatal error. This is because
2330 applications requesting a renegotiation might well expect it to succeed
2331 and would have no code in place to handle the server denying it so the
2332 only safe thing to do is to terminate the connection.
2335 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
2336 peer supports secure renegotiation and 0 otherwise. Print out peer
2337 renegotiation support in s_client/s_server.
2340 *) Replace the highly broken and deprecated SPKAC certification method with
2341 the updated NID creation version. This should correctly handle UTF8.
2344 *) Implement RFC5746. Re-enable renegotiation but require the extension
2345 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
2346 turns out to be a bad idea. It has been replaced by
2347 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
2348 SSL_CTX_set_options(). This is really not recommended unless you
2349 know what you are doing.
2350 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
2352 *) Fixes to stateless session resumption handling. Use initial_ctx when
2353 issuing and attempting to decrypt tickets in case it has changed during
2354 servername handling. Use a non-zero length session ID when attempting
2355 stateless session resumption: this makes it possible to determine if
2356 a resumption has occurred immediately after receiving server hello
2357 (several places in OpenSSL subtly assume this) instead of later in
2361 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
2362 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
2363 fixes for a few places where the return code is not checked
2365 [Julia Lawall <julia@diku.dk>]
2367 *) Add --strict-warnings option to Configure script to include devteam
2368 warnings in other configurations.
2371 *) Add support for --libdir option and LIBDIR variable in makefiles. This
2372 makes it possible to install openssl libraries in locations which
2373 have names other than "lib", for example "/usr/lib64" which some
2375 [Steve Henson, based on patch from Jeremy Utley]
2377 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
2378 X690 8.9.12 and can produce some misleading textual output of OIDs.
2379 [Steve Henson, reported by Dan Kaminsky]
2381 *) Delete MD2 from algorithm tables. This follows the recommendation in
2382 several standards that it is not used in new applications due to
2383 several cryptographic weaknesses. For binary compatibility reasons
2384 the MD2 API is still compiled in by default.
2387 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
2391 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
2392 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
2394 [Guenter <lists@gknw.net>]
2396 *) Fix the server certificate chain building code to use X509_verify_cert(),
2397 it used to have an ad-hoc builder which was unable to cope with anything
2398 other than a simple chain.
2399 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
2401 *) Don't check self signed certificate signatures in X509_verify_cert()
2402 by default (a flag can override this): it just wastes time without
2403 adding any security. As a useful side effect self signed root CAs
2404 with non-FIPS digests are now usable in FIPS mode.
2407 *) In dtls1_process_out_of_seq_message() the check if the current message
2408 is already buffered was missing. For every new message was memory
2409 allocated, allowing an attacker to perform an denial of service attack
2410 with sending out of seq handshake messages until there is no memory
2411 left. Additionally every future messege was buffered, even if the
2412 sequence number made no sense and would be part of another handshake.
2413 So only messages with sequence numbers less than 10 in advance will be
2414 buffered. (CVE-2009-1378)
2415 [Robin Seggelmann, discovered by Daniel Mentz]
2417 *) Records are buffered if they arrive with a future epoch to be
2418 processed after finishing the corresponding handshake. There is
2419 currently no limitation to this buffer allowing an attacker to perform
2420 a DOS attack with sending records with future epochs until there is no
2421 memory left. This patch adds the pqueue_size() function to detemine
2422 the size of a buffer and limits the record buffer to 100 entries.
2424 [Robin Seggelmann, discovered by Daniel Mentz]
2426 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
2427 parent structure is freed. (CVE-2009-1379)
2430 *) Handle non-blocking I/O properly in SSL_shutdown() call.
2431 [Darryl Miles <darryl-mailinglists@netbauds.net>]
2434 [Ilya O. <vrghost@gmail.com>]
2436 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
2438 *) Disable renegotiation completely - this fixes a severe security
2439 problem (CVE-2009-3555) at the cost of breaking all
2440 renegotiation. Renegotiation can be re-enabled by setting
2441 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
2442 run-time. This is really not recommended unless you know what
2446 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
2448 *) Don't set val to NULL when freeing up structures, it is freed up by
2449 underlying code. If sizeof(void *) > sizeof(long) this can result in
2450 zeroing past the valid field. (CVE-2009-0789)
2451 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
2453 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
2454 checked correctly. This would allow some invalid signed attributes to
2455 appear to verify correctly. (CVE-2009-0591)
2456 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2458 *) Reject UniversalString and BMPString types with invalid lengths. This
2459 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
2460 a legal length. (CVE-2009-0590)
2463 *) Set S/MIME signing as the default purpose rather than setting it
2464 unconditionally. This allows applications to override it at the store
2468 *) Permit restricted recursion of ASN1 strings. This is needed in practice
2469 to handle some structures.
2472 *) Improve efficiency of mem_gets: don't search whole buffer each time
2474 [Jeremy Shapiro <jnshapir@us.ibm.com>]
2476 *) New -hex option for openssl rand.
2479 *) Print out UTF8String and NumericString when parsing ASN1.
2482 *) Support NumericString type for name components.
2485 *) Allow CC in the environment to override the automatically chosen
2486 compiler. Note that nothing is done to ensure flags work with the
2490 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2492 *) Properly check EVP_VerifyFinal() and similar return values
2494 [Ben Laurie, Bodo Moeller, Google Security Team]
2496 *) Enable TLS extensions by default.
2499 *) Allow the CHIL engine to be loaded, whether the application is
2500 multithreaded or not. (This does not release the developer from the
2501 obligation to set up the dynamic locking callbacks.)
2502 [Sander Temme <sander@temme.net>]
2504 *) Use correct exit code if there is an error in dgst command.
2505 [Steve Henson; problem pointed out by Roland Dirlewanger]
2507 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2508 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2511 *) Add experimental JPAKE support, including demo authentication in
2512 s_client and s_server.
2515 *) Set the comparison function in v3_addr_canonize().
2516 [Rob Austein <sra@hactrn.net>]
2518 *) Add support for XMPP STARTTLS in s_client.
2519 [Philip Paeps <philip@freebsd.org>]
2521 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
2522 to ensure that even with this option, only ciphersuites in the
2523 server's preference list will be accepted. (Note that the option
2524 applies only when resuming a session, so the earlier behavior was
2525 just about the algorithm choice for symmetric cryptography.)
2528 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
2530 *) Fix NULL pointer dereference if a DTLS server received
2531 ChangeCipherSpec as first record (CVE-2009-1386).
2534 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
2535 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2538 *) The fix in 0.9.8c that supposedly got rid of unsafe
2539 double-checked locking was incomplete for RSA blinding,
2540 addressing just one layer of what turns out to have been
2541 doubly unsafe triple-checked locking.
2543 So now fix this for real by retiring the MONT_HELPER macro
2544 in crypto/rsa/rsa_eay.c.
2546 [Bodo Moeller; problem pointed out by Marius Schilder]
2548 *) Various precautionary measures:
2550 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
2552 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
2553 (NB: This would require knowledge of the secret session ticket key
2554 to exploit, in which case you'd be SOL either way.)
2556 - Change bn_nist.c so that it will properly handle input BIGNUMs
2557 outside the expected range.
2559 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2562 [Neel Mehta, Bodo Moeller]
2564 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2565 the load fails. Useful for distros.
2566 [Ben Laurie and the FreeBSD team]
2568 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2571 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2574 *) Expand ENGINE to support engine supplied SSL client certificate functions.
2576 This work was sponsored by Logica.
2579 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
2580 keystores. Support for SSL/TLS client authentication too.
2581 Not compiled unless enable-capieng specified to Configure.
2583 This work was sponsored by Logica.
2586 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
2587 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
2588 attribute creation routines such as certifcate requests and PKCS#12
2592 Changes between 0.9.8g and 0.9.8h [28 May 2008]
2594 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
2595 handshake which could lead to a cilent crash as found using the
2596 Codenomicon TLS test suite (CVE-2008-1672)
2597 [Steve Henson, Mark Cox]
2599 *) Fix double free in TLS server name extensions which could lead to
2600 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
2603 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
2605 Clear the error queue to ensure that error entries left from
2606 older function calls do not interfere with the correct operation.
2607 [Lutz Jaenicke, Erik de Castro Lopo]
2609 *) Remove root CA certificates of commercial CAs:
2611 The OpenSSL project does not recommend any specific CA and does not
2612 have any policy with respect to including or excluding any CA.
2613 Therefore it does not make any sense to ship an arbitrary selection
2614 of root CA certificates with the OpenSSL software.
2617 *) RSA OAEP patches to fix two separate invalid memory reads.
2618 The first one involves inputs when 'lzero' is greater than
2619 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
2620 before the beginning of from). The second one involves inputs where
2621 the 'db' section contains nothing but zeroes (there is a one-byte
2622 invalid read after the end of 'db').
2623 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2625 *) Partial backport from 0.9.9-dev:
2627 Introduce bn_mul_mont (dedicated Montgomery multiplication
2628 procedure) as a candidate for BIGNUM assembler implementation.
2629 While 0.9.9-dev uses assembler for various architectures, only
2630 x86_64 is available by default here in the 0.9.8 branch, and
2631 32-bit x86 is available through a compile-time setting.
2633 To try the 32-bit x86 assembler implementation, use Configure
2634 option "enable-montasm" (which exists only for this backport).
2636 As "enable-montasm" for 32-bit x86 disclaims code stability
2637 anyway, in this constellation we activate additional code
2638 backported from 0.9.9-dev for further performance improvements,
2639 namely BN_from_montgomery_word. (To enable this otherwise,
2640 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
2642 [Andy Polyakov (backport partially by Bodo Moeller)]
2644 *) Add TLS session ticket callback. This allows an application to set
2645 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
2646 values. This is useful for key rollover for example where several key
2647 sets may exist with different names.
2650 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
2651 This was broken until now in 0.9.8 releases, such that the only way
2652 a registered ENGINE could be used (assuming it initialises
2653 successfully on the host) was to explicitly set it as the default
2654 for the relevant algorithms. This is in contradiction with 0.9.7
2655 behaviour and the documentation. With this fix, when an ENGINE is
2656 registered into a given algorithm's table of implementations, the
2657 'uptodate' flag is reset so that auto-discovery will be used next
2658 time a new context for that algorithm attempts to select an
2660 [Ian Lister (tweaked by Geoff Thorpe)]
2662 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
2663 implemention in the following ways:
2665 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
2668 Lack of BER streaming support means one pass streaming processing is
2669 only supported if data is detached: setting the streaming flag is
2670 ignored for embedded content.
2672 CMS support is disabled by default and must be explicitly enabled
2673 with the enable-cms configuration option.
2676 *) Update the GMP engine glue to do direct copies between BIGNUM and
2677 mpz_t when openssl and GMP use the same limb size. Otherwise the
2678 existing "conversion via a text string export" trick is still used.
2679 [Paul Sheer <paulsheer@gmail.com>]
2681 *) Zlib compression BIO. This is a filter BIO which compressed and
2682 uncompresses any data passed through it.
2685 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
2686 RFC3394 compatible AES key wrapping.
2689 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
2690 sets string data without copying. X509_ALGOR_set0() and
2691 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
2692 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
2693 from an X509_ATTRIBUTE structure optionally checking it occurs only
2694 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
2698 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
2699 to get the expected BN_FLG_CONSTTIME behavior.
2700 [Bodo Moeller (Google)]
2704 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
2705 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
2706 - added some more tests to do_tests.pl
2707 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
2708 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
2709 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
2710 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
2711 - various changes to netware.pl to enable gcc-cross builds on Win32
2713 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
2714 - various changes to fix missing prototype warnings
2715 - fixed x86nasm.pl to create correct asm files for NASM COFF output
2716 - added AES, WHIRLPOOL and CPUID assembler code to build files
2717 - added missing AES assembler make rules to mk1mf.pl
2718 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
2719 [Guenter Knauf <eflash@gmx.net>]
2721 *) Implement certificate status request TLS extension defined in RFC3546.
2722 A client can set the appropriate parameters and receive the encoded
2723 OCSP response via a callback. A server can query the supplied parameters
2724 and set the encoded OCSP response in the callback. Add simplified examples
2725 to s_client and s_server.
2728 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
2730 *) Fix various bugs:
2731 + Binary incompatibility of ssl_ctx_st structure
2732 + DTLS interoperation with non-compliant servers
2733 + Don't call get_session_cb() without proposed session
2734 + Fix ia64 assembler code
2735 [Andy Polyakov, Steve Henson]
2737 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
2739 *) DTLS Handshake overhaul. There were longstanding issues with
2740 OpenSSL DTLS implementation, which were making it impossible for
2741 RFC 4347 compliant client to communicate with OpenSSL server.
2742 Unfortunately just fixing these incompatibilities would "cut off"
2743 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
2744 server keeps tolerating non RFC compliant syntax. The opposite is
2745 not true, 0.9.8f client can not communicate with earlier server.
2746 This update even addresses CVE-2007-4995.
2749 *) Changes to avoid need for function casts in OpenSSL: some compilers
2750 (gcc 4.2 and later) reject their use.
2751 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
2754 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2755 RFC4507bis. The encrypted ticket format is an encrypted encoded
2756 SSL_SESSION structure, that way new session features are automatically
2759 If a client application caches session in an SSL_SESSION structure
2760 support is transparent because tickets are now stored in the encoded
2763 The SSL_CTX structure automatically generates keys for ticket
2764 protection in servers so again support should be possible
2765 with no application modification.
2767 If a client or server wishes to disable RFC4507 support then the option
2768 SSL_OP_NO_TICKET can be set.
2770 Add a TLS extension debugging callback to allow the contents of any client
2771 or server extensions to be examined.
2773 This work was sponsored by Google.
2776 *) Add initial support for TLS extensions, specifically for the server_name
2777 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2778 have new members for a host name. The SSL data structure has an
2779 additional member SSL_CTX *initial_ctx so that new sessions can be
2780 stored in that context to allow for session resumption, even after the
2781 SSL has been switched to a new SSL_CTX in reaction to a client's
2782 server_name extension.
2784 New functions (subject to change):
2786 SSL_get_servername()
2787 SSL_get_servername_type()
2790 New CTRL codes and macros (subject to change):
2792 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2793 - SSL_CTX_set_tlsext_servername_callback()
2794 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2795 - SSL_CTX_set_tlsext_servername_arg()
2796 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2798 openssl s_client has a new '-servername ...' option.
2800 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2801 '-key2 ...', '-servername_fatal' (subject to change). This allows
2802 testing the HostName extension for a specific single host name ('-cert'
2803 and '-key' remain fallbacks for handshakes without HostName
2804 negotiation). If the unrecogninzed_name alert has to be sent, this by
2805 default is a warning; it becomes fatal with the '-servername_fatal'
2808 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2810 *) Add AES and SSE2 assembly language support to VC++ build.
2813 *) Mitigate attack on final subtraction in Montgomery reduction.
2816 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2817 (which previously caused an internal error).
2820 *) Squeeze another 10% out of IGE mode when in != out.
2823 *) AES IGE mode speedup.
2824 [Dean Gaudet (Google)]
2826 *) Add the Korean symmetric 128-bit cipher SEED (see
2827 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2828 add SEED ciphersuites from RFC 4162:
2830 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2831 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2832 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2833 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2835 To minimize changes between patchlevels in the OpenSSL 0.9.8
2836 series, SEED remains excluded from compilation unless OpenSSL
2837 is configured with 'enable-seed'.
2838 [KISA, Bodo Moeller]
2840 *) Mitigate branch prediction attacks, which can be practical if a
2841 single processor is shared, allowing a spy process to extract
2842 information. For detailed background information, see
2843 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2844 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2845 and Necessary Software Countermeasures"). The core of the change
2846 are new versions BN_div_no_branch() and
2847 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2848 respectively, which are slower, but avoid the security-relevant
2849 conditional branches. These are automatically called by BN_div()
2850 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2851 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2852 remove a conditional branch.
2854 BN_FLG_CONSTTIME is the new name for the previous
2855 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2856 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2857 in the exponent causes BN_mod_exp_mont() to use the alternative
2858 implementation in BN_mod_exp_mont_consttime().) The old name
2859 remains as a deprecated alias.
2861 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2862 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2863 constant-time implementations for more than just exponentiation.
2864 Here too the old name is kept as a deprecated alias.
2866 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2867 the BN_BLINDING structure gets an independent copy of the
2868 modulus. This means that the previous "BIGNUM *m" argument to
2869 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2870 essentially becomes "const BIGNUM *m", although we can't actually
2871 change this in the header file before 0.9.9. It allows
2872 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2873 enable BN_FLG_CONSTTIME.
2875 [Matthew D Wood (Intel Corp)]
2877 *) In the SSL/TLS server implementation, be strict about session ID
2878 context matching (which matters if an application uses a single
2879 external cache for different purposes). Previously,
2880 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2881 set. This did ensure strict client verification, but meant that,
2882 with applications using a single external cache for quite
2883 different requirements, clients could circumvent ciphersuite
2884 restrictions for a given session ID context by starting a session
2885 in a different context.
2888 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2889 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2890 authentication-only ciphersuites.
2893 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2894 not complete and could lead to a possible single byte overflow
2895 (CVE-2007-5135) [Ben Laurie]
2897 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2899 *) Since AES128 and AES256 (and similarly Camellia128 and
2900 Camellia256) share a single mask bit in the logic of
2901 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2902 kludge to work properly if AES128 is available and AES256 isn't
2903 (or if Camellia128 is available and Camellia256 isn't).
2906 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2907 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2908 When a point or a seed is encoded in a BIT STRING, we need to
2909 prevent the removal of trailing zero bits to get the proper DER
2910 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2911 of a NamedBitList, for which trailing 0 bits need to be removed.)
2914 *) Have SSL/TLS server implementation tolerate "mismatched" record
2915 protocol version while receiving ClientHello even if the
2916 ClientHello is fragmented. (The server can't insist on the
2917 particular protocol version it has chosen before the ServerHello
2918 message has informed the client about his choice.)
2921 *) Add RFC 3779 support.
2922 [Rob Austein for ARIN, Ben Laurie]
2924 *) Load error codes if they are not already present instead of using a
2925 static variable. This allows them to be cleanly unloaded and reloaded.
2926 Improve header file function name parsing.
2929 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2930 or CAPABILITY handshake as required by RFCs.
2933 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2935 *) Introduce limits to prevent malicious keys being able to
2936 cause a denial of service. (CVE-2006-2940)
2937 [Steve Henson, Bodo Moeller]
2939 *) Fix ASN.1 parsing of certain invalid structures that can result
2940 in a denial of service. (CVE-2006-2937) [Steve Henson]
2942 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2943 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2945 *) Fix SSL client code which could crash if connecting to a
2946 malicious SSLv2 server. (CVE-2006-4343)
2947 [Tavis Ormandy and Will Drewry, Google Security Team]
2949 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2950 match only those. Before that, "AES256-SHA" would be interpreted
2951 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2952 the same strength classification in 0.9.7h) as we currently only
2953 have a single AES bit in the ciphersuite description bitmap.
2954 That change, however, also applied to ciphersuite strings such as
2955 "RC4-MD5" that intentionally matched multiple ciphersuites --
2956 namely, SSL 2.0 ciphersuites in addition to the more common ones
2957 from SSL 3.0/TLS 1.0.
2959 So we change the selection algorithm again: Naming an explicit
2960 ciphersuite selects this one ciphersuite, and any other similar
2961 ciphersuite (same bitmap) from *other* protocol versions.
2962 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2963 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2965 Since SSL 2.0 does not have any ciphersuites for which the
2966 128/256 bit distinction would be relevant, this works for now.
2967 The proper fix will be to use different bits for AES128 and
2968 AES256, which would have avoided the problems from the beginning;
2969 however, bits are scarce, so we can only do this in a new release
2970 (not just a patchlevel) when we can change the SSL_CIPHER
2971 definition to split the single 'unsigned long mask' bitmap into
2972 multiple values to extend the available space.
2976 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2978 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2979 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2981 *) Add AES IGE and biIGE modes.
2984 *) Change the Unix randomness entropy gathering to use poll() when
2985 possible instead of select(), since the latter has some
2986 undesirable limitations.
2987 [Darryl Miles via Richard Levitte and Bodo Moeller]
2989 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2990 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2991 cannot be implicitly activated as part of, e.g., the "AES" alias.
2992 However, please upgrade to OpenSSL 0.9.9[-dev] for
2993 non-experimental use of the ECC ciphersuites to get TLS extension
2994 support, which is required for curve and point format negotiation
2995 to avoid potential handshake problems.
2998 *) Disable rogue ciphersuites:
3000 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3001 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3002 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3004 The latter two were purportedly from
3005 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3008 Also deactivate the remaining ciphersuites from
3009 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3010 unofficial, and the ID has long expired.
3013 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3014 dual-core machines) and other potential thread-safety issues.
3017 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3018 versions), which is now available for royalty-free use
3019 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3020 Also, add Camellia TLS ciphersuites from RFC 4132.
3022 To minimize changes between patchlevels in the OpenSSL 0.9.8
3023 series, Camellia remains excluded from compilation unless OpenSSL
3024 is configured with 'enable-camellia'.
3027 *) Disable the padding bug check when compression is in use. The padding
3028 bug check assumes the first packet is of even length, this is not
3029 necessarily true if compresssion is enabled and can result in false
3030 positives causing handshake failure. The actual bug test is ancient
3031 code so it is hoped that implementations will either have fixed it by
3032 now or any which still have the bug do not support compression.
3035 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3037 *) When applying a cipher rule check to see if string match is an explicit
3038 cipher suite and only match that one cipher suite if it is.
3041 *) Link in manifests for VC++ if needed.
3042 [Austin Ziegler <halostatue@gmail.com>]
3044 *) Update support for ECC-based TLS ciphersuites according to
3045 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3046 TLS extensions, which are supported starting with the 0.9.9
3047 branch, not in the OpenSSL 0.9.8 branch).
3050 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3051 opaque EVP_CIPHER_CTX handling.
3054 *) Fixes and enhancements to zlib compression code. We now only use
3055 "zlib1.dll" and use the default __cdecl calling convention on Win32
3056 to conform with the standards mentioned here:
3057 http://www.zlib.net/DLL_FAQ.txt
3058 Static zlib linking now works on Windows and the new --with-zlib-include
3059 --with-zlib-lib options to Configure can be used to supply the location
3060 of the headers and library. Gracefully handle case where zlib library
3064 *) Several fixes and enhancements to the OID generation code. The old code
3065 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3066 handle numbers larger than ULONG_MAX, truncated printing and had a
3067 non standard OBJ_obj2txt() behaviour.
3070 *) Add support for building of engines under engine/ as shared libraries
3071 under VC++ build system.
3074 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3075 Hopefully, we will not see any false combination of paths any more.
3078 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3080 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3081 (part of SSL_OP_ALL). This option used to disable the
3082 countermeasure against man-in-the-middle protocol-version
3083 rollback in the SSL 2.0 server implementation, which is a bad
3084 idea. (CVE-2005-2969)
3086 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3087 for Information Security, National Institute of Advanced Industrial
3088 Science and Technology [AIST], Japan)]
3090 *) Add two function to clear and return the verify parameter flags.
3093 *) Keep cipherlists sorted in the source instead of sorting them at
3094 runtime, thus removing the need for a lock.
3097 *) Avoid some small subgroup attacks in Diffie-Hellman.
3098 [Nick Mathewson and Ben Laurie]
3100 *) Add functions for well-known primes.
3103 *) Extended Windows CE support.
3104 [Satoshi Nakamura and Andy Polyakov]
3106 *) Initialize SSL_METHOD structures at compile time instead of during
3107 runtime, thus removing the need for a lock.
3110 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3111 attempting to decrypt each encrypted key in turn. Add support to
3115 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3117 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3120 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3123 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3124 key into the same file any more.
3127 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
3130 *) Add -utf8 command line and config file option to 'ca'.
3131 [Stefan <stf@udoma.org]
3133 *) Removed the macro des_crypt(), as it seems to conflict with some
3134 libraries. Use DES_crypt().
3137 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
3138 involves renaming the source and generated shared-libs for
3139 both. The engines will accept the corrected or legacy ids
3140 ('ncipher' and '4758_cca' respectively) when binding. NB,
3141 this only applies when building 'shared'.
3142 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
3144 *) Add attribute functions to EVP_PKEY structure. Modify
3145 PKCS12_create() to recognize a CSP name attribute and
3146 use it. Make -CSP option work again in pkcs12 utility.
3149 *) Add new functionality to the bn blinding code:
3150 - automatic re-creation of the BN_BLINDING parameters after
3151 a fixed number of uses (currently 32)
3152 - add new function for parameter creation
3153 - introduce flags to control the update behaviour of the
3154 BN_BLINDING parameters
3155 - hide BN_BLINDING structure
3156 Add a second BN_BLINDING slot to the RSA structure to improve
3157 performance when a single RSA object is shared among several
3161 *) Add support for DTLS.
3162 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
3164 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
3165 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
3168 *) Remove buggy and incompletet DH cert support from
3169 ssl/ssl_rsa.c and ssl/s3_both.c
3172 *) Use SHA-1 instead of MD5 as the default digest algorithm for
3173 the apps/openssl applications.
3176 *) Compile clean with "-Wall -Wmissing-prototypes
3177 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
3178 DEBUG_SAFESTACK must also be set.
3181 *) Change ./Configure so that certain algorithms can be disabled by default.
3182 The new counterpiece to "no-xxx" is "enable-xxx".
3184 The patented RC5 and MDC2 algorithms will now be disabled unless
3185 "enable-rc5" and "enable-mdc2", respectively, are specified.
3187 (IDEA remains enabled despite being patented. This is because IDEA
3188 is frequently required for interoperability, and there is no license
3189 fee for non-commercial use. As before, "no-idea" can be used to
3190 avoid this algorithm.)
3194 *) Add processing of proxy certificates (see RFC 3820). This work was
3195 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
3196 EGEE (Enabling Grids for E-science in Europe).
3199 *) RC4 performance overhaul on modern architectures/implementations, such
3200 as Intel P4, IA-64 and AMD64.
3203 *) New utility extract-section.pl. This can be used specify an alternative
3204 section number in a pod file instead of having to treat each file as
3205 a separate case in Makefile. This can be done by adding two lines to the
3208 =for comment openssl_section:XXX
3210 The blank line is mandatory.
3214 *) New arguments -certform, -keyform and -pass for s_client and s_server
3215 to allow alternative format key and certificate files and passphrase
3219 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
3220 update associated structures and add various utility functions.
3222 Add new policy related verify parameters, include policy checking in
3223 standard verify code. Enhance 'smime' application with extra parameters
3224 to support policy checking and print out.
3227 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
3228 Nehemiah processors. These extensions support AES encryption in hardware
3229 as well as RNG (though RNG support is currently disabled).
3230 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
3232 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
3235 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
3236 [Andy Polyakov and a number of other people]
3238 *) Improved PowerPC platform support. Most notably BIGNUM assembler
3239 implementation contributed by IBM.
3240 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
3242 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
3243 exponent rather than 'unsigned long'. There is a corresponding change to
3244 the new 'rsa_keygen' element of the RSA_METHOD structure.
3245 [Jelte Jansen, Geoff Thorpe]
3247 *) Functionality for creating the initial serial number file is now
3248 moved from CA.pl to the 'ca' utility with a new option -create_serial.
3250 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
3251 number file to 1, which is bound to cause problems. To avoid
3252 the problems while respecting compatibility between different 0.9.7
3253 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
3254 CA.pl for serial number initialization. With the new release 0.9.8,
3255 we can fix the problem directly in the 'ca' utility.)
3258 *) Reduced header interdepencies by declaring more opaque objects in
3259 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
3260 give fewer recursive includes, which could break lazy source code - so
3261 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
3262 developers should define this symbol when building and using openssl to
3263 ensure they track the recommended behaviour, interfaces, [etc], but
3264 backwards-compatible behaviour prevails when this isn't defined.
3267 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
3270 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
3271 This will generate a random key of the appropriate length based on the
3272 cipher context. The EVP_CIPHER can provide its own random key generation
3273 routine to support keys of a specific form. This is used in the des and
3274 3des routines to generate a key of the correct parity. Update S/MIME
3275 code to use new functions and hence generate correct parity DES keys.
3276 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
3277 valid (weak or incorrect parity).
3280 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
3281 as looking them up. This is useful when the verified structure may contain
3282 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
3283 present unless the new PKCS7_NO_CRL flag is asserted.
3286 *) Extend ASN1 oid configuration module. It now additionally accepts the
3289 shortName = some long name, 1.2.3.4
3292 *) Reimplemented the BN_CTX implementation. There is now no more static
3293 limitation on the number of variables it can handle nor the depth of the
3294 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
3295 information can now expand as required, and rather than having a single
3296 static array of bignums, BN_CTX now uses a linked-list of such arrays
3297 allowing it to expand on demand whilst maintaining the usefulness of
3298 BN_CTX's "bundling".
3301 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
3302 to allow all RSA operations to function using a single BN_CTX.
3305 *) Preliminary support for certificate policy evaluation and checking. This
3306 is initially intended to pass the tests outlined in "Conformance Testing
3307 of Relying Party Client Certificate Path Processing Logic" v1.07.
3310 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
3311 remained unused and not that useful. A variety of other little bignum
3312 tweaks and fixes have also been made continuing on from the audit (see
3316 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
3317 associated ASN1, EVP and SSL functions and old ASN1 macros.
3320 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
3321 and this should never fail. So the return value from the use of
3322 BN_set_word() (which can fail due to needless expansion) is now deprecated;
3323 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
3326 *) BN_CTX_get() should return zero-valued bignums, providing the same
3327 initialised value as BN_new().
3328 [Geoff Thorpe, suggested by Ulf Möller]
3330 *) Support for inhibitAnyPolicy certificate extension.
3333 *) An audit of the BIGNUM code is underway, for which debugging code is
3334 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
3335 is considered valid when processing BIGNUMs, and causes execution to
3336 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
3337 further steps are taken to deliberately pollute unused data in BIGNUM
3338 structures to try and expose faulty code further on. For now, openssl will
3339 (in its default mode of operation) continue to tolerate the inconsistent
3340 forms that it has tolerated in the past, but authors and packagers should
3341 consider trying openssl and their own applications when compiled with
3342 these debugging symbols defined. It will help highlight potential bugs in
3343 their own code, and will improve the test coverage for OpenSSL itself. At
3344 some point, these tighter rules will become openssl's default to improve
3345 maintainability, though the assert()s and other overheads will remain only
3346 in debugging configurations. See bn.h for more details.
3347 [Geoff Thorpe, Nils Larsch, Ulf Möller]
3349 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
3350 that can only be obtained through BN_CTX_new() (which implicitly
3351 initialises it). The presence of this function only made it possible
3352 to overwrite an existing structure (and cause memory leaks).
3355 *) Because of the callback-based approach for implementing LHASH as a
3356 template type, lh_insert() adds opaque objects to hash-tables and
3357 lh_doall() or lh_doall_arg() are typically used with a destructor callback
3358 to clean up those corresponding objects before destroying the hash table
3359 (and losing the object pointers). So some over-zealous constifications in
3360 LHASH have been relaxed so that lh_insert() does not take (nor store) the
3361 objects as "const" and the lh_doall[_arg] callback wrappers are not
3362 prototyped to have "const" restrictions on the object pointers they are
3363 given (and so aren't required to cast them away any more).
3366 *) The tmdiff.h API was so ugly and minimal that our own timing utility
3367 (speed) prefers to use its own implementation. The two implementations
3368 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
3369 its object type properly exposed (MS_TM) instead of casting to/from "char
3370 *". This may still change yet if someone realises MS_TM and "ms_time_***"
3371 aren't necessarily the greatest nomenclatures - but this is what was used
3372 internally to the implementation so I've used that for now.
3375 *) Ensure that deprecated functions do not get compiled when
3376 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
3377 the self-tests were still using deprecated key-generation functions so
3378 these have been updated also.
3381 *) Reorganise PKCS#7 code to separate the digest location functionality
3382 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
3383 New function PKCS7_set_digest() to set the digest type for PKCS#7
3384 digestedData type. Add additional code to correctly generate the
3385 digestedData type and add support for this type in PKCS7 initialization
3389 *) New function PKCS7_set0_type_other() this initializes a PKCS7
3390 structure of type "other".
3393 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
3394 sure the loop does correctly stop and breaking ("division by zero")
3395 modulus operations are not performed. The (pre-generated) prime
3396 table crypto/bn/bn_prime.h was already correct, but it could not be
3397 re-generated on some platforms because of the "division by zero"
3398 situation in the script.
3399 [Ralf S. Engelschall]
3401 *) Update support for ECC-based TLS ciphersuites according to
3402 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
3403 SHA-1 now is only used for "small" curves (where the
3404 representation of a field element takes up to 24 bytes); for
3405 larger curves, the field element resulting from ECDH is directly
3406 used as premaster secret.
3407 [Douglas Stebila (Sun Microsystems Laboratories)]
3409 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
3410 curve secp160r1 to the tests.
3411 [Douglas Stebila (Sun Microsystems Laboratories)]
3413 *) Add the possibility to load symbols globally with DSO.
3414 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
3416 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
3417 control of the error stack.
3420 *) Add support for STORE in ENGINE.
3423 *) Add the STORE type. The intention is to provide a common interface
3424 to certificate and key stores, be they simple file-based stores, or
3425 HSM-type store, or LDAP stores, or...
3426 NOTE: The code is currently UNTESTED and isn't really used anywhere.
3429 *) Add a generic structure called OPENSSL_ITEM. This can be used to
3430 pass a list of arguments to any function as well as provide a way
3431 for a function to pass data back to the caller.
3434 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
3435 works like BUF_strdup() but can be used to duplicate a portion of
3436 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
3440 *) Add the function sk_find_ex() which works like sk_find(), but will
3441 return an index to an element even if an exact match couldn't be
3442 found. The index is guaranteed to point at the element where the
3443 searched-for key would be inserted to preserve sorting order.
3446 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
3447 takes an extra flags argument for optional functionality. Currently,
3448 the following flags are defined:
3450 OBJ_BSEARCH_VALUE_ON_NOMATCH
3451 This one gets OBJ_bsearch_ex() to return a pointer to the first
3452 element where the comparing function returns a negative or zero
3455 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
3456 This one gets OBJ_bsearch_ex() to return a pointer to the first
3457 element where the comparing function returns zero. This is useful
3458 if there are more than one element where the comparing function
3462 *) Make it possible to create self-signed certificates with 'openssl ca'
3463 in such a way that the self-signed certificate becomes part of the
3464 CA database and uses the same mechanisms for serial number generation
3465 as all other certificate signing. The new flag '-selfsign' enables
3466 this functionality. Adapt CA.sh and CA.pl.in.
3469 *) Add functionality to check the public key of a certificate request
3470 against a given private. This is useful to check that a certificate
3471 request can be signed by that key (self-signing).
3474 *) Make it possible to have multiple active certificates with the same
3475 subject in the CA index file. This is done only if the keyword
3476 'unique_subject' is set to 'no' in the main CA section (default
3477 if 'CA_default') of the configuration file. The value is saved
3478 with the database itself in a separate index attribute file,
3479 named like the index file with '.attr' appended to the name.
3482 *) Generate muti valued AVAs using '+' notation in config files for
3486 *) Support for nameConstraints certificate extension.
3489 *) Support for policyConstraints certificate extension.
3492 *) Support for policyMappings certificate extension.
3495 *) Make sure the default DSA_METHOD implementation only uses its
3496 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
3497 and change its own handlers to be NULL so as to remove unnecessary
3498 indirection. This lets alternative implementations fallback to the
3499 default implementation more easily.
3502 *) Support for directoryName in GeneralName related extensions
3506 *) Make it possible to link applications using Makefile.shared.
3507 Make that possible even when linking against static libraries!
3510 *) Support for single pass processing for S/MIME signing. This now
3511 means that S/MIME signing can be done from a pipe, in addition
3512 cleartext signing (multipart/signed type) is effectively streaming
3513 and the signed data does not need to be all held in memory.
3515 This is done with a new flag PKCS7_STREAM. When this flag is set
3516 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
3517 is done after the data is output (and digests calculated) in
3518 SMIME_write_PKCS7().
3521 *) Add full support for -rpath/-R, both in shared libraries and
3522 applications, at least on the platforms where it's known how
3526 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
3527 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
3528 will now compute a table of multiples of the generator that
3529 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
3530 faster (notably in the case of a single point multiplication,
3531 scalar * generator).
3532 [Nils Larsch, Bodo Moeller]
3534 *) IPv6 support for certificate extensions. The various extensions
3535 which use the IP:a.b.c.d can now take IPv6 addresses using the
3536 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
3540 *) Added an ENGINE that implements RSA by performing private key
3541 exponentiations with the GMP library. The conversions to and from
3542 GMP's mpz_t format aren't optimised nor are any montgomery forms
3543 cached, and on x86 it appears OpenSSL's own performance has caught up.
3544 However there are likely to be other architectures where GMP could
3545 provide a boost. This ENGINE is not built in by default, but it can be
3546 specified at Configure time and should be accompanied by the necessary
3547 linker additions, eg;
3548 ./config -DOPENSSL_USE_GMP -lgmp
3551 *) "openssl engine" will not display ENGINE/DSO load failure errors when
3552 testing availability of engines with "-t" - the old behaviour is
3553 produced by increasing the feature's verbosity with "-tt".
3556 *) ECDSA routines: under certain error conditions uninitialized BN objects
3557 could be freed. Solution: make sure initialization is performed early
3558 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
3562 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3563 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3564 software implementations. For DSA and DH, parameter generation can
3565 also be overriden by providing the appropriate method callbacks.
3568 *) Change the "progress" mechanism used in key-generation and
3569 primality testing to functions that take a new BN_GENCB pointer in
3570 place of callback/argument pairs. The new API functions have "_ex"
3571 postfixes and the older functions are reimplemented as wrappers for
3572 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
3573 declarations of the old functions to help (graceful) attempts to
3574 migrate to the new functions. Also, the new key-generation API
3575 functions operate on a caller-supplied key-structure and return
3576 success/failure rather than returning a key or NULL - this is to
3577 help make "keygen" another member function of RSA_METHOD etc.
3579 Example for using the new callback interface:
3581 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
3585 BN_GENCB_set(&my_cb, my_callback, my_arg);
3587 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
3588 /* For the meaning of a, b in calls to my_callback(), see the
3589 * documentation of the function that calls the callback.
3590 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
3591 * my_callback should return 1 if it wants BN_is_prime_ex()
3592 * to continue, or 0 to stop.
3597 *) Change the ZLIB compression method to be stateful, and make it
3598 available to TLS with the number defined in
3599 draft-ietf-tls-compression-04.txt.
3602 *) Add the ASN.1 structures and functions for CertificatePair, which
3603 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
3605 CertificatePair ::= SEQUENCE {
3606 forward [0] Certificate OPTIONAL,
3607 reverse [1] Certificate OPTIONAL,
3608 -- at least one of the pair shall be present -- }
3610 Also implement the PEM functions to read and write certificate
3611 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
3613 This needed to be defined, mostly for the sake of the LDAP
3614 attribute crossCertificatePair, but may prove useful elsewhere as
3618 *) Make it possible to inhibit symlinking of shared libraries in
3619 Makefile.shared, for Cygwin's sake.
3622 *) Extend the BIGNUM API by creating a function
3623 void BN_set_negative(BIGNUM *a, int neg);
3624 and a macro that behave like
3625 int BN_is_negative(const BIGNUM *a);
3627 to avoid the need to access 'a->neg' directly in applications.
3630 *) Implement fast modular reduction for pseudo-Mersenne primes
3631 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
3632 EC_GROUP_new_curve_GFp() will now automatically use this
3634 [Nils Larsch <nla@trustcenter.de>]
3636 *) Add new lock type (CRYPTO_LOCK_BN).
3639 *) Change the ENGINE framework to automatically load engines
3640 dynamically from specific directories unless they could be
3641 found to already be built in or loaded. Move all the
3642 current engines except for the cryptodev one to a new
3644 The engines in engines/ are built as shared libraries if
3645 the "shared" options was given to ./Configure or ./config.
3646 Otherwise, they are inserted in libcrypto.a.
3647 /usr/local/ssl/engines is the default directory for dynamic
3648 engines, but that can be overriden at configure time through
3649 the usual use of --prefix and/or --openssldir, and at run
3650 time with the environment variable OPENSSL_ENGINES.
3651 [Geoff Thorpe and Richard Levitte]
3653 *) Add Makefile.shared, a helper makefile to build shared
3654 libraries. Addapt Makefile.org.
3657 *) Add version info to Win32 DLLs.
3658 [Peter 'Luna' Runestig" <peter@runestig.com>]
3660 *) Add new 'medium level' PKCS#12 API. Certificates and keys
3661 can be added using this API to created arbitrary PKCS#12
3662 files while avoiding the low level API.
3664 New options to PKCS12_create(), key or cert can be NULL and
3665 will then be omitted from the output file. The encryption
3666 algorithm NIDs can be set to -1 for no encryption, the mac
3667 iteration count can be set to 0 to omit the mac.
3669 Enhance pkcs12 utility by making the -nokeys and -nocerts
3670 options work when creating a PKCS#12 file. New option -nomac
3671 to omit the mac, NONE can be set for an encryption algorithm.
3672 New code is modified to use the enhanced PKCS12_create()
3673 instead of the low level API.
3676 *) Extend ASN1 encoder to support indefinite length constructed
3677 encoding. This can output sequences tags and octet strings in
3678 this form. Modify pk7_asn1.c to support indefinite length
3679 encoding. This is experimental and needs additional code to
3680 be useful, such as an ASN1 bio and some enhanced streaming
3683 Extend template encode functionality so that tagging is passed
3684 down to the template encoder.
3687 *) Let 'openssl req' fail if an argument to '-newkey' is not
3688 recognized instead of using RSA as a default.
3691 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
3692 As these are not official, they are not included in "ALL";
3693 the "ECCdraft" ciphersuite group alias can be used to select them.
3694 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
3696 *) Add ECDH engine support.
3697 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
3699 *) Add ECDH in new directory crypto/ecdh/.
3700 [Douglas Stebila (Sun Microsystems Laboratories)]
3702 *) Let BN_rand_range() abort with an error after 100 iterations
3703 without success (which indicates a broken PRNG).
3706 *) Change BN_mod_sqrt() so that it verifies that the input value
3707 is really the square of the return value. (Previously,
3708 BN_mod_sqrt would show GIGO behaviour.)
3711 *) Add named elliptic curves over binary fields from X9.62, SECG,
3712 and WAP/WTLS; add OIDs that were still missing.
3714 [Sheueling Chang Shantz and Douglas Stebila
3715 (Sun Microsystems Laboratories)]
3717 *) Extend the EC library for elliptic curves over binary fields
3718 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
3721 EC_GF2m_simple_method
3725 EC_GROUP_new_curve_GF2m
3726 EC_GROUP_set_curve_GF2m
3727 EC_GROUP_get_curve_GF2m
3728 EC_POINT_set_affine_coordinates_GF2m
3729 EC_POINT_get_affine_coordinates_GF2m
3730 EC_POINT_set_compressed_coordinates_GF2m
3732 Point compression for binary fields is disabled by default for
3733 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
3736 As binary polynomials are represented as BIGNUMs, various members
3737 of the EC_GROUP and EC_POINT data structures can be shared
3738 between the implementations for prime fields and binary fields;
3739 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
3740 are essentially identical to their ..._GFp counterparts.
3741 (For simplicity, the '..._GFp' prefix has been dropped from
3742 various internal method names.)
3744 An internal 'field_div' method (similar to 'field_mul' and
3745 'field_sqr') has been added; this is used only for binary fields.
3747 [Sheueling Chang Shantz and Douglas Stebila
3748 (Sun Microsystems Laboratories)]
3750 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
3751 through methods ('mul', 'precompute_mult').
3753 The generic implementations (now internally called 'ec_wNAF_mul'
3754 and 'ec_wNAF_precomputed_mult') remain the default if these
3755 methods are undefined.
3757 [Sheueling Chang Shantz and Douglas Stebila
3758 (Sun Microsystems Laboratories)]
3760 *) New function EC_GROUP_get_degree, which is defined through
3761 EC_METHOD. For curves over prime fields, this returns the bit
3762 length of the modulus.
3764 [Sheueling Chang Shantz and Douglas Stebila
3765 (Sun Microsystems Laboratories)]
3767 *) New functions EC_GROUP_dup, EC_POINT_dup.
3768 (These simply call ..._new and ..._copy).
3770 [Sheueling Chang Shantz and Douglas Stebila
3771 (Sun Microsystems Laboratories)]
3773 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
3774 Polynomials are represented as BIGNUMs (where the sign bit is not
3775 used) in the following functions [macros]:
3778 BN_GF2m_sub [= BN_GF2m_add]
3779 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
3780 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
3781 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3783 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3784 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3785 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3786 BN_GF2m_cmp [= BN_ucmp]
3788 (Note that only the 'mod' functions are actually for fields GF(2^m).
3789 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3791 For some functions, an the irreducible polynomial defining a
3792 field can be given as an 'unsigned int[]' with strictly
3793 decreasing elements giving the indices of those bits that are set;
3794 i.e., p[] represents the polynomial
3795 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3797 p[0] > p[1] > ... > p[k] = 0.
3798 This applies to the following functions:
3803 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3804 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3806 BN_GF2m_mod_sqrt_arr
3807 BN_GF2m_mod_solve_quad_arr
3811 Conversion can be performed by the following functions:
3816 bntest.c has additional tests for binary polynomial arithmetic.
3818 Two implementations for BN_GF2m_mod_div() are available.
3819 The default algorithm simply uses BN_GF2m_mod_inv() and
3820 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3821 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3822 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3824 [Sheueling Chang Shantz and Douglas Stebila
3825 (Sun Microsystems Laboratories)]
3827 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3828 functionality is disabled at compile-time.
3829 [Douglas Stebila <douglas.stebila@sun.com>]
3831 *) Change default behaviour of 'openssl asn1parse' so that more
3832 information is visible when viewing, e.g., a certificate:
3834 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3835 mode the content of non-printable OCTET STRINGs is output in a
3836 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3837 avoid the appearance of a printable string.
3838 [Nils Larsch <nla@trustcenter.de>]
3840 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3842 EC_GROUP_set_asn1_flag()
3843 EC_GROUP_get_asn1_flag()
3844 EC_GROUP_set_point_conversion_form()
3845 EC_GROUP_get_point_conversion_form()
3846 These control ASN1 encoding details:
3847 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3848 has been set to OPENSSL_EC_NAMED_CURVE.
3849 - Points are encoded in uncompressed form by default; options for
3850 asn1_for are as for point2oct, namely
3851 POINT_CONVERSION_COMPRESSED
3852 POINT_CONVERSION_UNCOMPRESSED
3853 POINT_CONVERSION_HYBRID
3855 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3858 EC_GROUP_get0_seed()
3859 EC_GROUP_get_seed_len()
3860 This is used only for ASN1 purposes (so far).
3861 [Nils Larsch <nla@trustcenter.de>]
3863 *) Add 'field_type' member to EC_METHOD, which holds the NID
3864 of the appropriate field type OID. The new function
3865 EC_METHOD_get_field_type() returns this value.
3866 [Nils Larsch <nla@trustcenter.de>]
3871 EC_POINT_point2hex()
3872 EC_POINT_hex2point()
3873 providing useful interfaces to EC_POINT_point2oct() and
3874 EC_POINT_oct2point().
3875 [Nils Larsch <nla@trustcenter.de>]
3877 *) Change internals of the EC library so that the functions
3878 EC_GROUP_set_generator()
3879 EC_GROUP_get_generator()
3880 EC_GROUP_get_order()
3881 EC_GROUP_get_cofactor()
3882 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3883 to methods, which would lead to unnecessary code duplication when
3884 adding different types of curves.
3885 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3887 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3888 arithmetic, and such that modified wNAFs are generated
3889 (which avoid length expansion in many cases).
3892 *) Add a function EC_GROUP_check_discriminant() (defined via
3893 EC_METHOD) that verifies that the curve discriminant is non-zero.
3895 Add a function EC_GROUP_check() that makes some sanity tests
3896 on a EC_GROUP, its generator and order. This includes
3897 EC_GROUP_check_discriminant().
3898 [Nils Larsch <nla@trustcenter.de>]
3900 *) Add ECDSA in new directory crypto/ecdsa/.
3902 Add applications 'openssl ecparam' and 'openssl ecdsa'
3903 (these are based on 'openssl dsaparam' and 'openssl dsa').
3905 ECDSA support is also included in various other files across the
3906 library. Most notably,
3907 - 'openssl req' now has a '-newkey ecdsa:file' option;
3908 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3909 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3910 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3911 them suitable for ECDSA where domain parameters must be
3912 extracted before the specific public key;
3913 - ECDSA engine support has been added.
3914 [Nils Larsch <nla@trustcenter.de>]
3916 *) Include some named elliptic curves, and add OIDs from X9.62,
3917 SECG, and WAP/WTLS. Each curve can be obtained from the new
3919 EC_GROUP_new_by_curve_name(),
3920 and the list of available named curves can be obtained with
3921 EC_get_builtin_curves().
3922 Also add a 'curve_name' member to EC_GROUP objects, which can be
3924 EC_GROUP_set_curve_name()
3925 EC_GROUP_get_curve_name()
3926 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3928 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3929 was actually never needed) and in BN_mul(). The removal in BN_mul()
3930 required a small change in bn_mul_part_recursive() and the addition
3931 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3932 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3933 bn_sub_words() and bn_add_words() except they take arrays with
3937 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3939 *) Cleanse PEM buffers before freeing them since they may contain
3941 [Benjamin Bennett <ben@psc.edu>]
3943 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3944 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3945 authentication-only ciphersuites.
3948 *) Since AES128 and AES256 share a single mask bit in the logic of
3949 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3950 kludge to work properly if AES128 is available and AES256 isn't.
3953 *) Expand security boundary to match 1.1.1 module.
3956 *) Remove redundant features: hash file source, editing of test vectors
3957 modify fipsld to use external fips_premain.c signature.
3960 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3961 run algorithm test programs.
3964 *) Make algorithm test programs more tolerant of whitespace.
3967 *) Have SSL/TLS server implementation tolerate "mismatched" record
3968 protocol version while receiving ClientHello even if the
3969 ClientHello is fragmented. (The server can't insist on the
3970 particular protocol version it has chosen before the ServerHello
3971 message has informed the client about his choice.)
3974 *) Load error codes if they are not already present instead of using a
3975 static variable. This allows them to be cleanly unloaded and reloaded.
3978 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3980 *) Introduce limits to prevent malicious keys being able to
3981 cause a denial of service. (CVE-2006-2940)
3982 [Steve Henson, Bodo Moeller]
3984 *) Fix ASN.1 parsing of certain invalid structures that can result
3985 in a denial of service. (CVE-2006-2937) [Steve Henson]
3987 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3988 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3990 *) Fix SSL client code which could crash if connecting to a
3991 malicious SSLv2 server. (CVE-2006-4343)
3992 [Tavis Ormandy and Will Drewry, Google Security Team]
3994 *) Change ciphersuite string processing so that an explicit
3995 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3996 will no longer include "AES128-SHA"), and any other similar
3997 ciphersuite (same bitmap) from *other* protocol versions (so that
3998 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3999 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
4000 changes from 0.9.8b and 0.9.8d.
4003 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
4005 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4006 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4008 *) Change the Unix randomness entropy gathering to use poll() when
4009 possible instead of select(), since the latter has some
4010 undesirable limitations.
4011 [Darryl Miles via Richard Levitte and Bodo Moeller]
4013 *) Disable rogue ciphersuites:
4015 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4016 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4017 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4019 The latter two were purportedly from
4020 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4023 Also deactive the remaining ciphersuites from
4024 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4025 unofficial, and the ID has long expired.
4028 *) Fix RSA blinding Heisenbug (problems sometimes occured on
4029 dual-core machines) and other potential thread-safety issues.
4032 Changes between 0.9.7i and 0.9.7j [04 May 2006]
4034 *) Adapt fipsld and the build system to link against the validated FIPS
4035 module in FIPS mode.
4038 *) Fixes for VC++ 2005 build under Windows.
4041 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
4042 from a Windows bash shell such as MSYS. It is autodetected from the
4043 "config" script when run from a VC++ environment. Modify standard VC++
4044 build to use fipscanister.o from the GNU make build.
4047 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
4049 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
4050 The value now differs depending on if you build for FIPS or not.
4051 BEWARE! A program linked with a shared FIPSed libcrypto can't be
4052 safely run with a non-FIPSed libcrypto, as it may crash because of
4053 the difference induced by this change.
4056 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
4058 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4059 (part of SSL_OP_ALL). This option used to disable the
4060 countermeasure against man-in-the-middle protocol-version
4061 rollback in the SSL 2.0 server implementation, which is a bad
4062 idea. (CVE-2005-2969)
4064 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4065 for Information Security, National Institute of Advanced Industrial
4066 Science and Technology [AIST], Japan)]
4068 *) Minimal support for X9.31 signatures and PSS padding modes. This is
4069 mainly for FIPS compliance and not fully integrated at this stage.
4072 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
4073 the exponentiation using a fixed-length exponent. (Otherwise,
4074 the information leaked through timing could expose the secret key
4075 after many signatures; cf. Bleichenbacher's attack on DSA with
4079 *) Make a new fixed-window mod_exp implementation the default for
4080 RSA, DSA, and DH private-key operations so that the sequence of
4081 squares and multiplies and the memory access pattern are
4082 independent of the particular secret key. This will mitigate
4083 cache-timing and potential related attacks.
4085 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
4086 and this is automatically used by BN_mod_exp_mont() if the new flag
4087 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
4088 will use this BN flag for private exponents unless the flag
4089 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
4090 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
4092 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
4094 *) Change the client implementation for SSLv23_method() and
4095 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
4096 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
4097 (Previously, the SSL 2.0 backwards compatible Client Hello
4098 message format would be used even with SSL_OP_NO_SSLv2.)
4101 *) Add support for smime-type MIME parameter in S/MIME messages which some
4105 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
4106 a threadsafe manner. Modify rsa code to use new function and add calls
4107 to dsa and dh code (which had race conditions before).
4110 *) Include the fixed error library code in the C error file definitions
4111 instead of fixing them up at runtime. This keeps the error code
4112 structures constant.
4115 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
4117 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
4120 *) Fixes for newer kerberos headers. NB: the casts are needed because
4121 the 'length' field is signed on one version and unsigned on another
4122 with no (?) obvious way to tell the difference, without these VC++
4123 complains. Also the "definition" of FAR (blank) is no longer included
4124 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
4125 some needed definitions.
4128 *) Undo Cygwin change.
4131 *) Added support for proxy certificates according to RFC 3820.
4132 Because they may be a security thread to unaware applications,
4133 they must be explicitely allowed in run-time. See
4134 docs/HOWTO/proxy_certificates.txt for further information.
4137 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
4139 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
4140 server and client random values. Previously
4141 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
4142 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
4144 This change has negligible security impact because:
4146 1. Server and client random values still have 24 bytes of pseudo random
4149 2. Server and client random values are sent in the clear in the initial
4152 3. The master secret is derived using the premaster secret (48 bytes in
4153 size for static RSA ciphersuites) as well as client server and random
4156 The OpenSSL team would like to thank the UK NISCC for bringing this issue
4159 [Stephen Henson, reported by UK NISCC]
4161 *) Use Windows randomness collection on Cygwin.
4164 *) Fix hang in EGD/PRNGD query when communication socket is closed
4165 prematurely by EGD/PRNGD.
4166 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
4168 *) Prompt for pass phrases when appropriate for PKCS12 input format.
4171 *) Back-port of selected performance improvements from development
4172 branch, as well as improved support for PowerPC platforms.
4175 *) Add lots of checks for memory allocation failure, error codes to indicate
4176 failure and freeing up memory if a failure occurs.
4177 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
4179 *) Add new -passin argument to dgst.
4182 *) Perform some character comparisons of different types in X509_NAME_cmp:
4183 this is needed for some certificates that reencode DNs into UTF8Strings
4184 (in violation of RFC3280) and can't or wont issue name rollover
4188 *) Make an explicit check during certificate validation to see that
4189 the CA setting in each certificate on the chain is correct. As a
4190 side effect always do the following basic checks on extensions,
4191 not just when there's an associated purpose to the check:
4193 - if there is an unhandled critical extension (unless the user
4194 has chosen to ignore this fault)
4195 - if the path length has been exceeded (if one is set at all)
4196 - that certain extensions fit the associated purpose (if one has
4200 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
4202 *) Avoid a race condition when CRLs are checked in a multi threaded
4203 environment. This would happen due to the reordering of the revoked
4204 entries during signature checking and serial number lookup. Now the
4205 encoding is cached and the serial number sort performed under a lock.
4206 Add new STACK function sk_is_sorted().
4209 *) Add Delta CRL to the extension code.
4212 *) Various fixes to s3_pkt.c so alerts are sent properly.
4213 [David Holmes <d.holmes@f5.com>]
4215 *) Reduce the chances of duplicate issuer name and serial numbers (in
4216 violation of RFC3280) using the OpenSSL certificate creation utilities.
4217 This is done by creating a random 64 bit value for the initial serial
4218 number when a serial number file is created or when a self signed
4219 certificate is created using 'openssl req -x509'. The initial serial
4220 number file is created using 'openssl x509 -next_serial' in CA.pl
4221 rather than being initialized to 1.
4224 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
4226 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4227 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4228 [Joe Orton, Steve Henson]
4230 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
4232 [Joe Orton, Steve Henson]
4234 *) Make it possible to have multiple active certificates with the same
4235 subject in the CA index file. This is done only if the keyword
4236 'unique_subject' is set to 'no' in the main CA section (default
4237 if 'CA_default') of the configuration file. The value is saved
4238 with the database itself in a separate index attribute file,
4239 named like the index file with '.attr' appended to the name.
4242 *) X509 verify fixes. Disable broken certificate workarounds when
4243 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
4244 keyUsage extension present. Don't accept CRLs with unhandled critical
4245 extensions: since verify currently doesn't process CRL extensions this
4246 rejects a CRL with *any* critical extensions. Add new verify error codes
4250 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
4251 A clarification of RFC2560 will require the use of OCTET STRINGs and
4252 some implementations cannot handle the current raw format. Since OpenSSL
4253 copies and compares OCSP nonces as opaque blobs without any attempt at
4254 parsing them this should not create any compatibility issues.
4257 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
4258 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
4259 this HMAC (and other) operations are several times slower than OpenSSL
4263 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
4264 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4266 *) Use the correct content when signing type "other".
4269 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
4271 *) Fix various bugs revealed by running the NISCC test suite:
4273 Stop out of bounds reads in the ASN1 code when presented with
4274 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4276 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
4278 If verify callback ignores invalid public key errors don't try to check
4279 certificate signature with the NULL public key.
4283 *) New -ignore_err option in ocsp application to stop the server
4284 exiting on the first error in a request.
4287 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4288 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4292 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4293 extra data after the compression methods not only for TLS 1.0
4294 but also for SSL 3.0 (as required by the specification).
4295 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4297 *) Change X509_certificate_type() to mark the key as exported/exportable
4298 when it's 512 *bits* long, not 512 bytes.
4301 *) Change AES_cbc_encrypt() so it outputs exact multiple of
4302 blocks during encryption.
4305 *) Various fixes to base64 BIO and non blocking I/O. On write
4306 flushes were not handled properly if the BIO retried. On read
4307 data was not being buffered properly and had various logic bugs.
4308 This also affects blocking I/O when the data being decoded is a
4312 *) Various S/MIME bugfixes and compatibility changes:
4313 output correct application/pkcs7 MIME type if
4314 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
4315 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
4316 of files as .eml work). Correctly handle very long lines in MIME
4320 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
4322 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4323 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4324 a protocol version number mismatch like a decryption error
4325 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4328 *) Turn on RSA blinding by default in the default implementation
4329 to avoid a timing attack. Applications that don't want it can call
4330 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4331 They would be ill-advised to do so in most cases.
4332 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4334 *) Change RSA blinding code so that it works when the PRNG is not
4335 seeded (in this case, the secret RSA exponent is abused as
4336 an unpredictable seed -- if it is not unpredictable, there
4337 is no point in blinding anyway). Make RSA blinding thread-safe
4338 by remembering the creator's thread ID in rsa->blinding and
4339 having all other threads use local one-time blinding factors
4340 (this requires more computation than sharing rsa->blinding, but
4341 avoids excessive locking; and if an RSA object is not shared
4342 between threads, blinding will still be very fast).
4345 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
4346 ENGINE as defaults for all supported algorithms irrespective of
4347 the 'flags' parameter. 'flags' is now honoured, so applications
4348 should make sure they are passing it correctly.
4351 *) Target "mingw" now allows native Windows code to be generated in
4352 the Cygwin environment as well as with the MinGW compiler.
4355 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
4357 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4358 via timing by performing a MAC computation even if incorrrect
4359 block cipher padding has been found. This is a countermeasure
4360 against active attacks where the attacker has to distinguish
4361 between bad padding and a MAC verification error. (CVE-2003-0078)
4363 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4364 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4365 Martin Vuagnoux (EPFL, Ilion)]
4367 *) Make the no-err option work as intended. The intention with no-err
4368 is not to have the whole error stack handling routines removed from
4369 libcrypto, it's only intended to remove all the function name and
4370 reason texts, thereby removing some of the footprint that may not
4371 be interesting if those errors aren't displayed anyway.
4373 NOTE: it's still possible for any application or module to have it's
4374 own set of error texts inserted. The routines are there, just not
4375 used by default when no-err is given.
4378 *) Add support for FreeBSD on IA64.
4379 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
4381 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
4382 Kerberos function mit_des_cbc_cksum(). Before this change,
4383 the value returned by DES_cbc_cksum() was like the one from
4384 mit_des_cbc_cksum(), except the bytes were swapped.
4385 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
4387 *) Allow an application to disable the automatic SSL chain building.
4388 Before this a rather primitive chain build was always performed in
4389 ssl3_output_cert_chain(): an application had no way to send the
4390 correct chain if the automatic operation produced an incorrect result.
4392 Now the chain builder is disabled if either:
4394 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
4396 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
4398 The reasoning behind this is that an application would not want the
4399 auto chain building to take place if extra chain certificates are
4400 present and it might also want a means of sending no additional
4401 certificates (for example the chain has two certificates and the
4405 *) Add the possibility to build without the ENGINE framework.
4406 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4408 *) Under Win32 gmtime() can return NULL: check return value in
4409 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
4412 *) DSA routines: under certain error conditions uninitialized BN objects
4413 could be freed. Solution: make sure initialization is performed early
4414 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
4415 Nils Larsch <nla@trustcenter.de> via PR#459)
4418 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
4419 checked on reconnect on the client side, therefore session resumption
4420 could still fail with a "ssl session id is different" error. This
4421 behaviour is masked when SSL_OP_ALL is used due to
4422 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
4423 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4424 followup to PR #377.
4427 *) IA-32 assembler support enhancements: unified ELF targets, support
4428 for SCO/Caldera platforms, fix for Cygwin shared build.
4431 *) Add support for FreeBSD on sparc64. As a consequence, support for
4432 FreeBSD on non-x86 processors is separate from x86 processors on
4433 the config script, much like the NetBSD support.
4434 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
4436 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
4438 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
4441 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
4442 code (06) was taken as the first octet of the session ID and the last
4443 octet was ignored consequently. As a result SSLv2 client side session
4444 caching could not have worked due to the session ID mismatch between
4446 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4450 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
4451 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
4455 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
4456 seems that in spite of existing for more than a year, many application
4457 author have done nothing to provide the necessary callbacks, which
4458 means that this particular engine will not work properly anywhere.
4459 This is a very unfortunate situation which forces us, in the name
4460 of usability, to give the hw_ncipher.c a static lock, which is part
4462 NOTE: This is for the 0.9.7 series ONLY. This hack will never
4463 appear in 0.9.8 or later. We EXPECT application authors to have
4464 dealt properly with this when 0.9.8 is released (unless we actually
4465 make such changes in the libcrypto locking code that changes will
4466 have to be made anyway).
4469 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
4470 octets have been read, EOF or an error occurs. Without this change
4471 some truncated ASN1 structures will not produce an error.
4474 *) Disable Heimdal support, since it hasn't been fully implemented.
4475 Still give the possibility to force the use of Heimdal, but with
4476 warnings and a request that patches get sent to openssl-dev.
4479 *) Add the VC-CE target, introduce the WINCE sysname, and add
4480 INSTALL.WCE and appropriate conditionals to make it build.
4481 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4483 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
4484 cygssl-x.y.z.dll, where x, y and z are the major, minor and
4485 edit numbers of the version.
4486 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4488 *) Introduce safe string copy and catenation functions
4489 (BUF_strlcpy() and BUF_strlcat()).
4490 [Ben Laurie (CHATS) and Richard Levitte]
4492 *) Avoid using fixed-size buffers for one-line DNs.
4493 [Ben Laurie (CHATS)]
4495 *) Add BUF_MEM_grow_clean() to avoid information leakage when
4496 resizing buffers containing secrets, and use where appropriate.
4497 [Ben Laurie (CHATS)]
4499 *) Avoid using fixed size buffers for configuration file location.
4500 [Ben Laurie (CHATS)]
4502 *) Avoid filename truncation for various CA files.
4503 [Ben Laurie (CHATS)]
4505 *) Use sizeof in preference to magic numbers.
4506 [Ben Laurie (CHATS)]
4508 *) Avoid filename truncation in cert requests.
4509 [Ben Laurie (CHATS)]
4511 *) Add assertions to check for (supposedly impossible) buffer
4513 [Ben Laurie (CHATS)]
4515 *) Don't cache truncated DNS entries in the local cache (this could
4516 potentially lead to a spoofing attack).
4517 [Ben Laurie (CHATS)]
4519 *) Fix various buffers to be large enough for hex/decimal
4520 representations in a platform independent manner.
4521 [Ben Laurie (CHATS)]
4523 *) Add CRYPTO_realloc_clean() to avoid information leakage when
4524 resizing buffers containing secrets, and use where appropriate.
4525 [Ben Laurie (CHATS)]
4527 *) Add BIO_indent() to avoid much slightly worrying code to do
4529 [Ben Laurie (CHATS)]
4531 *) Convert sprintf()/BIO_puts() to BIO_printf().
4532 [Ben Laurie (CHATS)]
4534 *) buffer_gets() could terminate with the buffer only half
4536 [Ben Laurie (CHATS)]
4538 *) Add assertions to prevent user-supplied crypto functions from
4539 overflowing internal buffers by having large block sizes, etc.
4540 [Ben Laurie (CHATS)]
4542 *) New OPENSSL_assert() macro (similar to assert(), but enabled
4544 [Ben Laurie (CHATS)]
4546 *) Eliminate unused copy of key in RC4.
4547 [Ben Laurie (CHATS)]
4549 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
4550 [Ben Laurie (CHATS)]
4552 *) Fix off-by-one error in EGD path.
4553 [Ben Laurie (CHATS)]
4555 *) If RANDFILE path is too long, ignore instead of truncating.
4556 [Ben Laurie (CHATS)]
4558 *) Eliminate unused and incorrectly sized X.509 structure
4560 [Ben Laurie (CHATS)]
4562 *) Eliminate unused and dangerous function knumber().
4563 [Ben Laurie (CHATS)]
4565 *) Eliminate unused and dangerous structure, KSSL_ERR.
4566 [Ben Laurie (CHATS)]
4568 *) Protect against overlong session ID context length in an encoded
4569 session object. Since these are local, this does not appear to be
4571 [Ben Laurie (CHATS)]
4573 *) Change from security patch (see 0.9.6e below) that did not affect
4574 the 0.9.6 release series:
4576 Remote buffer overflow in SSL3 protocol - an attacker could
4577 supply an oversized master key in Kerberos-enabled versions.
4579 [Ben Laurie (CHATS)]
4581 *) Change the SSL kerb5 codes to match RFC 2712.
4584 *) Make -nameopt work fully for req and add -reqopt switch.
4585 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
4587 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
4588 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
4590 *) Make sure tests can be performed even if the corresponding algorithms
4591 have been removed entirely. This was also the last step to make
4592 OpenSSL compilable with DJGPP under all reasonable conditions.
4593 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
4595 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
4596 to allow version independent disabling of normally unselected ciphers,
4597 which may be activated as a side-effect of selecting a single cipher.
4599 (E.g., cipher list string "RSA" enables ciphersuites that are left
4600 out of "ALL" because they do not provide symmetric encryption.
4601 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
4602 [Lutz Jaenicke, Bodo Moeller]
4604 *) Add appropriate support for separate platform-dependent build
4605 directories. The recommended way to make a platform-dependent
4606 build directory is the following (tested on Linux), maybe with
4609 # Place yourself outside of the OpenSSL source tree. In
4610 # this example, the environment variable OPENSSL_SOURCE
4611 # is assumed to contain the absolute OpenSSL source directory.
4612 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
4613 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
4614 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
4615 mkdir -p `dirname $F`
4616 ln -s $OPENSSL_SOURCE/$F $F
4619 To be absolutely sure not to disturb the source tree, a "make clean"
4620 is a good thing. If it isn't successfull, don't worry about it,
4621 it probably means the source directory is very clean.
4624 *) Make sure any ENGINE control commands make local copies of string
4625 pointers passed to them whenever necessary. Otherwise it is possible
4626 the caller may have overwritten (or deallocated) the original string
4627 data when a later ENGINE operation tries to use the stored values.
4628 [Götz Babin-Ebell <babinebell@trustcenter.de>]
4630 *) Improve diagnostics in file reading and command-line digests.
4631 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
4633 *) Add AES modes CFB and OFB to the object database. Correct an
4634 error in AES-CFB decryption.
4637 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
4638 allows existing EVP_CIPHER_CTX structures to be reused after
4639 calling EVP_*Final(). This behaviour is used by encryption
4640 BIOs and some applications. This has the side effect that
4641 applications must explicitly clean up cipher contexts with
4642 EVP_CIPHER_CTX_cleanup() or they will leak memory.
4645 *) Check the values of dna and dnb in bn_mul_recursive before calling
4646 bn_mul_comba (a non zero value means the a or b arrays do not contain
4647 n2 elements) and fallback to bn_mul_normal if either is not zero.
4650 *) Fix escaping of non-ASCII characters when using the -subj option
4651 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
4654 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
4655 form for "surname", serialNumber has no short form.
4656 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
4657 therefore remove "mail" short name for "internet 7".
4658 The OID for unique identifiers in X509 certificates is
4659 x500UniqueIdentifier, not uniqueIdentifier.
4660 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
4663 *) Add an "init" command to the ENGINE config module and auto initialize
4664 ENGINEs. Without any "init" command the ENGINE will be initialized
4665 after all ctrl commands have been executed on it. If init=1 the
4666 ENGINE is initailized at that point (ctrls before that point are run
4667 on the uninitialized ENGINE and after on the initialized one). If
4668 init=0 then the ENGINE will not be iniatialized at all.
4671 *) Fix the 'app_verify_callback' interface so that the user-defined
4672 argument is actually passed to the callback: In the
4673 SSL_CTX_set_cert_verify_callback() prototype, the callback
4674 declaration has been changed from
4677 int (*cb)(X509_STORE_CTX *,void *);
4678 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
4679 i=s->ctx->app_verify_callback(&ctx)
4680 has been changed into
4681 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
4683 To update applications using SSL_CTX_set_cert_verify_callback(),
4684 a dummy argument can be added to their callback functions.
4685 [D. K. Smetters <smetters@parc.xerox.com>]
4687 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
4688 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
4690 *) Add and OPENSSL_LOAD_CONF define which will cause
4691 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
4692 This allows older applications to transparently support certain
4693 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
4694 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
4695 load the config file and OPENSSL_add_all_algorithms_conf() which will
4696 always load it have also been added.
4699 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
4700 Adjust NIDs and EVP layer.
4701 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4703 *) Config modules support in openssl utility.
4705 Most commands now load modules from the config file,
4706 though in a few (such as version) this isn't done
4707 because it couldn't be used for anything.
4709 In the case of ca and req the config file used is
4710 the same as the utility itself: that is the -config
4711 command line option can be used to specify an
4715 *) Move default behaviour from OPENSSL_config(). If appname is NULL
4716 use "openssl_conf" if filename is NULL use default openssl config file.
4719 *) Add an argument to OPENSSL_config() to allow the use of an alternative
4720 config section name. Add a new flag to tolerate a missing config file
4721 and move code to CONF_modules_load_file().
4724 *) Support for crypto accelerator cards from Accelerated Encryption
4725 Processing, www.aep.ie. (Use engine 'aep')
4726 The support was copied from 0.9.6c [engine] and adapted/corrected
4727 to work with the new engine framework.
4728 [AEP Inc. and Richard Levitte]
4730 *) Support for SureWare crypto accelerator cards from Baltimore
4731 Technologies. (Use engine 'sureware')
4732 The support was copied from 0.9.6c [engine] and adapted
4733 to work with the new engine framework.
4736 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
4737 make the newer ENGINE framework commands for the CHIL engine work.
4738 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
4740 *) Make it possible to produce shared libraries on ReliantUNIX.
4741 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
4743 *) Add the configuration target debug-linux-ppro.
4744 Make 'openssl rsa' use the general key loading routines
4745 implemented in apps.c, and make those routines able to
4746 handle the key format FORMAT_NETSCAPE and the variant
4748 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4750 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4751 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4753 *) Add -keyform to rsautl, and document -engine.
4754 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
4756 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
4757 BIO_R_NO_SUCH_FILE error code rather than the generic
4758 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
4761 *) Add new functions
4763 ERR_peek_last_error_line
4764 ERR_peek_last_error_line_data.
4765 These are similar to
4768 ERR_peek_error_line_data,
4769 but report on the latest error recorded rather than the first one
4770 still in the error queue.
4771 [Ben Laurie, Bodo Moeller]
4773 *) default_algorithms option in ENGINE config module. This allows things
4775 default_algorithms = ALL
4776 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
4779 *) Prelminary ENGINE config module.
4782 *) New experimental application configuration code.
4785 *) Change the AES code to follow the same name structure as all other
4786 symmetric ciphers, and behave the same way. Move everything to
4787 the directory crypto/aes, thereby obsoleting crypto/rijndael.
4788 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4790 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
4791 [Ben Laurie and Theo de Raadt]
4793 *) Add option to output public keys in req command.
4794 [Massimiliano Pala madwolf@openca.org]
4796 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4797 (up to about 10% better than before for P-192 and P-224).
4800 *) New functions/macros
4802 SSL_CTX_set_msg_callback(ctx, cb)
4803 SSL_CTX_set_msg_callback_arg(ctx, arg)
4804 SSL_set_msg_callback(ssl, cb)
4805 SSL_set_msg_callback_arg(ssl, arg)
4807 to request calling a callback function
4809 void cb(int write_p, int version, int content_type,
4810 const void *buf, size_t len, SSL *ssl, void *arg)
4812 whenever a protocol message has been completely received
4813 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4814 protocol version according to which the SSL library interprets
4815 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4816 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4817 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4818 specification (change_cipher_spec(20), alert(21), handshake(22)).
4819 'buf' and 'len' point to the actual message, 'ssl' to the
4820 SSL object, and 'arg' is the application-defined value set by
4821 SSL[_CTX]_set_msg_callback_arg().
4823 'openssl s_client' and 'openssl s_server' have new '-msg' options
4824 to enable a callback that displays all protocol messages.
4827 *) Change the shared library support so shared libraries are built as
4828 soon as the corresponding static library is finished, and thereby get
4829 openssl and the test programs linked against the shared library.
4830 This still only happens when the keyword "shard" has been given to
4831 the configuration scripts.
4833 NOTE: shared library support is still an experimental thing, and
4834 backward binary compatibility is still not guaranteed.
4835 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4837 *) Add support for Subject Information Access extension.
4838 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4840 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4841 additional bytes when new memory had to be allocated, not just
4842 when reusing an existing buffer.
4845 *) New command line and configuration option 'utf8' for the req command.
4846 This allows field values to be specified as UTF8 strings.
4849 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4850 runs for the former and machine-readable output for the latter.
4853 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4854 of the e-mail address in the DN (i.e., it will go into a certificate
4855 extension only). The new configuration file option 'email_in_dn = no'
4856 has the same effect.
4857 [Massimiliano Pala madwolf@openca.org]
4859 *) Change all functions with names starting with des_ to be starting
4860 with DES_ instead. Add wrappers that are compatible with libdes,
4861 but are named _ossl_old_des_*. Finally, add macros that map the
4862 des_* symbols to the corresponding _ossl_old_des_* if libdes
4863 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4864 desired, the des_* symbols will be mapped to DES_*, with one
4867 Since we provide two compatibility mappings, the user needs to
4868 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4869 compatibility is desired. The default (i.e., when that macro
4870 isn't defined) is OpenSSL 0.9.6c compatibility.
4872 There are also macros that enable and disable the support of old
4873 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4874 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4875 are defined, the default will apply: to support the old des routines.
4877 In either case, one must include openssl/des.h to get the correct
4878 definitions. Do not try to just include openssl/des_old.h, that
4881 NOTE: This is a major break of an old API into a new one. Software
4882 authors are encouraged to switch to the DES_ style functions. Some
4883 time in the future, des_old.h and the libdes compatibility functions
4884 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4885 default), and then completely removed.
4888 *) Test for certificates which contain unsupported critical extensions.
4889 If such a certificate is found during a verify operation it is
4890 rejected by default: this behaviour can be overridden by either
4891 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4892 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4893 X509_supported_extension() has also been added which returns 1 if a
4894 particular extension is supported.
4897 *) Modify the behaviour of EVP cipher functions in similar way to digests
4898 to retain compatibility with existing code.
4901 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4902 compatibility with existing code. In particular the 'ctx' parameter does
4903 not have to be to be initialized before the call to EVP_DigestInit() and
4904 it is tidied up after a call to EVP_DigestFinal(). New function
4905 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4906 EVP_MD_CTX_copy() changed to not require the destination to be
4907 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4908 requires the destination to be valid.
4910 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4911 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4914 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4915 so that complete 'Handshake' protocol structures are kept in memory
4916 instead of overwriting 'msg_type' and 'length' with 'body' data.
4919 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4920 [Massimo Santin via Richard Levitte]
4922 *) Major restructuring to the underlying ENGINE code. This includes
4923 reduction of linker bloat, separation of pure "ENGINE" manipulation
4924 (initialisation, etc) from functionality dealing with implementations
4925 of specific crypto iterfaces. This change also introduces integrated
4926 support for symmetric ciphers and digest implementations - so ENGINEs
4927 can now accelerate these by providing EVP_CIPHER and EVP_MD
4928 implementations of their own. This is detailed in crypto/engine/README
4929 as it couldn't be adequately described here. However, there are a few
4930 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4931 were changed in the original introduction of ENGINE code have now
4932 reverted back - the hooking from this code to ENGINE is now a good
4933 deal more passive and at run-time, operations deal directly with
4934 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4935 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4936 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4937 they were not being used by the framework as there is no concept of a
4938 BIGNUM_METHOD and they could not be generalised to the new
4939 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4940 ENGINE_cpy() has been removed as it cannot be consistently defined in
4944 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4947 *) Change mkdef.pl to sort symbols that get the same entry number,
4948 and make sure the automatically generated functions ERR_load_*
4949 become part of libeay.num as well.
4952 *) New function SSL_renegotiate_pending(). This returns true once
4953 renegotiation has been requested (either SSL_renegotiate() call
4954 or HelloRequest/ClientHello receveived from the peer) and becomes
4955 false once a handshake has been completed.
4956 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4957 sends a HelloRequest, but does not ensure that a handshake takes
4958 place. SSL_renegotiate_pending() is useful for checking if the
4959 client has followed the request.)
4962 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4963 By default, clients may request session resumption even during
4964 renegotiation (if session ID contexts permit); with this option,
4965 session resumption is possible only in the first handshake.
4967 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4968 more bits available for options that should not be part of
4969 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4972 *) Add some demos for certificate and certificate request creation.
4975 *) Make maximum certificate chain size accepted from the peer application
4976 settable (SSL*_get/set_max_cert_list()), as proposed by
4977 "Douglas E. Engert" <deengert@anl.gov>.
4980 *) Add support for shared libraries for Unixware-7
4981 (Boyd Lynn Gerber <gerberb@zenez.com>).
4984 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4985 be done prior to destruction. Use this to unload error strings from
4986 ENGINEs that load their own error strings. NB: This adds two new API
4987 functions to "get" and "set" this destroy handler in an ENGINE.
4990 *) Alter all existing ENGINE implementations (except "openssl" and
4991 "openbsd") to dynamically instantiate their own error strings. This
4992 makes them more flexible to be built both as statically-linked ENGINEs
4993 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4994 Also, add stub code to each that makes building them as self-contained
4995 shared-libraries easier (see README.ENGINE).
4998 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4999 implementations into applications that are completely implemented in
5000 self-contained shared-libraries. The "dynamic" ENGINE exposes control
5001 commands that can be used to configure what shared-library to load and
5002 to control aspects of the way it is handled. Also, made an update to
5003 the README.ENGINE file that brings its information up-to-date and
5004 provides some information and instructions on the "dynamic" ENGINE
5005 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
5008 *) Make it possible to unload ranges of ERR strings with a new
5009 "ERR_unload_strings" function.
5012 *) Add a copy() function to EVP_MD.
5015 *) Make EVP_MD routines take a context pointer instead of just the
5016 md_data void pointer.
5019 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
5020 that the digest can only process a single chunk of data
5021 (typically because it is provided by a piece of
5022 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
5023 is only going to provide a single chunk of data, and hence the
5024 framework needn't accumulate the data for oneshot drivers.
5027 *) As with "ERR", make it possible to replace the underlying "ex_data"
5028 functions. This change also alters the storage and management of global
5029 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
5030 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
5031 index counters. The API functions that use this state have been changed
5032 to take a "class_index" rather than pointers to the class's local STACK
5033 and counter, and there is now an API function to dynamically create new
5034 classes. This centralisation allows us to (a) plug a lot of the
5035 thread-safety problems that existed, and (b) makes it possible to clean
5036 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
5037 such data would previously have always leaked in application code and
5038 workarounds were in place to make the memory debugging turn a blind eye
5039 to it. Application code that doesn't use this new function will still
5040 leak as before, but their memory debugging output will announce it now
5041 rather than letting it slide.
5043 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
5044 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
5045 has a return value to indicate success or failure.
5048 *) Make it possible to replace the underlying "ERR" functions such that the
5049 global state (2 LHASH tables and 2 locks) is only used by the "default"
5050 implementation. This change also adds two functions to "get" and "set"
5051 the implementation prior to it being automatically set the first time
5052 any other ERR function takes place. Ie. an application can call "get",
5053 pass the return value to a module it has just loaded, and that module
5054 can call its own "set" function using that value. This means the
5055 module's "ERR" operations will use (and modify) the error state in the
5056 application and not in its own statically linked copy of OpenSSL code.
5059 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
5060 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
5061 the operation, and provides a more encapsulated way for external code
5062 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
5063 to use these functions rather than manually incrementing the counts.
5065 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
5068 *) Add EVP test program.
5071 *) Add symmetric cipher support to ENGINE. Expect the API to change!
5074 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
5075 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
5076 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
5077 These allow a CRL to be built without having to access X509_CRL fields
5078 directly. Modify 'ca' application to use new functions.
5081 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
5082 bug workarounds. Rollback attack detection is a security feature.
5083 The problem will only arise on OpenSSL servers when TLSv1 is not
5084 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
5085 Software authors not wanting to support TLSv1 will have special reasons
5086 for their choice and can explicitly enable this option.
5087 [Bodo Moeller, Lutz Jaenicke]
5089 *) Rationalise EVP so it can be extended: don't include a union of
5090 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
5091 (similar to those existing for EVP_CIPHER_CTX).
5096 EVP_MD_CTX_init(&md); /* new function call */
5097 EVP_DigestInit(&md, EVP_sha1());
5098 EVP_DigestUpdate(&md, in, len);
5099 EVP_DigestFinal(&md, out, NULL);
5100 EVP_MD_CTX_cleanup(&md); /* new function call */
5104 *) Make DES key schedule conform to the usual scheme, as well as
5105 correcting its structure. This means that calls to DES functions
5106 now have to pass a pointer to a des_key_schedule instead of a
5107 plain des_key_schedule (which was actually always a pointer
5110 des_key_schedule ks;
5112 des_set_key_checked(..., &ks);
5113 des_ncbc_encrypt(..., &ks, ...);
5115 (Note that a later change renames 'des_...' into 'DES_...'.)
5118 *) Initial reduction of linker bloat: the use of some functions, such as
5119 PEM causes large amounts of unused functions to be linked in due to
5120 poor organisation. For example pem_all.c contains every PEM function
5121 which has a knock on effect of linking in large amounts of (unused)
5122 ASN1 code. Grouping together similar functions and splitting unrelated
5123 functions prevents this.
5126 *) Cleanup of EVP macros.
5129 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
5130 correct _ecb suffix.
5133 *) Add initial OCSP responder support to ocsp application. The
5134 revocation information is handled using the text based index
5135 use by the ca application. The responder can either handle
5136 requests generated internally, supplied in files (for example
5137 via a CGI script) or using an internal minimal server.
5140 *) Add configuration choices to get zlib compression for TLS.
5143 *) Changes to Kerberos SSL for RFC 2712 compliance:
5144 1. Implemented real KerberosWrapper, instead of just using
5145 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
5146 2. Implemented optional authenticator field of KerberosWrapper.
5148 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
5149 and authenticator structs; see crypto/krb5/.
5151 Generalized Kerberos calls to support multiple Kerberos libraries.
5152 [Vern Staats <staatsvr@asc.hpc.mil>,
5153 Jeffrey Altman <jaltman@columbia.edu>
5154 via Richard Levitte]
5156 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
5157 already does with RSA. testdsa.h now has 'priv_key/pub_key'
5158 values for each of the key sizes rather than having just
5159 parameters (and 'speed' generating keys each time).
5162 *) Speed up EVP routines.
5165 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
5166 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
5167 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
5168 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
5170 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
5171 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
5172 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
5175 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
5177 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
5180 *) Added the OS2-EMX target.
5181 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
5183 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
5184 to support NCONF routines in extension code. New function CONF_set_nconf()
5185 to allow functions which take an NCONF to also handle the old LHASH
5186 structure: this means that the old CONF compatible routines can be
5187 retained (in particular wrt extensions) without having to duplicate the
5188 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
5191 *) Enhance the general user interface with mechanisms for inner control
5192 and with possibilities to have yes/no kind of prompts.
5195 *) Change all calls to low level digest routines in the library and
5196 applications to use EVP. Add missing calls to HMAC_cleanup() and
5197 don't assume HMAC_CTX can be copied using memcpy().
5198 [Verdon Walker <VWalker@novell.com>, Steve Henson]
5200 *) Add the possibility to control engines through control names but with
5201 arbitrary arguments instead of just a string.
5202 Change the key loaders to take a UI_METHOD instead of a callback
5203 function pointer. NOTE: this breaks binary compatibility with earlier
5204 versions of OpenSSL [engine].
5205 Adapt the nCipher code for these new conditions and add a card insertion
5209 *) Enhance the general user interface with mechanisms to better support
5210 dialog box interfaces, application-defined prompts, the possibility
5211 to use defaults (for example default passwords from somewhere else)
5212 and interrupts/cancellations.
5215 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
5216 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
5219 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
5220 tidy up some unnecessarily weird code in 'sk_new()').
5221 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
5223 *) Change the key loading routines for ENGINEs to use the same kind
5224 callback (pem_password_cb) as all other routines that need this
5228 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
5229 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
5230 than this minimum value is recommended.
5233 *) New random seeder for OpenVMS, using the system process statistics
5234 that are easily reachable.
5237 *) Windows apparently can't transparently handle global
5238 variables defined in DLLs. Initialisations such as:
5240 const ASN1_ITEM *it = &ASN1_INTEGER_it;
5242 wont compile. This is used by the any applications that need to
5243 declare their own ASN1 modules. This was fixed by adding the option
5244 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
5245 needed for static libraries under Win32.
5248 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
5249 setting of purpose and trust fields. New X509_STORE trust and
5250 purpose functions and tidy up setting in other SSL functions.
5253 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
5254 structure. These are inherited by X509_STORE_CTX when it is
5255 initialised. This allows various defaults to be set in the
5256 X509_STORE structure (such as flags for CRL checking and custom
5257 purpose or trust settings) for functions which only use X509_STORE_CTX
5258 internally such as S/MIME.
5260 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
5261 trust settings if they are not set in X509_STORE. This allows X509_STORE
5262 purposes and trust (in S/MIME for example) to override any set by default.
5264 Add command line options for CRL checking to smime, s_client and s_server
5268 *) Initial CRL based revocation checking. If the CRL checking flag(s)
5269 are set then the CRL is looked up in the X509_STORE structure and
5270 its validity and signature checked, then if the certificate is found
5271 in the CRL the verify fails with a revoked error.
5273 Various new CRL related callbacks added to X509_STORE_CTX structure.
5275 Command line options added to 'verify' application to support this.
5277 This needs some additional work, such as being able to handle multiple
5278 CRLs with different times, extension based lookup (rather than just
5279 by subject name) and ultimately more complete V2 CRL extension
5283 *) Add a general user interface API (crypto/ui/). This is designed
5284 to replace things like des_read_password and friends (backward
5285 compatibility functions using this new API are provided).
5286 The purpose is to remove prompting functions from the DES code
5287 section as well as provide for prompting through dialog boxes in
5288 a window system and the like.
5291 *) Add "ex_data" support to ENGINE so implementations can add state at a
5292 per-structure level rather than having to store it globally.
5295 *) Make it possible for ENGINE structures to be copied when retrieved by
5296 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
5297 This causes the "original" ENGINE structure to act like a template,
5298 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
5299 operational state can be localised to each ENGINE structure, despite the
5300 fact they all share the same "methods". New ENGINE structures returned in
5301 this case have no functional references and the return value is the single
5302 structural reference. This matches the single structural reference returned
5303 by ENGINE_by_id() normally, when it is incremented on the pre-existing
5307 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
5308 needs to match any other type at all we need to manually clear the
5312 *) Changes to the "openssl engine" utility to include;
5313 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
5314 about an ENGINE's available control commands.
5315 - executing control commands from command line arguments using the
5316 '-pre' and '-post' switches. '-post' is only used if '-t' is
5317 specified and the ENGINE is successfully initialised. The syntax for
5318 the individual commands are colon-separated, for example;
5319 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
5322 *) New dynamic control command support for ENGINEs. ENGINEs can now
5323 declare their own commands (numbers), names (strings), descriptions,
5324 and input types for run-time discovery by calling applications. A
5325 subset of these commands are implicitly classed as "executable"
5326 depending on their input type, and only these can be invoked through
5327 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
5328 can be based on user input, config files, etc). The distinction is
5329 that "executable" commands cannot return anything other than a boolean
5330 result and can only support numeric or string input, whereas some
5331 discoverable commands may only be for direct use through
5332 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
5333 pointers, or other custom uses. The "executable" commands are to
5334 support parameterisations of ENGINE behaviour that can be
5335 unambiguously defined by ENGINEs and used consistently across any
5336 OpenSSL-based application. Commands have been added to all the
5337 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
5338 control over shared-library paths without source code alterations.
5341 *) Changed all ENGINE implementations to dynamically allocate their
5342 ENGINEs rather than declaring them statically. Apart from this being
5343 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
5344 this also allows the implementations to compile without using the
5345 internal engine_int.h header.
5348 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
5349 'const' value. Any code that should be able to modify a RAND_METHOD
5350 should already have non-const pointers to it (ie. they should only
5351 modify their own ones).
5354 *) Made a variety of little tweaks to the ENGINE code.
5355 - "atalla" and "ubsec" string definitions were moved from header files
5356 to C code. "nuron" string definitions were placed in variables
5357 rather than hard-coded - allowing parameterisation of these values
5358 later on via ctrl() commands.
5359 - Removed unused "#if 0"'d code.
5360 - Fixed engine list iteration code so it uses ENGINE_free() to release
5361 structural references.
5362 - Constified the RAND_METHOD element of ENGINE structures.
5363 - Constified various get/set functions as appropriate and added
5364 missing functions (including a catch-all ENGINE_cpy that duplicates
5365 all ENGINE values onto a new ENGINE except reference counts/state).
5366 - Removed NULL parameter checks in get/set functions. Setting a method
5367 or function to NULL is a way of cancelling out a previously set
5368 value. Passing a NULL ENGINE parameter is just plain stupid anyway
5369 and doesn't justify the extra error symbols and code.
5370 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
5371 flags from engine_int.h to engine.h.
5372 - Changed prototypes for ENGINE handler functions (init(), finish(),
5373 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
5376 *) Implement binary inversion algorithm for BN_mod_inverse in addition
5377 to the algorithm using long division. The binary algorithm can be
5378 used only if the modulus is odd. On 32-bit systems, it is faster
5379 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
5380 roughly 5-15% for 256-bit moduli), so we use it only for moduli
5381 up to 450 bits. In 64-bit environments, the binary algorithm
5382 appears to be advantageous for much longer moduli; here we use it
5383 for moduli up to 2048 bits.
5386 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
5387 could not support the combine flag in choice fields.
5390 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
5391 extensions from a certificate request to the certificate.
5394 *) Allow multiple 'certopt' and 'nameopt' options to be separated
5395 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
5396 file: this allows the display of the certificate about to be
5397 signed to be customised, to allow certain fields to be included
5398 or excluded and extension details. The old system didn't display
5399 multicharacter strings properly, omitted fields not in the policy
5400 and couldn't display additional details such as extensions.
5403 *) Function EC_POINTs_mul for multiple scalar multiplication
5404 of an arbitrary number of elliptic curve points
5405 \sum scalars[i]*points[i],
5406 optionally including the generator defined for the EC_GROUP:
5407 scalar*generator + \sum scalars[i]*points[i].
5409 EC_POINT_mul is a simple wrapper function for the typical case
5410 that the point list has just one item (besides the optional
5414 *) First EC_METHODs for curves over GF(p):
5416 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
5417 operations and provides various method functions that can also
5418 operate with faster implementations of modular arithmetic.
5420 EC_GFp_mont_method() reuses most functions that are part of
5421 EC_GFp_simple_method, but uses Montgomery arithmetic.
5423 [Bodo Moeller; point addition and point doubling
5424 implementation directly derived from source code provided by
5425 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
5427 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
5428 crypto/ec/ec_lib.c):
5430 Curves are EC_GROUP objects (with an optional group generator)
5431 based on EC_METHODs that are built into the library.
5433 Points are EC_POINT objects based on EC_GROUP objects.
5435 Most of the framework would be able to handle curves over arbitrary
5436 finite fields, but as there are no obvious types for fields other
5437 than GF(p), some functions are limited to that for now.
5440 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
5441 that the file contains a complete HTTP response.
5444 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
5445 change the def and num file printf format specifier from "%-40sXXX"
5446 to "%-39s XXX". The latter will always guarantee a space after the
5447 field while the former will cause them to run together if the field
5448 is 40 of more characters long.
5451 *) Constify the cipher and digest 'method' functions and structures
5452 and modify related functions to take constant EVP_MD and EVP_CIPHER
5456 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
5457 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
5460 *) Modify EVP_Digest*() routines so they now return values. Although the
5461 internal software routines can never fail additional hardware versions
5465 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
5467 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
5468 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
5471 ERR_R_NESTED_ASN1_ERROR
5473 ERR_R_MISSING_ASN1_EOS
5474 were 4 .. 9, conflicting with
5475 ERR_LIB_RSA (= ERR_R_RSA_LIB)
5477 ERR_LIB_PEM (= ERR_R_PEM_LIB).
5478 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
5480 Add new error code 'ERR_R_INTERNAL_ERROR'.
5483 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
5487 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
5488 sets the subject name for a new request or supersedes the
5489 subject name in a given request. Formats that can be parsed are
5490 'CN=Some Name, OU=myOU, C=IT'
5492 'CN=Some Name/OU=myOU/C=IT'.
5494 Add options '-batch' and '-verbose' to 'openssl req'.
5495 [Massimiliano Pala <madwolf@hackmasters.net>]
5497 *) Introduce the possibility to access global variables through
5498 functions on platform were that's the best way to handle exporting
5499 global variables in shared libraries. To enable this functionality,
5500 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
5501 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
5502 is normally done by Configure or something similar).
5504 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
5505 in the source file (foo.c) like this:
5507 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
5508 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
5510 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
5511 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
5513 OPENSSL_DECLARE_GLOBAL(int,foo);
5514 #define foo OPENSSL_GLOBAL_REF(foo)
5515 OPENSSL_DECLARE_GLOBAL(double,bar);
5516 #define bar OPENSSL_GLOBAL_REF(bar)
5518 The #defines are very important, and therefore so is including the
5519 header file everywhere where the defined globals are used.
5521 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
5522 of ASN.1 items, but that structure is a bit different.
5524 The largest change is in util/mkdef.pl which has been enhanced with
5525 better and easier to understand logic to choose which symbols should
5526 go into the Windows .def files as well as a number of fixes and code
5527 cleanup (among others, algorithm keywords are now sorted
5528 lexicographically to avoid constant rewrites).
5531 *) In BN_div() keep a copy of the sign of 'num' before writing the
5532 result to 'rm' because if rm==num the value will be overwritten
5533 and produce the wrong result if 'num' is negative: this caused
5534 problems with BN_mod() and BN_nnmod().
5537 *) Function OCSP_request_verify(). This checks the signature on an
5538 OCSP request and verifies the signer certificate. The signer
5539 certificate is just checked for a generic purpose and OCSP request
5543 *) Add OCSP_check_validity() function to check the validity of OCSP
5544 responses. OCSP responses are prepared in real time and may only
5545 be a few seconds old. Simply checking that the current time lies
5546 between thisUpdate and nextUpdate max reject otherwise valid responses
5547 caused by either OCSP responder or client clock inaccuracy. Instead
5548 we allow thisUpdate and nextUpdate to fall within a certain period of
5549 the current time. The age of the response can also optionally be
5550 checked. Two new options -validity_period and -status_age added to
5554 *) If signature or public key algorithm is unrecognized print out its
5555 OID rather that just UNKNOWN.
5558 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
5559 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
5560 ID to be generated from the issuer certificate alone which can then be
5561 passed to OCSP_id_issuer_cmp().
5564 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
5565 ASN1 modules to export functions returning ASN1_ITEM pointers
5566 instead of the ASN1_ITEM structures themselves. This adds several
5567 new macros which allow the underlying ASN1 function/structure to
5568 be accessed transparently. As a result code should not use ASN1_ITEM
5569 references directly (such as &X509_it) but instead use the relevant
5570 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
5571 use of the new ASN1 code on platforms where exporting structures
5572 is problematical (for example in shared libraries) but exporting
5573 functions returning pointers to structures is not.
5576 *) Add support for overriding the generation of SSL/TLS session IDs.
5577 These callbacks can be registered either in an SSL_CTX or per SSL.
5578 The purpose of this is to allow applications to control, if they wish,
5579 the arbitrary values chosen for use as session IDs, particularly as it
5580 can be useful for session caching in multiple-server environments. A
5581 command-line switch for testing this (and any client code that wishes
5582 to use such a feature) has been added to "s_server".
5583 [Geoff Thorpe, Lutz Jaenicke]
5585 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
5586 of the form '#if defined(...) || defined(...) || ...' and
5587 '#if !defined(...) && !defined(...) && ...'. This also avoids
5588 the growing number of special cases it was previously handling.
5591 *) Make all configuration macros available for application by making
5592 sure they are available in opensslconf.h, by giving them names starting
5593 with "OPENSSL_" to avoid conflicts with other packages and by making
5594 sure e_os2.h will cover all platform-specific cases together with
5596 Additionally, it is now possible to define configuration/platform-
5597 specific names (called "system identities"). In the C code, these
5598 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
5599 macro with the name beginning with "OPENSSL_SYS_", which is determined
5600 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
5604 *) New option -set_serial to 'req' and 'x509' this allows the serial
5605 number to use to be specified on the command line. Previously self
5606 signed certificates were hard coded with serial number 0 and the
5607 CA options of 'x509' had to use a serial number in a file which was
5611 *) New options to 'ca' utility to support V2 CRL entry extensions.
5612 Currently CRL reason, invalidity date and hold instruction are
5613 supported. Add new CRL extensions to V3 code and some new objects.
5616 *) New function EVP_CIPHER_CTX_set_padding() this is used to
5617 disable standard block padding (aka PKCS#5 padding) in the EVP
5618 API, which was previously mandatory. This means that the data is
5619 not padded in any way and so the total length much be a multiple
5620 of the block size, otherwise an error occurs.
5623 *) Initial (incomplete) OCSP SSL support.
5626 *) New function OCSP_parse_url(). This splits up a URL into its host,
5627 port and path components: primarily to parse OCSP URLs. New -url
5628 option to ocsp utility.
5631 *) New nonce behavior. The return value of OCSP_check_nonce() now
5632 reflects the various checks performed. Applications can decide
5633 whether to tolerate certain situations such as an absent nonce
5634 in a response when one was present in a request: the ocsp application
5635 just prints out a warning. New function OCSP_add1_basic_nonce()
5636 this is to allow responders to include a nonce in a response even if
5637 the request is nonce-less.
5640 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
5641 skipped when using openssl x509 multiple times on a single input file,
5642 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
5645 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
5646 set string type: to handle setting ASN1_TIME structures. Fix ca
5647 utility to correctly initialize revocation date of CRLs.
5650 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
5651 the clients preferred ciphersuites and rather use its own preferences.
5652 Should help to work around M$ SGC (Server Gated Cryptography) bug in
5653 Internet Explorer by ensuring unchanged hash method during stepup.
5654 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
5657 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
5658 to aes and add a new 'exist' option to print out symbols that don't
5662 *) Additional options to ocsp utility to allow flags to be set and
5663 additional certificates supplied.
5666 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
5667 OCSP client a number of certificate to only verify the response
5671 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
5672 handle the new API. Currently only ECB, CBC modes supported. Add new
5675 Add TLS AES ciphersuites as described in RFC3268, "Advanced
5676 Encryption Standard (AES) Ciphersuites for Transport Layer
5677 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
5678 not enabled by default and were not part of the "ALL" ciphersuite
5679 alias because they were not yet official; they could be
5680 explicitly requested by specifying the "AESdraft" ciphersuite
5681 group alias. In the final release of OpenSSL 0.9.7, the group
5682 alias is called "AES" and is part of "ALL".)
5683 [Ben Laurie, Steve Henson, Bodo Moeller]
5685 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
5686 request to response.
5689 *) Functions for OCSP responders. OCSP_request_onereq_count(),
5690 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
5691 extract information from a certificate request. OCSP_response_create()
5692 creates a response and optionally adds a basic response structure.
5693 OCSP_basic_add1_status() adds a complete single response to a basic
5694 response and returns the OCSP_SINGLERESP structure just added (to allow
5695 extensions to be included for example). OCSP_basic_add1_cert() adds a
5696 certificate to a basic response and OCSP_basic_sign() signs a basic
5697 response with various flags. New helper functions ASN1_TIME_check()
5698 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
5699 (converts ASN1_TIME to GeneralizedTime).
5702 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
5703 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
5704 structure from a certificate. X509_pubkey_digest() digests the public_key
5705 contents: this is used in various key identifiers.
5708 *) Make sk_sort() tolerate a NULL argument.
5709 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
5711 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
5712 passed by the function are trusted implicitly. If any of them signed the
5713 response then it is assumed to be valid and is not verified.
5716 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
5717 to data. This was previously part of the PKCS7 ASN1 code. This
5718 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
5719 [Steve Henson, reported by Kenneth R. Robinette
5720 <support@securenetterm.com>]
5722 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
5723 routines: without these tracing memory leaks is very painful.
5724 Fix leaks in PKCS12 and PKCS7 routines.
5727 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
5728 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
5729 effectively meant GeneralizedTime would never be used. Now it
5730 is initialised to -1 but X509_time_adj() now has to check the value
5731 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
5732 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
5733 [Steve Henson, reported by Kenneth R. Robinette
5734 <support@securenetterm.com>]
5736 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
5737 result in a zero length in the ASN1_INTEGER structure which was
5738 not consistent with the structure when d2i_ASN1_INTEGER() was used
5739 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
5740 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
5741 where it did not print out a minus for negative ASN1_INTEGER.
5744 *) Add summary printout to ocsp utility. The various functions which
5745 convert status values to strings have been renamed to:
5746 OCSP_response_status_str(), OCSP_cert_status_str() and
5747 OCSP_crl_reason_str() and are no longer static. New options
5748 to verify nonce values and to disable verification. OCSP response
5749 printout format cleaned up.
5752 *) Add additional OCSP certificate checks. These are those specified
5753 in RFC2560. This consists of two separate checks: the CA of the
5754 certificate being checked must either be the OCSP signer certificate
5755 or the issuer of the OCSP signer certificate. In the latter case the
5756 OCSP signer certificate must contain the OCSP signing extended key
5757 usage. This check is performed by attempting to match the OCSP
5758 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
5759 in the OCSP_CERTID structures of the response.
5762 *) Initial OCSP certificate verification added to OCSP_basic_verify()
5763 and related routines. This uses the standard OpenSSL certificate
5764 verify routines to perform initial checks (just CA validity) and
5765 to obtain the certificate chain. Then additional checks will be
5766 performed on the chain. Currently the root CA is checked to see
5767 if it is explicitly trusted for OCSP signing. This is used to set
5768 a root CA as a global signing root: that is any certificate that
5769 chains to that CA is an acceptable OCSP signing certificate.
5772 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
5773 extensions from a separate configuration file.
5774 As when reading extensions from the main configuration file,
5775 the '-extensions ...' option may be used for specifying the
5777 [Massimiliano Pala <madwolf@comune.modena.it>]
5779 *) New OCSP utility. Allows OCSP requests to be generated or
5780 read. The request can be sent to a responder and the output
5781 parsed, outputed or printed in text form. Not complete yet:
5782 still needs to check the OCSP response validity.
5785 *) New subcommands for 'openssl ca':
5786 'openssl ca -status <serial>' prints the status of the cert with
5787 the given serial number (according to the index file).
5788 'openssl ca -updatedb' updates the expiry status of certificates
5790 [Massimiliano Pala <madwolf@comune.modena.it>]
5792 *) New '-newreq-nodes' command option to CA.pl. This is like
5793 '-newreq', but calls 'openssl req' with the '-nodes' option
5794 so that the resulting key is not encrypted.
5795 [Damien Miller <djm@mindrot.org>]
5797 *) New configuration for the GNU Hurd.
5798 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5800 *) Initial code to implement OCSP basic response verify. This
5801 is currently incomplete. Currently just finds the signer's
5802 certificate and verifies the signature on the response.
5805 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5806 value of OPENSSLDIR. This is available via the new '-d' option
5807 to 'openssl version', and is also included in 'openssl version -a'.
5810 *) Allowing defining memory allocation callbacks that will be given
5811 file name and line number information in additional arguments
5812 (a const char* and an int). The basic functionality remains, as
5813 well as the original possibility to just replace malloc(),
5814 realloc() and free() by functions that do not know about these
5815 additional arguments. To register and find out the current
5816 settings for extended allocation functions, the following
5817 functions are provided:
5819 CRYPTO_set_mem_ex_functions
5820 CRYPTO_set_locked_mem_ex_functions
5821 CRYPTO_get_mem_ex_functions
5822 CRYPTO_get_locked_mem_ex_functions
5824 These work the same way as CRYPTO_set_mem_functions and friends.
5825 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5826 extended allocation function is enabled.
5827 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5828 a conventional allocation function is enabled.
5829 [Richard Levitte, Bodo Moeller]
5831 *) Finish off removing the remaining LHASH function pointer casts.
5832 There should no longer be any prototype-casting required when using
5833 the LHASH abstraction, and any casts that remain are "bugs". See
5834 the callback types and macros at the head of lhash.h for details
5835 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5838 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5839 If /dev/[u]random devices are not available or do not return enough
5840 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5842 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5843 /etc/entropy will be queried once each in this sequence, quering stops
5844 when enough entropy was collected without querying more sockets.
5847 *) Change the Unix RAND_poll() variant to be able to poll several
5848 random devices, as specified by DEVRANDOM, until a sufficient amount
5849 of data has been collected. We spend at most 10 ms on each file
5850 (select timeout) and read in non-blocking mode. DEVRANDOM now
5851 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5852 (previously it was just the string "/dev/urandom"), so on typical
5853 platforms the 10 ms delay will never occur.
5854 Also separate out the Unix variant to its own file, rand_unix.c.
5855 For VMS, there's a currently-empty rand_vms.c.
5858 *) Move OCSP client related routines to ocsp_cl.c. These
5859 provide utility functions which an application needing
5860 to issue a request to an OCSP responder and analyse the
5861 response will typically need: as opposed to those which an
5862 OCSP responder itself would need which will be added later.
5864 OCSP_request_sign() signs an OCSP request with an API similar
5865 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5866 response. OCSP_response_get1_basic() extracts basic response
5867 from response. OCSP_resp_find_status(): finds and extracts status
5868 information from an OCSP_CERTID structure (which will be created
5869 when the request structure is built). These are built from lower
5870 level functions which work on OCSP_SINGLERESP structures but
5871 wont normally be used unless the application wishes to examine
5872 extensions in the OCSP response for example.
5874 Replace nonce routines with a pair of functions.
5875 OCSP_request_add1_nonce() adds a nonce value and optionally
5876 generates a random value. OCSP_check_nonce() checks the
5877 validity of the nonce in an OCSP response.
5880 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5881 This doesn't copy the supplied OCSP_CERTID and avoids the
5882 need to free up the newly created id. Change return type
5883 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5884 This can then be used to add extensions to the request.
5885 Deleted OCSP_request_new(), since most of its functionality
5886 is now in OCSP_REQUEST_new() (and the case insensitive name
5887 clash) apart from the ability to set the request name which
5888 will be added elsewhere.
5891 *) Update OCSP API. Remove obsolete extensions argument from
5892 various functions. Extensions are now handled using the new
5893 OCSP extension code. New simple OCSP HTTP function which
5894 can be used to send requests and parse the response.
5897 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5898 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5899 uses the special reorder version of SET OF to sort the attributes
5900 and reorder them to match the encoded order. This resolves a long
5901 standing problem: a verify on a PKCS7 structure just after signing
5902 it used to fail because the attribute order did not match the
5903 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5904 it uses the received order. This is necessary to tolerate some broken
5905 software that does not order SET OF. This is handled by encoding
5906 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5907 to produce the required SET OF.
5910 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5911 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5912 files to get correct declarations of the ASN.1 item variables.
5915 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5916 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5917 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5918 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5919 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5920 ASN1_ITEM and no wrapper functions.
5923 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5924 replace the old function pointer based I/O routines. Change most of
5925 the *_d2i_bio() and *_d2i_fp() functions to use these.
5928 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5929 lines, recognice more "algorithms" that can be deselected, and make
5930 it complain about algorithm deselection that isn't recognised.
5933 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5934 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5935 to use new functions. Add NO_ASN1_OLD which can be set to remove
5936 some old style ASN1 functions: this can be used to determine if old
5937 code will still work when these eventually go away.
5940 *) New extension functions for OCSP structures, these follow the
5941 same conventions as certificates and CRLs.
5944 *) New function X509V3_add1_i2d(). This automatically encodes and
5945 adds an extension. Its behaviour can be customised with various
5946 flags to append, replace or delete. Various wrappers added for
5947 certifcates and CRLs.
5950 *) Fix to avoid calling the underlying ASN1 print routine when
5951 an extension cannot be parsed. Correct a typo in the
5952 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5955 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5956 entries for variables.
5959 *) Add functionality to apps/openssl.c for detecting locking
5960 problems: As the program is single-threaded, all we have
5961 to do is register a locking callback using an array for
5962 storing which locks are currently held by the program.
5965 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5966 SSL_get_ex_data_X509_STORE_idx(), which is used in
5967 ssl_verify_cert_chain() and thus can be called at any time
5968 during TLS/SSL handshakes so that thread-safety is essential.
5969 Unfortunately, the ex_data design is not at all suited
5970 for multi-threaded use, so it probably should be abolished.
5973 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5974 [Broadcom, tweaked and integrated by Geoff Thorpe]
5976 *) Move common extension printing code to new function
5977 X509V3_print_extensions(). Reorganise OCSP print routines and
5978 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5981 *) New function X509_signature_print() to remove duplication in some
5985 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5986 set (this was treated exactly the same as SET OF previously). This
5987 is used to reorder the STACK representing the structure to match the
5988 encoding. This will be used to get round a problem where a PKCS7
5989 structure which was signed could not be verified because the STACK
5990 order did not reflect the encoded order.
5993 *) Reimplement the OCSP ASN1 module using the new code.
5996 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5997 for its ASN1 operations. The old style function pointers still exist
5998 for now but they will eventually go away.
6001 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
6002 completely replaces the old ASN1 functionality with a table driven
6003 encoder and decoder which interprets an ASN1_ITEM structure describing
6004 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
6005 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
6006 has also been converted to the new form.
6009 *) Change BN_mod_exp_recp so that negative moduli are tolerated
6010 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
6011 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
6012 for negative moduli.
6015 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
6016 of not touching the result's sign bit.
6019 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
6023 *) Changed the LHASH code to use prototypes for callbacks, and created
6024 macros to declare and implement thin (optionally static) functions
6025 that provide type-safety and avoid function pointer casting for the
6026 type-specific callbacks.
6029 *) Added Kerberos Cipher Suites to be used with TLS, as written in
6031 [Veers Staats <staatsvr@asc.hpc.mil>,
6032 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
6034 *) Reformat the FAQ so the different questions and answers can be divided
6035 in sections depending on the subject.
6038 *) Have the zlib compression code load ZLIB.DLL dynamically under
6042 *) New function BN_mod_sqrt for computing square roots modulo a prime
6043 (using the probabilistic Tonelli-Shanks algorithm unless
6044 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
6045 be handled deterministically).
6046 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6048 *) Make BN_mod_inverse faster by explicitly handling small quotients
6049 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
6050 512 bits], about 30% for larger ones [1024 or 2048 bits].)
6053 *) New function BN_kronecker.
6056 *) Fix BN_gcd so that it works on negative inputs; the result is
6057 positive unless both parameters are zero.
6058 Previously something reasonably close to an infinite loop was
6059 possible because numbers could be growing instead of shrinking
6060 in the implementation of Euclid's algorithm.
6063 *) Fix BN_is_word() and BN_is_one() macros to take into account the
6064 sign of the number in question.
6066 Fix BN_is_word(a,w) to work correctly for w == 0.
6068 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
6069 because its test if the absolute value of 'a' equals 'w'.
6070 Note that BN_abs_is_word does *not* handle w == 0 reliably;
6071 it exists mostly for use in the implementations of BN_is_zero(),
6072 BN_is_one(), and BN_is_word().
6075 *) New function BN_swap.
6078 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
6079 the exponentiation functions are more likely to produce reasonable
6080 results on negative inputs.
6083 *) Change BN_mod_mul so that the result is always non-negative.
6084 Previously, it could be negative if one of the factors was negative;
6085 I don't think anyone really wanted that behaviour.
6088 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
6089 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
6090 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
6091 and add new functions:
6100 BN_mod_lshift1_quick
6104 These functions always generate non-negative results.
6106 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
6107 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
6109 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
6110 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
6111 be reduced modulo m.
6112 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6115 The following entry accidentily appeared in the CHANGES file
6116 distributed with OpenSSL 0.9.7. The modifications described in
6117 it do *not* apply to OpenSSL 0.9.7.
6119 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6120 was actually never needed) and in BN_mul(). The removal in BN_mul()
6121 required a small change in bn_mul_part_recursive() and the addition
6122 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6123 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6124 bn_sub_words() and bn_add_words() except they take arrays with
6129 *) In 'openssl passwd', verify passwords read from the terminal
6130 unless the '-salt' option is used (which usually means that
6131 verification would just waste user's time since the resulting
6132 hash is going to be compared with some given password hash)
6133 or the new '-noverify' option is used.
6135 This is an incompatible change, but it does not affect
6136 non-interactive use of 'openssl passwd' (passwords on the command
6137 line, '-stdin' option, '-in ...' option) and thus should not
6141 *) Remove all references to RSAref, since there's no more need for it.
6144 *) Make DSO load along a path given through an environment variable
6145 (SHLIB_PATH) with shl_load().
6148 *) Constify the ENGINE code as a result of BIGNUM constification.
6149 Also constify the RSA code and most things related to it. In a
6150 few places, most notable in the depth of the ASN.1 code, ugly
6151 casts back to non-const were required (to be solved at a later
6155 *) Make it so the openssl application has all engines loaded by default.
6158 *) Constify the BIGNUM routines a little more.
6161 *) Add the following functions:
6163 ENGINE_load_cswift()
6165 ENGINE_load_atalla()
6167 ENGINE_load_builtin_engines()
6169 That way, an application can itself choose if external engines that
6170 are built-in in OpenSSL shall ever be used or not. The benefit is
6171 that applications won't have to be linked with libdl or other dso
6172 libraries unless it's really needed.
6174 Changed 'openssl engine' to load all engines on demand.
6175 Changed the engine header files to avoid the duplication of some
6176 declarations (they differed!).
6179 *) 'openssl engine' can now list capabilities.
6182 *) Better error reporting in 'openssl engine'.
6185 *) Never call load_dh_param(NULL) in s_server.
6188 *) Add engine application. It can currently list engines by name and
6189 identity, and test if they are actually available.
6192 *) Improve RPM specification file by forcing symbolic linking and making
6193 sure the installed documentation is also owned by root.root.
6194 [Damien Miller <djm@mindrot.org>]
6196 *) Give the OpenSSL applications more possibilities to make use of
6197 keys (public as well as private) handled by engines.
6200 *) Add OCSP code that comes from CertCo.
6203 *) Add VMS support for the Rijndael code.
6206 *) Added untested support for Nuron crypto accelerator.
6209 *) Add support for external cryptographic devices. This code was
6210 previously distributed separately as the "engine" branch.
6211 [Geoff Thorpe, Richard Levitte]
6213 *) Rework the filename-translation in the DSO code. It is now possible to
6214 have far greater control over how a "name" is turned into a filename
6215 depending on the operating environment and any oddities about the
6216 different shared library filenames on each system.
6219 *) Support threads on FreeBSD-elf in Configure.
6222 *) Fix for SHA1 assembly problem with MASM: it produces
6223 warnings about corrupt line number information when assembling
6224 with debugging information. This is caused by the overlapping
6226 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
6229 NCONF_get_number() has no error checking at all. As a replacement,
6230 NCONF_get_number_e() is defined (_e for "error checking") and is
6231 promoted strongly. The old NCONF_get_number is kept around for
6232 binary backward compatibility.
6233 Make it possible for methods to load from something other than a BIO,
6234 by providing a function pointer that is given a name instead of a BIO.
6235 For example, this could be used to load configuration data from an
6239 *) Fix for non blocking accept BIOs. Added new I/O special reason
6240 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
6241 with non blocking I/O was not possible because no retry code was
6242 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
6246 *) Added the beginnings of Rijndael support.
6249 *) Fix for bug in DirectoryString mask setting. Add support for
6250 X509_NAME_print_ex() in 'req' and X509_print_ex() function
6251 to allow certificate printing to more controllable, additional
6252 'certopt' option to 'x509' to allow new printing options to be
6256 *) Clean old EAY MD5 hack from e_os.h.
6259 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
6261 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6262 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6263 [Joe Orton, Steve Henson]
6265 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
6267 *) Fix additional bug revealed by the NISCC test suite:
6269 Stop bug triggering large recursion when presented with
6270 certain ASN.1 tags (CVE-2003-0851)
6273 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
6275 *) Fix various bugs revealed by running the NISCC test suite:
6277 Stop out of bounds reads in the ASN1 code when presented with
6278 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6280 If verify callback ignores invalid public key errors don't try to check
6281 certificate signature with the NULL public key.
6285 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6286 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6290 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6291 extra data after the compression methods not only for TLS 1.0
6292 but also for SSL 3.0 (as required by the specification).
6293 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6295 *) Change X509_certificate_type() to mark the key as exported/exportable
6296 when it's 512 *bits* long, not 512 bytes.
6299 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
6301 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6302 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6303 a protocol version number mismatch like a decryption error
6304 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6307 *) Turn on RSA blinding by default in the default implementation
6308 to avoid a timing attack. Applications that don't want it can call
6309 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6310 They would be ill-advised to do so in most cases.
6311 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6313 *) Change RSA blinding code so that it works when the PRNG is not
6314 seeded (in this case, the secret RSA exponent is abused as
6315 an unpredictable seed -- if it is not unpredictable, there
6316 is no point in blinding anyway). Make RSA blinding thread-safe
6317 by remembering the creator's thread ID in rsa->blinding and
6318 having all other threads use local one-time blinding factors
6319 (this requires more computation than sharing rsa->blinding, but
6320 avoids excessive locking; and if an RSA object is not shared
6321 between threads, blinding will still be very fast).
6324 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
6326 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6327 via timing by performing a MAC computation even if incorrrect
6328 block cipher padding has been found. This is a countermeasure
6329 against active attacks where the attacker has to distinguish
6330 between bad padding and a MAC verification error. (CVE-2003-0078)
6332 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6333 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6334 Martin Vuagnoux (EPFL, Ilion)]
6336 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
6338 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
6339 memory from it's contents. This is done with a counter that will
6340 place alternating values in each byte. This can be used to solve
6341 two issues: 1) the removal of calls to memset() by highly optimizing
6342 compilers, and 2) cleansing with other values than 0, since those can
6343 be read through on certain media, for example a swap space on disk.
6346 *) Bugfix: client side session caching did not work with external caching,
6347 because the session->cipher setting was not restored when reloading
6348 from the external cache. This problem was masked, when
6349 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
6350 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
6353 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
6354 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
6355 [Zeev Lieber <zeev-l@yahoo.com>]
6357 *) Undo an undocumented change introduced in 0.9.6e which caused
6358 repeated calls to OpenSSL_add_all_ciphers() and
6359 OpenSSL_add_all_digests() to be ignored, even after calling
6363 *) Change the default configuration reader to deal with last line not
6364 being properly terminated.
6367 *) Change X509_NAME_cmp() so it applies the special rules on handling
6368 DN values that are of type PrintableString, as well as RDNs of type
6369 emailAddress where the value has the type ia5String.
6370 [stefank@valicert.com via Richard Levitte]
6372 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
6373 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
6374 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
6375 the bitwise-OR of the two for use by the majority of applications
6376 wanting this behaviour, and update the docs. The documented
6377 behaviour and actual behaviour were inconsistent and had been
6378 changing anyway, so this is more a bug-fix than a behavioural
6380 [Geoff Thorpe, diagnosed by Nadav Har'El]
6382 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
6383 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
6386 *) Fix initialization code race conditions in
6387 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
6388 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
6389 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
6390 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
6391 ssl2_get_cipher_by_char(),
6392 ssl3_get_cipher_by_char().
6393 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
6395 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
6396 the cached sessions are flushed, as the remove_cb() might use ex_data
6397 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
6398 (see [openssl.org #212]).
6399 [Geoff Thorpe, Lutz Jaenicke]
6401 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
6402 length, instead of the encoding length to d2i_ASN1_OBJECT.
6405 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
6407 *) [In 0.9.6g-engine release:]
6408 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
6409 [Lynn Gazis <lgazis@rainbow.com>]
6411 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
6413 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
6414 and get fix the header length calculation.
6415 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
6416 Alon Kantor <alonk@checkpoint.com> (and others),
6419 *) Use proper error handling instead of 'assertions' in buffer
6420 overflow checks added in 0.9.6e. This prevents DoS (the
6421 assertions could call abort()).
6422 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
6424 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
6426 *) Add various sanity checks to asn1_get_length() to reject
6427 the ASN1 length bytes if they exceed sizeof(long), will appear
6428 negative or the content length exceeds the length of the
6430 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6432 *) Fix cipher selection routines: ciphers without encryption had no flags
6433 for the cipher strength set and where therefore not handled correctly
6434 by the selection routines (PR #130).
6437 *) Fix EVP_dsa_sha macro.
6441 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
6442 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
6443 that was added in OpenSSL 0.9.6d.
6445 As the countermeasure turned out to be incompatible with some
6446 broken SSL implementations, the new option is part of SSL_OP_ALL.
6447 SSL_OP_ALL is usually employed when compatibility with weird SSL
6448 implementations is desired (e.g. '-bugs' option to 's_client' and
6449 's_server'), so the new option is automatically set in many
6453 *) Changes in security patch:
6455 Changes marked "(CHATS)" were sponsored by the Defense Advanced
6456 Research Projects Agency (DARPA) and Air Force Research Laboratory,
6457 Air Force Materiel Command, USAF, under agreement number
6460 *) Add various sanity checks to asn1_get_length() to reject
6461 the ASN1 length bytes if they exceed sizeof(long), will appear
6462 negative or the content length exceeds the length of the
6463 supplied buffer. (CVE-2002-0659)
6464 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6466 *) Assertions for various potential buffer overflows, not known to
6468 [Ben Laurie (CHATS)]
6470 *) Various temporary buffers to hold ASCII versions of integers were
6471 too small for 64 bit platforms. (CVE-2002-0655)
6472 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
6474 *) Remote buffer overflow in SSL3 protocol - an attacker could
6475 supply an oversized session ID to a client. (CVE-2002-0656)
6476 [Ben Laurie (CHATS)]
6478 *) Remote buffer overflow in SSL2 protocol - an attacker could
6479 supply an oversized client master key. (CVE-2002-0656)
6480 [Ben Laurie (CHATS)]
6482 Changes between 0.9.6c and 0.9.6d [9 May 2002]
6484 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
6485 encoded as NULL) with id-dsa-with-sha1.
6486 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
6488 *) Check various X509_...() return values in apps/req.c.
6489 [Nils Larsch <nla@trustcenter.de>]
6491 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
6492 an end-of-file condition would erronously be flagged, when the CRLF
6493 was just at the end of a processed block. The bug was discovered when
6494 processing data through a buffering memory BIO handing the data to a
6495 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
6496 <ptsekov@syntrex.com> and Nedelcho Stanev.
6499 *) Implement a countermeasure against a vulnerability recently found
6500 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
6501 before application data chunks to avoid the use of known IVs
6502 with data potentially chosen by the attacker.
6505 *) Fix length checks in ssl3_get_client_hello().
6508 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
6509 to prevent ssl3_read_internal() from incorrectly assuming that
6510 ssl3_read_bytes() found application data while handshake
6511 processing was enabled when in fact s->s3->in_read_app_data was
6512 merely automatically cleared during the initial handshake.
6513 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
6515 *) Fix object definitions for Private and Enterprise: they were not
6516 recognized in their shortname (=lowercase) representation. Extend
6517 obj_dat.pl to issue an error when using undefined keywords instead
6518 of silently ignoring the problem (Svenning Sorensen
6519 <sss@sss.dnsalias.net>).
6522 *) Fix DH_generate_parameters() so that it works for 'non-standard'
6523 generators, i.e. generators other than 2 and 5. (Previously, the
6524 code did not properly initialise the 'add' and 'rem' values to
6525 BN_generate_prime().)
6527 In the new general case, we do not insist that 'generator' is
6528 actually a primitive root: This requirement is rather pointless;
6529 a generator of the order-q subgroup is just as good, if not
6533 *) Map new X509 verification errors to alerts. Discovered and submitted by
6534 Tom Wu <tom@arcot.com>.
6537 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
6538 returning non-zero before the data has been completely received
6539 when using non-blocking I/O.
6540 [Bodo Moeller; problem pointed out by John Hughes]
6542 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
6543 [Ben Laurie, Lutz Jaenicke]
6545 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
6546 Yoram Zahavi <YoramZ@gilian.com>).
6549 *) Add information about CygWin 1.3 and on, and preserve proper
6550 configuration for the versions before that.
6551 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6553 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
6554 check whether we deal with a copy of a session and do not delete from
6555 the cache in this case. Problem reported by "Izhar Shoshani Levi"
6556 <izhar@checkpoint.com>.
6559 *) Do not store session data into the internal session cache, if it
6560 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
6561 flag is set). Proposed by Aslam <aslam@funk.com>.
6564 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
6568 *) [In 0.9.6d-engine release:]
6569 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6570 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6572 *) Add the configuration target linux-s390x.
6573 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
6575 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
6576 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
6577 variable as an indication that a ClientHello message has been
6578 received. As the flag value will be lost between multiple
6579 invocations of ssl3_accept when using non-blocking I/O, the
6580 function may not be aware that a handshake has actually taken
6581 place, thus preventing a new session from being added to the
6584 To avoid this problem, we now set s->new_session to 2 instead of
6585 using a local variable.
6586 [Lutz Jaenicke, Bodo Moeller]
6588 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
6589 if the SSL_R_LENGTH_MISMATCH error is detected.
6590 [Geoff Thorpe, Bodo Moeller]
6592 *) New 'shared_ldflag' column in Configure platform table.
6595 *) Fix EVP_CIPHER_mode macro.
6596 ["Dan S. Camper" <dan@bti.net>]
6598 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
6599 type, we must throw them away by setting rr->length to 0.
6600 [D P Chang <dpc@qualys.com>]
6602 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
6604 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
6605 <Dominikus.Scherkl@biodata.com>. (The previous implementation
6606 worked incorrectly for those cases where range = 10..._2 and
6607 3*range is two bits longer than range.)
6610 *) Only add signing time to PKCS7 structures if it is not already
6614 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
6615 OBJ_ld_ce should be OBJ_id_ce.
6616 Also some ip-pda OIDs in crypto/objects/objects.txt were
6617 incorrect (cf. RFC 3039).
6618 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
6620 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
6621 returns early because it has nothing to do.
6622 [Andy Schneider <andy.schneider@bjss.co.uk>]
6624 *) [In 0.9.6c-engine release:]
6625 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
6626 [Andy Schneider <andy.schneider@bjss.co.uk>]
6628 *) [In 0.9.6c-engine release:]
6629 Add support for Cryptographic Appliance's keyserver technology.
6630 (Use engine 'keyclient')
6631 [Cryptographic Appliances and Geoff Thorpe]
6633 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
6634 is called via tools/c89.sh because arguments have to be
6635 rearranged (all '-L' options must appear before the first object
6637 [Richard Shapiro <rshapiro@abinitio.com>]
6639 *) [In 0.9.6c-engine release:]
6640 Add support for Broadcom crypto accelerator cards, backported
6642 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
6644 *) [In 0.9.6c-engine release:]
6645 Add support for SureWare crypto accelerator cards from
6646 Baltimore Technologies. (Use engine 'sureware')
6647 [Baltimore Technologies and Mark Cox]
6649 *) [In 0.9.6c-engine release:]
6650 Add support for crypto accelerator cards from Accelerated
6651 Encryption Processing, www.aep.ie. (Use engine 'aep')
6652 [AEP Inc. and Mark Cox]
6654 *) Add a configuration entry for gcc on UnixWare.
6655 [Gary Benson <gbenson@redhat.com>]
6657 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
6658 messages are stored in a single piece (fixed-length part and
6659 variable-length part combined) and fix various bugs found on the way.
6662 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
6663 instead. BIO_gethostbyname() does not know what timeouts are
6664 appropriate, so entries would stay in cache even when they have
6666 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
6668 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
6669 faced with a pathologically small ClientHello fragment that does
6670 not contain client_version: Instead of aborting with an error,
6671 simply choose the highest available protocol version (i.e.,
6672 TLS 1.0 unless it is disabled). In practice, ClientHello
6673 messages are never sent like this, but this change gives us
6674 strictly correct behaviour at least for TLS.
6677 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
6678 never resets s->method to s->ctx->method when called from within
6679 one of the SSL handshake functions.
6680 [Bodo Moeller; problem pointed out by Niko Baric]
6682 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
6683 (sent using the client's version number) if client_version is
6684 smaller than the protocol version in use. Also change
6685 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
6686 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
6687 the client will at least see that alert.
6690 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
6694 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
6695 client receives HelloRequest while in a handshake.
6696 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
6698 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
6699 should end in 'break', not 'goto end' which circuments various
6700 cleanups done in state SSL_ST_OK. But session related stuff
6701 must be disabled for SSL_ST_OK in the case that we just sent a
6704 Also avoid some overhead by not calling ssl_init_wbio_buffer()
6705 before just sending a HelloRequest.
6706 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
6708 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
6709 reveal whether illegal block cipher padding was found or a MAC
6710 verification error occured. (Neither SSLerr() codes nor alerts
6711 are directly visible to potential attackers, but the information
6712 may leak via logfiles.)
6714 Similar changes are not required for the SSL 2.0 implementation
6715 because the number of padding bytes is sent in clear for SSL 2.0,
6716 and the extra bytes are just ignored. However ssl/s2_pkt.c
6717 failed to verify that the purported number of padding bytes is in
6721 *) Add OpenUNIX-8 support including shared libraries
6722 (Boyd Lynn Gerber <gerberb@zenez.com>).
6725 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
6726 'wristwatch attack' using huge encoding parameters (cf.
6727 James H. Manger's CRYPTO 2001 paper). Note that the
6728 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
6729 encoding parameters and hence was not vulnerable.
6732 *) BN_sqr() bug fix.
6733 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
6735 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
6736 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
6737 followed by modular reduction.
6738 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
6740 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
6741 equivalent based on BN_pseudo_rand() instead of BN_rand().
6744 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
6745 This function was broken, as the check for a new client hello message
6746 to handle SGC did not allow these large messages.
6747 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
6750 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
6753 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
6754 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
6757 *) Rework the configuration and shared library support for Tru64 Unix.
6758 The configuration part makes use of modern compiler features and
6759 still retains old compiler behavior for those that run older versions
6760 of the OS. The shared library support part includes a variant that
6761 uses the RPATH feature, and is available through the special
6762 configuration target "alpha-cc-rpath", which will never be selected
6764 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
6766 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
6767 with the same message size as in ssl3_get_certificate_request().
6768 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
6769 messages might inadvertently be reject as too long.
6770 [Petr Lampa <lampa@fee.vutbr.cz>]
6772 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
6775 *) Modified SSL library such that the verify_callback that has been set
6776 specificly for an SSL object with SSL_set_verify() is actually being
6777 used. Before the change, a verify_callback set with this function was
6778 ignored and the verify_callback() set in the SSL_CTX at the time of
6779 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
6780 to allow the necessary settings.
6783 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
6784 explicitly to NULL, as at least on Solaris 8 this seems not always to be
6785 done automatically (in contradiction to the requirements of the C
6786 standard). This made problems when used from OpenSSH.
6789 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
6790 dh->length and always used
6792 BN_rand_range(priv_key, dh->p).
6794 BN_rand_range() is not necessary for Diffie-Hellman, and this
6795 specific range makes Diffie-Hellman unnecessarily inefficient if
6796 dh->length (recommended exponent length) is much smaller than the
6797 length of dh->p. We could use BN_rand_range() if the order of
6798 the subgroup was stored in the DH structure, but we only have
6803 BN_rand(priv_key, l, ...)
6805 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6811 RSA_eay_public_encrypt
6812 RSA_eay_private_decrypt
6813 RSA_eay_private_encrypt (signing)
6814 RSA_eay_public_decrypt (signature verification)
6816 (default implementations for RSA_public_encrypt,
6817 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6818 always reject numbers >= n.
6821 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6822 to synchronize access to 'locking_thread'. This is necessary on
6823 systems where access to 'locking_thread' (an 'unsigned long'
6824 variable) is not atomic.
6827 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6828 *before* setting the 'crypto_lock_rand' flag. The previous code had
6829 a race condition if 0 is a valid thread ID.
6830 [Travis Vitek <vitek@roguewave.com>]
6832 *) Add support for shared libraries under Irix.
6833 [Albert Chin-A-Young <china@thewrittenword.com>]
6835 *) Add configuration option to build on Linux on both big-endian and
6837 [Ralf Baechle <ralf@uni-koblenz.de>]
6839 *) Add the possibility to create shared libraries on HP-UX.
6842 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6844 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6845 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6846 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6847 PRNG state recovery was possible based on the output of
6848 one PRNG request appropriately sized to gain knowledge on
6849 'md' followed by enough consecutive 1-byte PRNG requests
6850 to traverse all of 'state'.
6852 1. When updating 'md_local' (the current thread's copy of 'md')
6853 during PRNG output generation, hash all of the previous
6854 'md_local' value, not just the half used for PRNG output.
6856 2. Make the number of bytes from 'state' included into the hash
6857 independent from the number of PRNG bytes requested.
6859 The first measure alone would be sufficient to avoid
6860 Markku-Juhani's attack. (Actually it had never occurred
6861 to me that the half of 'md_local' used for chaining was the
6862 half from which PRNG output bytes were taken -- I had always
6863 assumed that the secret half would be used.) The second
6864 measure makes sure that additional data from 'state' is never
6865 mixed into 'md_local' in small portions; this heuristically
6866 further strengthens the PRNG.
6869 *) Fix crypto/bn/asm/mips3.s.
6872 *) When only the key is given to "enc", the IV is undefined. Print out
6873 an error message in this case.
6876 *) Handle special case when X509_NAME is empty in X509 printing routines.
6879 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6880 positive and less than q.
6883 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6884 used: it isn't thread safe and the add_lock_callback should handle
6886 [Paul Rose <Paul.Rose@bridge.com>]
6888 *) Verify that incoming data obeys the block size in
6889 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6893 [Ulf Möller, Bodo Möller]
6895 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6896 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6897 when fixing the server behaviour for backwards-compatible 'client
6898 hello' messages. (Note that the attack is impractical against
6899 SSL 3.0 and TLS 1.0 anyway because length and version checking
6900 means that the probability of guessing a valid ciphertext is
6901 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6904 Before 0.9.5, the countermeasure (hide the error by generating a
6905 random 'decryption result') did not work properly because
6906 ERR_clear_error() was missing, meaning that SSL_get_error() would
6907 detect the supposedly ignored error.
6909 Both problems are now fixed.
6912 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6913 (previously it was 1024).
6916 *) Fix for compatibility mode trust settings: ignore trust settings
6917 unless some valid trust or reject settings are present.
6920 *) Fix for blowfish EVP: its a variable length cipher.
6923 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6924 parameters in DSA public key structures and return an error in the
6925 DSA routines if parameters are absent.
6928 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6929 in the current directory if neither $RANDFILE nor $HOME was set.
6930 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6931 caused some confusion to Windows users who haven't defined $HOME.
6932 Thus RAND_file_name() is changed again: e_os.h can define a
6933 DEFAULT_HOME, which will be used if $HOME is not set.
6934 For Windows, we use "C:"; on other platforms, we still require
6935 environment variables.
6937 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6938 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6939 having multiple threads call RAND_poll() concurrently.
6942 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6943 combination of a flag and a thread ID variable.
6944 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6945 flag), *other* threads can enter ssleay_add_bytes without obeying
6946 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6947 that they do not hold after the first thread unsets add_do_not_lock).
6950 *) Change bctest again: '-x' expressions are not available in all
6954 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6956 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6957 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6959 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6960 the default extension for executables, if any. Also, make the perl
6961 scripts that use symlink() to test if it really exists and use "cp"
6962 if it doesn't. All this made OpenSSL compilable and installable in
6966 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6967 If SEQUENCE is length is indefinite just set c->slen to the total
6968 amount of data available.
6969 [Steve Henson, reported by shige@FreeBSD.org]
6970 [This change does not apply to 0.9.7.]
6972 *) Change bctest to avoid here-documents inside command substitution
6973 (workaround for FreeBSD /bin/sh bug).
6974 For compatibility with Ultrix, avoid shell functions (introduced
6975 in the bctest version that searches along $PATH).
6978 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6979 with des_encrypt() defined on some operating systems, like Solaris
6983 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6984 On the Importance of Eliminating Errors in Cryptographic
6985 Computations, J. Cryptology 14 (2001) 2, 101-119,
6986 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6989 *) MIPS assembler BIGNUM division bug fix.
6992 *) Disabled incorrect Alpha assembler code.
6995 *) Fix PKCS#7 decode routines so they correctly update the length
6996 after reading an EOC for the EXPLICIT tag.
6998 [This change does not apply to 0.9.7.]
7000 *) Fix bug in PKCS#12 key generation routines. This was triggered
7001 if a 3DES key was generated with a 0 initial byte. Include
7002 PKCS12_BROKEN_KEYGEN compilation option to retain the old
7003 (but broken) behaviour.
7006 *) Enhance bctest to search for a working bc along $PATH and print
7008 [Tim Rice <tim@multitalents.net> via Richard Levitte]
7010 *) Fix memory leaks in err.c: free err_data string if necessary;
7011 don't write to the wrong index in ERR_set_error_data.
7014 *) Implement ssl23_peek (analogous to ssl23_read), which previously
7018 *) Replace rdtsc with _emit statements for VC++ version 5.
7019 [Jeremy Cooper <jeremy@baymoo.org>]
7021 *) Make it possible to reuse SSLv2 sessions.
7024 *) In copy_email() check for >= 0 as a return value for
7025 X509_NAME_get_index_by_NID() since 0 is a valid index.
7026 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
7028 *) Avoid coredump with unsupported or invalid public keys by checking if
7029 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
7030 PKCS7_verify() fails with non detached data.
7033 *) Don't use getenv in library functions when run as setuid/setgid.
7034 New function OPENSSL_issetugid().
7037 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7038 due to incorrect handling of multi-threading:
7040 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
7042 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
7044 3. Count how many times MemCheck_off() has been called so that
7045 nested use can be treated correctly. This also avoids
7046 inband-signalling in the previous code (which relied on the
7047 assumption that thread ID 0 is impossible).
7050 *) Add "-rand" option also to s_client and s_server.
7053 *) Fix CPU detection on Irix 6.x.
7054 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
7055 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7057 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
7060 [This change does not apply to 0.9.7.]
7062 *) Use the cached encoding of an X509_NAME structure rather than
7063 copying it. This is apparently the reason for the libsafe "errors"
7064 but the code is actually correct.
7067 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
7068 Bleichenbacher's DSA attack.
7069 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
7070 to be set and top=0 forces the highest bit to be set; top=-1 is new
7071 and leaves the highest bit random.
7072 [Ulf Moeller, Bodo Moeller]
7074 *) In the NCONF_...-based implementations for CONF_... queries
7075 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
7076 a temporary CONF structure with the data component set to NULL
7077 (which gives segmentation faults in lh_retrieve).
7078 Instead, use NULL for the CONF pointer in CONF_get_string and
7079 CONF_get_number (which may use environment variables) and directly
7080 return NULL from CONF_get_section.
7083 *) Fix potential buffer overrun for EBCDIC.
7086 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7087 keyUsage if basicConstraints absent for a CA.
7090 *) Make SMIME_write_PKCS7() write mail header values with a format that
7091 is more generally accepted (no spaces before the semicolon), since
7092 some programs can't parse those values properly otherwise. Also make
7093 sure BIO's that break lines after each write do not create invalid
7097 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
7098 macros previously used would not encode an empty SEQUENCE OF
7099 and break the signature.
7101 [This change does not apply to 0.9.7.]
7103 *) Zero the premaster secret after deriving the master secret in
7107 *) Add some EVP_add_digest_alias registrations (as found in
7108 OpenSSL_add_all_digests()) to SSL_library_init()
7109 aka OpenSSL_add_ssl_algorithms(). This provides improved
7110 compatibility with peers using X.509 certificates
7111 with unconventional AlgorithmIdentifier OIDs.
7114 *) Fix for Irix with NO_ASM.
7115 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7117 *) ./config script fixes.
7118 [Ulf Moeller, Richard Levitte]
7120 *) Fix 'openssl passwd -1'.
7123 *) Change PKCS12_key_gen_asc() so it can cope with non null
7124 terminated strings whose length is passed in the passlen
7125 parameter, for example from PEM callbacks. This was done
7126 by adding an extra length parameter to asc2uni().
7127 [Steve Henson, reported by <oddissey@samsung.co.kr>]
7129 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
7130 call failed, free the DSA structure.
7133 *) Fix to uni2asc() to cope with zero length Unicode strings.
7134 These are present in some PKCS#12 files.
7137 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
7138 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
7139 when writing a 32767 byte record.
7140 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
7142 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
7143 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
7145 (RSA objects have a reference count access to which is protected
7146 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
7147 so they are meant to be shared between threads.)
7148 [Bodo Moeller, Geoff Thorpe; original patch submitted by
7149 "Reddie, Steven" <Steven.Reddie@ca.com>]
7151 *) Fix a deadlock in CRYPTO_mem_leaks().
7154 *) Use better test patterns in bntest.
7157 *) rand_win.c fix for Borland C.
7160 *) BN_rshift bugfix for n == 0.
7163 *) Add a 'bctest' script that checks for some known 'bc' bugs
7164 so that 'make test' does not abort just because 'bc' is broken.
7167 *) Store verify_result within SSL_SESSION also for client side to
7168 avoid potential security hole. (Re-used sessions on the client side
7169 always resulted in verify_result==X509_V_OK, not using the original
7170 result of the server certificate verification.)
7173 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
7174 SSL3_RT_APPLICATION_DATA, return 0.
7175 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
7179 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
7180 releases, have been re-implemented by renaming the previous
7181 implementations of ssl2_read and ssl3_read to ssl2_read_internal
7182 and ssl3_read_internal, respectively, and adding 'peek' parameters
7183 to them. The new ssl[23]_{read,peek} functions are calls to
7184 ssl[23]_read_internal with the 'peek' flag set appropriately.
7185 A 'peek' parameter has also been added to ssl3_read_bytes, which
7186 does the actual work for ssl3_read_internal.
7189 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
7190 the method-specific "init()" handler. Also clean up ex_data after
7191 calling the method-specific "finish()" handler. Previously, this was
7192 happening the other way round.
7195 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
7196 The previous value, 12, was not always sufficient for BN_mod_exp().
7199 *) Make sure that shared libraries get the internal name engine with
7200 the full version number and not just 0. This should mark the
7201 shared libraries as not backward compatible. Of course, this should
7202 be changed again when we can guarantee backward binary compatibility.
7205 *) Fix typo in get_cert_by_subject() in by_dir.c
7206 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
7208 *) Rework the system to generate shared libraries:
7210 - Make note of the expected extension for the shared libraries and
7211 if there is a need for symbolic links from for example libcrypto.so.0
7212 to libcrypto.so.0.9.7. There is extended info in Configure for
7215 - Make as few rebuilds of the shared libraries as possible.
7217 - Still avoid linking the OpenSSL programs with the shared libraries.
7219 - When installing, install the shared libraries separately from the
7223 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
7225 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
7226 and not in SSL_clear because the latter is also used by the
7227 accept/connect functions; previously, the settings made by
7228 SSL_set_read_ahead would be lost during the handshake.
7229 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
7231 *) Correct util/mkdef.pl to be selective about disabled algorithms.
7232 Previously, it would create entries for disableed algorithms no
7236 *) Added several new manual pages for SSL_* function.
7239 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
7241 *) In ssl23_get_client_hello, generate an error message when faced
7242 with an initial SSL 3.0/TLS record that is too small to contain the
7243 first two bytes of the ClientHello message, i.e. client_version.
7244 (Note that this is a pathologic case that probably has never happened
7245 in real life.) The previous approach was to use the version number
7246 from the record header as a substitute; but our protocol choice
7247 should not depend on that one because it is not authenticated
7248 by the Finished messages.
7251 *) More robust randomness gathering functions for Windows.
7252 [Jeffrey Altman <jaltman@columbia.edu>]
7254 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
7255 not set then we don't setup the error code for issuer check errors
7256 to avoid possibly overwriting other errors which the callback does
7257 handle. If an application does set the flag then we assume it knows
7258 what it is doing and can handle the new informational codes
7262 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
7263 a general "ANY" type, as such it should be able to decode anything
7264 including tagged types. However it didn't check the class so it would
7265 wrongly interpret tagged types in the same way as their universal
7266 counterpart and unknown types were just rejected. Changed so that the
7267 tagged and unknown types are handled in the same way as a SEQUENCE:
7268 that is the encoding is stored intact. There is also a new type
7269 "V_ASN1_OTHER" which is used when the class is not universal, in this
7270 case we have no idea what the actual type is so we just lump them all
7274 *) On VMS, stdout may very well lead to a file that is written to
7275 in a record-oriented fashion. That means that every write() will
7276 write a separate record, which will be read separately by the
7277 programs trying to read from it. This can be very confusing.
7279 The solution is to put a BIO filter in the way that will buffer
7280 text until a linefeed is reached, and then write everything a
7281 line at a time, so every record written will be an actual line,
7282 not chunks of lines and not (usually doesn't happen, but I've
7283 seen it once) several lines in one record. BIO_f_linebuffer() is
7286 Currently, it's a VMS-only method, because that's where it has
7287 been tested well enough.
7290 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
7291 it can return incorrect results.
7292 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
7293 but it was in 0.9.6-beta[12].)
7296 *) Disable the check for content being present when verifying detached
7297 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
7298 include zero length content when signing messages.
7301 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
7302 BIO_ctrl (for BIO pairs).
7305 *) Add DSO method for VMS.
7308 *) Bug fix: Montgomery multiplication could produce results with the
7312 *) Add RPM specification openssl.spec and modify it to build three
7313 packages. The default package contains applications, application
7314 documentation and run-time libraries. The devel package contains
7315 include files, static libraries and function documentation. The
7316 doc package contains the contents of the doc directory. The original
7317 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
7320 *) Add a large number of documentation files for many SSL routines.
7321 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7323 *) Add a configuration entry for Sony News 4.
7324 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
7326 *) Don't set the two most significant bits to one when generating a
7327 random number < q in the DSA library.
7330 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
7331 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
7332 the underlying transport is blocking) if a handshake took place.
7333 (The default behaviour is needed by applications such as s_client
7334 and s_server that use select() to determine when to use SSL_read;
7335 but for applications that know in advance when to expect data, it
7336 just makes things more complicated.)
7339 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
7343 *) Add a few more EBCDIC conditionals that make `req' and `x509'
7344 work better on such systems.
7345 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7347 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
7348 Update PKCS12_parse() so it copies the friendlyName and the
7349 keyid to the certificates aux info.
7352 *) Fix bug in PKCS7_verify() which caused an infinite loop
7353 if there was more than one signature.
7354 [Sven Uszpelkat <su@celocom.de>]
7356 *) Major change in util/mkdef.pl to include extra information
7357 about each symbol, as well as presentig variables as well
7358 as functions. This change means that there's n more need
7359 to rebuild the .num files when some algorithms are excluded.
7362 *) Allow the verify time to be set by an application,
7363 rather than always using the current time.
7366 *) Phase 2 verify code reorganisation. The certificate
7367 verify code now looks up an issuer certificate by a
7368 number of criteria: subject name, authority key id
7369 and key usage. It also verifies self signed certificates
7370 by the same criteria. The main comparison function is
7371 X509_check_issued() which performs these checks.
7373 Lot of changes were necessary in order to support this
7374 without completely rewriting the lookup code.
7376 Authority and subject key identifier are now cached.
7378 The LHASH 'certs' is X509_STORE has now been replaced
7379 by a STACK_OF(X509_OBJECT). This is mainly because an
7380 LHASH can't store or retrieve multiple objects with
7381 the same hash value.
7383 As a result various functions (which were all internal
7384 use only) have changed to handle the new X509_STORE
7385 structure. This will break anything that messed round
7386 with X509_STORE internally.
7388 The functions X509_STORE_add_cert() now checks for an
7389 exact match, rather than just subject name.
7391 The X509_STORE API doesn't directly support the retrieval
7392 of multiple certificates matching a given criteria, however
7393 this can be worked round by performing a lookup first
7394 (which will fill the cache with candidate certificates)
7395 and then examining the cache for matches. This is probably
7396 the best we can do without throwing out X509_LOOKUP
7397 entirely (maybe later...).
7399 The X509_VERIFY_CTX structure has been enhanced considerably.
7401 All certificate lookup operations now go via a get_issuer()
7402 callback. Although this currently uses an X509_STORE it
7403 can be replaced by custom lookups. This is a simple way
7404 to bypass the X509_STORE hackery necessary to make this
7405 work and makes it possible to use more efficient techniques
7406 in future. A very simple version which uses a simple
7407 STACK for its trusted certificate store is also provided
7408 using X509_STORE_CTX_trusted_stack().
7410 The verify_cb() and verify() callbacks now have equivalents
7411 in the X509_STORE_CTX structure.
7413 X509_STORE_CTX also has a 'flags' field which can be used
7414 to customise the verify behaviour.
7417 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
7418 excludes S/MIME capabilities.
7421 *) When a certificate request is read in keep a copy of the
7422 original encoding of the signed data and use it when outputing
7423 again. Signatures then use the original encoding rather than
7424 a decoded, encoded version which may cause problems if the
7425 request is improperly encoded.
7428 *) For consistency with other BIO_puts implementations, call
7429 buffer_write(b, ...) directly in buffer_puts instead of calling
7432 In BIO_puts, increment b->num_write as in BIO_write.
7433 [Peter.Sylvester@EdelWeb.fr]
7435 *) Fix BN_mul_word for the case where the word is 0. (We have to use
7436 BN_zero, we may not return a BIGNUM with an array consisting of
7440 *) Avoid calling abort() from within the library when problems are
7441 detected, except if preprocessor symbols have been defined
7442 (such as REF_CHECK, BN_DEBUG etc.).
7445 *) New openssl application 'rsautl'. This utility can be
7446 used for low level RSA operations. DER public key
7447 BIO/fp routines also added.
7450 *) New Configure entry and patches for compiling on QNX 4.
7451 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
7453 *) A demo state-machine implementation was sponsored by
7454 Nuron (http://www.nuron.com/) and is now available in
7455 demos/state_machine.
7458 *) New options added to the 'dgst' utility for signature
7459 generation and verification.
7462 *) Unrecognized PKCS#7 content types are now handled via a
7463 catch all ASN1_TYPE structure. This allows unsupported
7464 types to be stored as a "blob" and an application can
7465 encode and decode it manually.
7468 *) Fix various signed/unsigned issues to make a_strex.c
7470 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
7472 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
7473 length if passed a buffer. ASN1_INTEGER_to_BN failed
7474 if passed a NULL BN and its argument was negative.
7475 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
7477 *) Modification to PKCS#7 encoding routines to output definite
7478 length encoding. Since currently the whole structures are in
7479 memory there's not real point in using indefinite length
7480 constructed encoding. However if OpenSSL is compiled with
7481 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
7484 *) Added BIO_vprintf() and BIO_vsnprintf().
7487 *) Added more prefixes to parse for in the the strings written
7488 through a logging bio, to cover all the levels that are available
7489 through syslog. The prefixes are now:
7491 PANIC, EMERG, EMR => LOG_EMERG
7492 ALERT, ALR => LOG_ALERT
7493 CRIT, CRI => LOG_CRIT
7494 ERROR, ERR => LOG_ERR
7495 WARNING, WARN, WAR => LOG_WARNING
7496 NOTICE, NOTE, NOT => LOG_NOTICE
7497 INFO, INF => LOG_INFO
7498 DEBUG, DBG => LOG_DEBUG
7500 and as before, if none of those prefixes are present at the
7501 beginning of the string, LOG_ERR is chosen.
7503 On Win32, the LOG_* levels are mapped according to this:
7505 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
7506 LOG_WARNING => EVENTLOG_WARNING_TYPE
7507 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
7511 *) Made it possible to reconfigure with just the configuration
7512 argument "reconf" or "reconfigure". The command line arguments
7513 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
7514 and are retrieved from there when reconfiguring.
7518 [Assar Westerlund <assar@sics.se>, Richard Levitte]
7520 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
7523 *) The obj_dat.pl script was messing up the sorting of object
7524 names. The reason was that it compared the quoted version
7525 of strings as a result "OCSP" > "OCSP Signing" because
7526 " > SPACE. Changed script to store unquoted versions of
7527 names and add quotes on output. It was also omitting some
7528 names from the lookup table if they were given a default
7529 value (that is if SN is missing it is given the same
7530 value as LN and vice versa), these are now added on the
7531 grounds that if an object has a name we should be able to
7532 look it up. Finally added warning output when duplicate
7533 short or long names are found.
7536 *) Changes needed for Tandem NSK.
7537 [Scott Uroff <scott@xypro.com>]
7539 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
7540 RSA_padding_check_SSLv23(), special padding was never detected
7541 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
7542 version rollback attacks was not effective.
7544 In s23_clnt.c, don't use special rollback-attack detection padding
7545 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
7546 client; similarly, in s23_srvr.c, don't do the rollback check if
7547 SSL 2.0 is the only protocol enabled in the server.
7550 *) Make it possible to get hexdumps of unprintable data with 'openssl
7551 asn1parse'. By implication, the functions ASN1_parse_dump() and
7552 BIO_dump_indent() are added.
7555 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
7556 these print out strings and name structures based on various
7557 flags including RFC2253 support and proper handling of
7558 multibyte characters. Added options to the 'x509' utility
7559 to allow the various flags to be set.
7562 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
7563 Also change the functions X509_cmp_current_time() and
7564 X509_gmtime_adj() work with an ASN1_TIME structure,
7565 this will enable certificates using GeneralizedTime in validity
7566 dates to be checked.
7569 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
7570 negative public key encodings) on by default,
7571 NO_NEG_PUBKEY_BUG can be set to disable it.
7574 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
7575 content octets. An i2c_ASN1_OBJECT is unnecessary because
7576 the encoding can be trivially obtained from the structure.
7579 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
7580 not read locks (CRYPTO_r_[un]lock).
7583 *) A first attempt at creating official support for shared
7584 libraries through configuration. I've kept it so the
7585 default is static libraries only, and the OpenSSL programs
7586 are always statically linked for now, but there are
7587 preparations for dynamic linking in place.
7588 This has been tested on Linux and Tru64.
7591 *) Randomness polling function for Win9x, as described in:
7592 Peter Gutmann, Software Generation of Practically Strong
7596 *) Fix so PRNG is seeded in req if using an already existing
7600 *) New options to smime application. -inform and -outform
7601 allow alternative formats for the S/MIME message including
7602 PEM and DER. The -content option allows the content to be
7603 specified separately. This should allow things like Netscape
7604 form signing output easier to verify.
7607 *) Fix the ASN1 encoding of tags using the 'long form'.
7610 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
7611 STRING types. These convert content octets to and from the
7612 underlying type. The actual tag and length octets are
7613 already assumed to have been read in and checked. These
7614 are needed because all other string types have virtually
7615 identical handling apart from the tag. By having versions
7616 of the ASN1 functions that just operate on content octets
7617 IMPLICIT tagging can be handled properly. It also allows
7618 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
7619 and ASN1_INTEGER are identical apart from the tag.
7622 *) Change the handling of OID objects as follows:
7624 - New object identifiers are inserted in objects.txt, following
7625 the syntax given in objects.README.
7626 - objects.pl is used to process obj_mac.num and create a new
7628 - obj_dat.pl is used to create a new obj_dat.h, using the data in
7631 This is currently kind of a hack, and the perl code in objects.pl
7632 isn't very elegant, but it works as I intended. The simplest way
7633 to check that it worked correctly is to look in obj_dat.h and
7634 check the array nid_objs and make sure the objects haven't moved
7635 around (this is important!). Additions are OK, as well as
7636 consistent name changes.
7639 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
7642 *) Addition of the command line parameter '-rand file' to 'openssl req'.
7643 The given file adds to whatever has already been seeded into the
7644 random pool through the RANDFILE configuration file option or
7645 environment variable, or the default random state file.
7648 *) mkstack.pl now sorts each macro group into lexical order.
7649 Previously the output order depended on the order the files
7650 appeared in the directory, resulting in needless rewriting
7654 *) Patches to make OpenSSL compile under Win32 again. Mostly
7655 work arounds for the VC++ problem that it treats func() as
7656 func(void). Also stripped out the parts of mkdef.pl that
7657 added extra typesafe functions: these no longer exist.
7660 *) Reorganisation of the stack code. The macros are now all
7661 collected in safestack.h . Each macro is defined in terms of
7662 a "stack macro" of the form SKM_<name>(type, a, b). The
7663 DEBUG_SAFESTACK is now handled in terms of function casts,
7664 this has the advantage of retaining type safety without the
7665 use of additional functions. If DEBUG_SAFESTACK is not defined
7666 then the non typesafe macros are used instead. Also modified the
7667 mkstack.pl script to handle the new form. Needs testing to see
7668 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
7669 the default if no major problems. Similar behaviour for ASN1_SET_OF
7670 and PKCS12_STACK_OF.
7673 *) When some versions of IIS use the 'NET' form of private key the
7674 key derivation algorithm is different. Normally MD5(password) is
7675 used as a 128 bit RC4 key. In the modified case
7676 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
7677 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
7678 as the old Netscape_RSA functions except they have an additional
7679 'sgckey' parameter which uses the modified algorithm. Also added
7680 an -sgckey command line option to the rsa utility. Thanks to
7681 Adrian Peck <bertie@ncipher.com> for posting details of the modified
7682 algorithm to openssl-dev.
7685 *) The evp_local.h macros were using 'c.##kname' which resulted in
7686 invalid expansion on some systems (SCO 5.0.5 for example).
7687 Corrected to 'c.kname'.
7688 [Phillip Porch <root@theporch.com>]
7690 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
7691 a STACK of email addresses from a certificate or request, these look
7692 in the subject name and the subject alternative name extensions and
7693 omit any duplicate addresses.
7696 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
7697 This makes DSA verification about 2 % faster.
7700 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
7701 (meaning that now 2^5 values will be precomputed, which is only 4 KB
7702 plus overhead for 1024 bit moduli).
7703 This makes exponentiations about 0.5 % faster for 1024 bit
7704 exponents (as measured by "openssl speed rsa2048").
7707 *) Rename memory handling macros to avoid conflicts with other
7709 Malloc => OPENSSL_malloc
7710 Malloc_locked => OPENSSL_malloc_locked
7711 Realloc => OPENSSL_realloc
7712 Free => OPENSSL_free
7715 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
7716 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
7719 *) CygWin32 support.
7720 [John Jarvie <jjarvie@newsguy.com>]
7722 *) The type-safe stack code has been rejigged. It is now only compiled
7723 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
7724 by default all type-specific stack functions are "#define"d back to
7725 standard stack functions. This results in more streamlined output
7726 but retains the type-safety checking possibilities of the original
7730 *) The STACK code has been cleaned up, and certain type declarations
7731 that didn't make a lot of sense have been brought in line. This has
7732 also involved a cleanup of sorts in safestack.h to more correctly
7733 map type-safe stack functions onto their plain stack counterparts.
7734 This work has also resulted in a variety of "const"ifications of
7735 lots of the code, especially "_cmp" operations which should normally
7736 be prototyped with "const" parameters anyway.
7739 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
7740 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
7741 (The PRNG state consists of two parts, the large pool 'state' and 'md',
7742 where all of 'md' is used each time the PRNG is used, but 'state'
7743 is used only indexed by a cyclic counter. As entropy may not be
7744 well distributed from the beginning, 'md' is important as a
7745 chaining variable. However, the output function chains only half
7746 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
7747 all of 'md', and seeding with STATE_SIZE dummy bytes will result
7748 in all of 'state' being rewritten, with the new values depending
7749 on virtually all of 'md'. This overcomes the 80 bit limitation.)
7752 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
7753 the handshake is continued after ssl_verify_cert_chain();
7754 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
7755 can lead to 'unexplainable' connection aborts later.
7756 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
7758 *) Major EVP API cipher revision.
7759 Add hooks for extra EVP features. This allows various cipher
7760 parameters to be set in the EVP interface. Support added for variable
7761 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
7762 setting of RC2 and RC5 parameters.
7764 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
7767 Remove lots of duplicated code from the EVP library. For example *every*
7768 cipher init() function handles the 'iv' in the same way according to the
7769 cipher mode. They also all do nothing if the 'key' parameter is NULL and
7770 for CFB and OFB modes they zero ctx->num.
7772 New functionality allows removal of S/MIME code RC2 hack.
7774 Most of the routines have the same form and so can be declared in terms
7777 By shifting this to the top level EVP_CipherInit() it can be removed from
7778 all individual ciphers. If the cipher wants to handle IVs or keys
7779 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
7782 Change lots of functions like EVP_EncryptUpdate() to now return a
7783 value: although software versions of the algorithms cannot fail
7784 any installed hardware versions can.
7787 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
7788 this option is set, tolerate broken clients that send the negotiated
7789 protocol version number instead of the requested protocol version
7793 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7794 i.e. non-zero for export ciphersuites, zero otherwise.
7795 Previous versions had this flag inverted, inconsistent with
7796 rsa_tmp_cb (..._TMP_RSA_CB).
7797 [Bodo Moeller; problem reported by Amit Chopra]
7799 *) Add missing DSA library text string. Work around for some IIS
7800 key files with invalid SEQUENCE encoding.
7803 *) Add a document (doc/standards.txt) that list all kinds of standards
7804 and so on that are implemented in OpenSSL.
7807 *) Enhance c_rehash script. Old version would mishandle certificates
7808 with the same subject name hash and wouldn't handle CRLs at all.
7809 Added -fingerprint option to crl utility, to support new c_rehash
7813 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7816 *) Fix for SSL server purpose checking. Server checking was
7817 rejecting certificates which had extended key usage present
7818 but no ssl client purpose.
7819 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7821 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7822 is a little unclear about how a blank password is handled.
7823 Since the password in encoded as a BMPString with terminating
7824 double NULL a zero length password would end up as just the
7825 double NULL. However no password at all is different and is
7826 handled differently in the PKCS#12 key generation code. NS
7827 treats a blank password as zero length. MSIE treats it as no
7828 password on export: but it will try both on import. We now do
7829 the same: PKCS12_parse() tries zero length and no password if
7830 the password is set to "" or NULL (NULL is now a valid password:
7831 it wasn't before) as does the pkcs12 application.
7834 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7835 perror when PEM_read_bio_X509_REQ fails, the error message must
7836 be obtained from the error queue.
7839 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7840 it in ERR_remove_state if appropriate, and change ERR_get_state
7841 accordingly to avoid race conditions (this is necessary because
7842 thread_hash is no longer constant once set).
7845 *) Bugfix for linux-elf makefile.one.
7848 *) RSA_get_default_method() will now cause a default
7849 RSA_METHOD to be chosen if one doesn't exist already.
7850 Previously this was only set during a call to RSA_new()
7851 or RSA_new_method(NULL) meaning it was possible for
7852 RSA_get_default_method() to return NULL.
7855 *) Added native name translation to the existing DSO code
7856 that will convert (if the flag to do so is set) filenames
7857 that are sufficiently small and have no path information
7858 into a canonical native form. Eg. "blah" converted to
7859 "libblah.so" or "blah.dll" etc.
7862 *) New function ERR_error_string_n(e, buf, len) which is like
7863 ERR_error_string(e, buf), but writes at most 'len' bytes
7864 including the 0 terminator. For ERR_error_string_n, 'buf'
7866 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7868 *) CONF library reworked to become more general. A new CONF
7869 configuration file reader "class" is implemented as well as a
7870 new functions (NCONF_*, for "New CONF") to handle it. The now
7871 old CONF_* functions are still there, but are reimplemented to
7872 work in terms of the new functions. Also, a set of functions
7873 to handle the internal storage of the configuration data is
7874 provided to make it easier to write new configuration file
7875 reader "classes" (I can definitely see something reading a
7876 configuration file in XML format, for example), called _CONF_*,
7877 or "the configuration storage API"...
7879 The new configuration file reading functions are:
7881 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7882 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7884 NCONF_default, NCONF_WIN32
7886 NCONF_dump_fp, NCONF_dump_bio
7888 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7889 NCONF_new creates a new CONF object. This works in the same way
7890 as other interfaces in OpenSSL, like the BIO interface.
7891 NCONF_dump_* dump the internal storage of the configuration file,
7892 which is useful for debugging. All other functions take the same
7893 arguments as the old CONF_* functions wth the exception of the
7894 first that must be a `CONF *' instead of a `LHASH *'.
7896 To make it easer to use the new classes with the old CONF_* functions,
7897 the function CONF_set_default_method is provided.
7900 *) Add '-tls1' option to 'openssl ciphers', which was already
7901 mentioned in the documentation but had not been implemented.
7902 (This option is not yet really useful because even the additional
7903 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7906 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7907 OpenSSL-based applications) load shared libraries and bind to
7908 them in a portable way.
7909 [Geoff Thorpe, with contributions from Richard Levitte]
7911 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7913 *) Make sure _lrotl and _lrotr are only used with MSVC.
7915 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7916 (the default implementation of RAND_status).
7918 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7919 to '-clrext' (= clear extensions), as intended and documented.
7920 [Bodo Moeller; inconsistency pointed out by Michael Attili
7923 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7924 was larger than the MD block size.
7925 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7927 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7928 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7929 using the passed key: if the passed key was a private key the result
7930 of X509_print(), for example, would be to print out all the private key
7934 *) des_quad_cksum() byte order bug fix.
7935 [Ulf Möller, using the problem description in krb4-0.9.7, where
7936 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7938 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7940 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7942 *) For easily testing in shell scripts whether some command
7943 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7944 returns with exit code 0 iff no command of the given name is available.
7945 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7946 the output goes to stdout and nothing is printed to stderr.
7947 Additional arguments are always ignored.
7949 Since for each cipher there is a command of the same name,
7950 the 'no-cipher' compilation switches can be tested this way.
7952 ('openssl no-XXX' is not able to detect pseudo-commands such
7953 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7956 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7959 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7960 is set; it will be thrown away anyway because each handshake creates
7962 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7963 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7964 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7965 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7968 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7969 'Q' and 'R' lose their special meanings (quit/renegotiate).
7970 This is part of what -quiet does; unlike -quiet, -ign_eof
7971 does not suppress any output.
7974 *) Add compatibility options to the purpose and trust code. The
7975 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7976 accepts a certificate or CA, this was the previous behaviour,
7977 with all the associated security issues.
7979 X509_TRUST_COMPAT is the old trust behaviour: only and
7980 automatically trust self signed roots in certificate store. A
7981 new trust setting X509_TRUST_DEFAULT is used to specify that
7982 a purpose has no associated trust setting and it should instead
7983 use the value in the default purpose.
7986 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7987 and fix a memory leak.
7990 *) In util/mkerr.pl (which implements 'make errors'), preserve
7991 reason strings from the previous version of the .c file, as
7992 the default to have only downcase letters (and digits) in
7993 automatically generated reasons codes is not always appropriate.
7996 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7997 using strerror. Previously, ERR_reason_error_string() returned
7998 library names as reason strings for SYSerr; but SYSerr is a special
7999 case where small numbers are errno values, not library numbers.
8002 *) Add '-dsaparam' option to 'openssl dhparam' application. This
8003 converts DSA parameters into DH parameters. (When creating parameters,
8004 DSA_generate_parameters is used.)
8007 *) Include 'length' (recommended exponent length) in C code generated
8008 by 'openssl dhparam -C'.
8011 *) The second argument to set_label in perlasm was already being used
8012 so couldn't be used as a "file scope" flag. Moved to third argument
8016 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
8017 instead of RAND_bytes for encryption IVs and salts.
8020 *) Include RAND_status() into RAND_METHOD instead of implementing
8021 it only for md_rand.c Otherwise replacing the PRNG by calling
8022 RAND_set_rand_method would be impossible.
8025 *) Don't let DSA_generate_key() enter an infinite loop if the random
8026 number generation fails.
8029 *) New 'rand' application for creating pseudo-random output.
8032 *) Added configuration support for Linux/IA64
8033 [Rolf Haberrecker <rolf@suse.de>]
8035 *) Assembler module support for Mingw32.
8038 *) Shared library support for HPUX (in shlib/).
8039 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
8041 *) Shared library support for Solaris gcc.
8042 [Lutz Behnke <behnke@trustcenter.de>]
8044 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
8046 *) PKCS7_encrypt() was adding text MIME headers twice because they
8047 were added manually and by SMIME_crlf_copy().
8050 *) In bntest.c don't call BN_rand with zero bits argument.
8051 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
8053 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
8054 case was implemented. This caused BN_div_recp() to fail occasionally.
8057 *) Add an optional second argument to the set_label() in the perl
8058 assembly language builder. If this argument exists and is set
8059 to 1 it signals that the assembler should use a symbol whose
8060 scope is the entire file, not just the current function. This
8061 is needed with MASM which uses the format label:: for this scope.
8062 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
8064 *) Change the ASN1 types so they are typedefs by default. Before
8065 almost all types were #define'd to ASN1_STRING which was causing
8066 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
8070 *) Change names of new functions to the new get1/get0 naming
8071 convention: After 'get1', the caller owns a reference count
8072 and has to call ..._free; 'get0' returns a pointer to some
8073 data structure without incrementing reference counters.
8074 (Some of the existing 'get' functions increment a reference
8075 counter, some don't.)
8076 Similarly, 'set1' and 'add1' functions increase reference
8077 counters or duplicate objects.
8080 *) Allow for the possibility of temp RSA key generation failure:
8081 the code used to assume it always worked and crashed on failure.
8084 *) Fix potential buffer overrun problem in BIO_printf().
8085 [Ulf Möller, using public domain code by Patrick Powell; problem
8086 pointed out by David Sacerdote <das33@cornell.edu>]
8088 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
8089 RAND_egd() and RAND_status(). In the command line application,
8090 the EGD socket can be specified like a seed file using RANDFILE
8094 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
8095 Some CAs (e.g. Verisign) distribute certificates in this form.
8098 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
8099 list to exclude them. This means that no special compilation option
8100 is needed to use anonymous DH: it just needs to be included in the
8104 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
8105 EVP_MD_type. The old functionality is available in a new macro called
8106 EVP_MD_md(). Change code that uses it and update docs.
8109 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
8110 where the 'void *' argument is replaced by a function pointer argument.
8111 Previously 'void *' was abused to point to functions, which works on
8112 many platforms, but is not correct. As these functions are usually
8113 called by macros defined in OpenSSL header files, most source code
8114 should work without changes.
8117 *) <openssl/opensslconf.h> (which is created by Configure) now contains
8118 sections with information on -D... compiler switches used for
8119 compiling the library so that applications can see them. To enable
8120 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
8121 must be defined. E.g.,
8122 #define OPENSSL_ALGORITHM_DEFINES
8123 #include <openssl/opensslconf.h>
8124 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
8125 [Richard Levitte, Ulf and Bodo Möller]
8127 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
8131 *) Change the 'other' type in certificate aux info to a STACK_OF
8132 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
8133 the required ASN1 format: arbitrary types determined by an OID.
8136 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
8137 argument to 'req'. This is not because the function is newer or
8138 better than others it just uses the work 'NEW' in the certificate
8139 request header lines. Some software needs this.
8142 *) Reorganise password command line arguments: now passwords can be
8143 obtained from various sources. Delete the PEM_cb function and make
8144 it the default behaviour: i.e. if the callback is NULL and the
8145 usrdata argument is not NULL interpret it as a null terminated pass
8146 phrase. If usrdata and the callback are NULL then the pass phrase
8147 is prompted for as usual.
8150 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
8151 the support is automatically enabled. The resulting binaries will
8152 autodetect the card and use it if present.
8153 [Ben Laurie and Compaq Inc.]
8155 *) Work around for Netscape hang bug. This sends certificate request
8156 and server done in one record. Since this is perfectly legal in the
8157 SSL/TLS protocol it isn't a "bug" option and is on by default. See
8158 the bugs/SSLv3 entry for more info.
8161 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
8164 *) Add -rand argument to smime and pkcs12 applications and read/write
8168 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
8171 *) Add command line password options to the remaining applications.
8174 *) Bug fix for BN_div_recp() for numerators with an even number of
8178 *) More tests in bntest.c, and changed test_bn output.
8181 *) ./config recognizes MacOS X now.
8184 *) Bug fix for BN_div() when the first words of num and divsor are
8185 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
8188 *) Add support for various broken PKCS#8 formats, and command line
8189 options to produce them.
8192 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
8193 get temporary BIGNUMs from a BN_CTX.
8196 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
8200 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
8201 include a #define from the old name to the new. The original intent
8202 was that statically linked binaries could for example just call
8203 SSLeay_add_all_ciphers() to just add ciphers to the table and not
8204 link with digests. This never worked becayse SSLeay_add_all_digests()
8205 and SSLeay_add_all_ciphers() were in the same source file so calling
8206 one would link with the other. They are now in separate source files.
8209 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
8212 *) Use a less unusual form of the Miller-Rabin primality test (it used
8213 a binary algorithm for exponentiation integrated into the Miller-Rabin
8214 loop, our standard modexp algorithms are faster).
8217 *) Support for the EBCDIC character set completed.
8218 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
8220 *) Source code cleanups: use const where appropriate, eliminate casts,
8221 use void * instead of char * in lhash.
8224 *) Bugfix: ssl3_send_server_key_exchange was not restartable
8225 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
8226 this the server could overwrite ephemeral keys that the client
8230 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
8231 using 50 iterations of the Rabin-Miller test.
8233 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
8234 iterations of the Rabin-Miller test as required by the appendix
8235 to FIPS PUB 186[-1]) instead of DSA_is_prime.
8236 As BN_is_prime_fasttest includes trial division, DSA parameter
8237 generation becomes much faster.
8239 This implies a change for the callback functions in DSA_is_prime
8240 and DSA_generate_parameters: The callback function is called once
8241 for each positive witness in the Rabin-Miller test, not just
8242 occasionally in the inner loop; and the parameters to the
8243 callback function now provide an iteration count for the outer
8244 loop rather than for the current invocation of the inner loop.
8245 DSA_generate_parameters additionally can call the callback
8246 function with an 'iteration count' of -1, meaning that a
8247 candidate has passed the trial division test (when q is generated
8248 from an application-provided seed, trial division is skipped).
8251 *) New function BN_is_prime_fasttest that optionally does trial
8252 division before starting the Rabin-Miller test and has
8253 an additional BN_CTX * argument (whereas BN_is_prime always
8254 has to allocate at least one BN_CTX).
8255 'callback(1, -1, cb_arg)' is called when a number has passed the
8256 trial division stage.
8259 *) Fix for bug in CRL encoding. The validity dates weren't being handled
8263 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
8266 *) New function BN_pseudo_rand().
8269 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
8270 bignum version of BN_from_montgomery() with the working code from
8271 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
8275 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
8276 made it impossible to use the same SSL_SESSION data structure in
8277 SSL2 clients in multiple threads.
8280 *) The return value of RAND_load_file() no longer counts bytes obtained
8281 by stat(). RAND_load_file(..., -1) is new and uses the complete file
8282 to seed the PRNG (previously an explicit byte count was required).
8283 [Ulf Möller, Bodo Möller]
8285 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
8286 used (char *) instead of (void *) and had casts all over the place.
8289 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
8292 *) Retain source code compatibility for BN_prime_checks macro:
8293 BN_is_prime(..., BN_prime_checks, ...) now uses
8294 BN_prime_checks_for_size to determine the appropriate number of
8295 Rabin-Miller iterations.
8298 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
8299 DH_CHECK_P_NOT_SAFE_PRIME.
8300 (Check if this is true? OpenPGP calls them "strong".)
8303 *) Merge the functionality of "dh" and "gendh" programs into a new program
8304 "dhparam". The old programs are retained for now but will handle DH keys
8305 (instead of parameters) in future.
8308 *) Make the ciphers, s_server and s_client programs check the return values
8309 when a new cipher list is set.
8312 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
8313 ciphers. Before when the 56bit ciphers were enabled the sorting was
8316 The syntax for the cipher sorting has been extended to support sorting by
8317 cipher-strength (using the strength_bits hard coded in the tables).
8318 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
8320 Fix a bug in the cipher-command parser: when supplying a cipher command
8321 string with an "undefined" symbol (neither command nor alphanumeric
8322 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
8323 an error is flagged.
8325 Due to the strength-sorting extension, the code of the
8326 ssl_create_cipher_list() function was completely rearranged. I hope that
8327 the readability was also increased :-)
8328 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8330 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
8331 for the first serial number and places 2 in the serial number file. This
8332 avoids problems when the root CA is created with serial number zero and
8333 the first user certificate has the same issuer name and serial number
8337 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
8338 the new code. Add documentation for this stuff.
8341 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
8342 X509_*() to X509at_*() on the grounds that they don't handle X509
8343 structures and behave in an analagous way to the X509v3 functions:
8344 they shouldn't be called directly but wrapper functions should be used
8347 So we also now have some wrapper functions that call the X509at functions
8348 when passed certificate requests. (TO DO: similar things can be done with
8349 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
8350 things. Some of these need some d2i or i2d and print functionality
8351 because they handle more complex structures.)
8354 *) Add missing #ifndefs that caused missing symbols when building libssl
8355 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
8356 NO_RSA in ssl/s2*.c.
8357 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
8359 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
8360 has a return value which indicates the quality of the random data
8361 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
8362 error queue. New function RAND_pseudo_bytes() generates output that is
8363 guaranteed to be unique but not unpredictable. RAND_add is like
8364 RAND_seed, but takes an extra argument for an entropy estimate
8365 (RAND_seed always assumes full entropy).
8368 *) Do more iterations of Rabin-Miller probable prime test (specifically,
8369 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
8370 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
8371 in crypto/bn/bn_prime.c for the complete table). This guarantees a
8372 false-positive rate of at most 2^-80 for random input.
8375 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
8378 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
8379 in the 0.9.5 release), this returns the chain
8380 from an X509_CTX structure with a dup of the stack and all
8381 the X509 reference counts upped: so the stack will exist
8382 after X509_CTX_cleanup() has been called. Modify pkcs12.c
8385 Also make SSL_SESSION_print() print out the verify return
8389 *) Add manpage for the pkcs12 command. Also change the default
8390 behaviour so MAC iteration counts are used unless the new
8391 -nomaciter option is used. This improves file security and
8392 only older versions of MSIE (4.0 for example) need it.
8395 *) Honor the no-xxx Configure options when creating .DEF files.
8398 *) Add PKCS#10 attributes to field table: challengePassword,
8399 unstructuredName and unstructuredAddress. These are taken from
8400 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
8401 international characters are used.
8403 More changes to X509_ATTRIBUTE code: allow the setting of types
8404 based on strings. Remove the 'loc' parameter when adding
8405 attributes because these will be a SET OF encoding which is sorted
8409 *) Initial changes to the 'req' utility to allow request generation
8410 automation. This will allow an application to just generate a template
8411 file containing all the field values and have req construct the
8414 Initial support for X509_ATTRIBUTE handling. Stacks of these are
8415 used all over the place including certificate requests and PKCS#7
8416 structures. They are currently handled manually where necessary with
8417 some primitive wrappers for PKCS#7. The new functions behave in a
8418 manner analogous to the X509 extension functions: they allow
8419 attributes to be looked up by NID and added.
8421 Later something similar to the X509V3 code would be desirable to
8422 automatically handle the encoding, decoding and printing of the
8423 more complex types. The string types like challengePassword can
8424 be handled by the string table functions.
8426 Also modified the multi byte string table handling. Now there is
8427 a 'global mask' which masks out certain types. The table itself
8428 can use the flag STABLE_NO_MASK to ignore the mask setting: this
8429 is useful when for example there is only one permissible type
8430 (as in countryName) and using the mask might result in no valid
8434 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
8435 SSL_get_peer_finished to allow applications to obtain the latest
8436 Finished messages sent to the peer or expected from the peer,
8437 respectively. (SSL_get_peer_finished is usually the Finished message
8438 actually received from the peer, otherwise the protocol will be aborted.)
8440 As the Finished message are message digests of the complete handshake
8441 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
8442 be used for external authentication procedures when the authentication
8443 provided by SSL/TLS is not desired or is not enough.
8446 *) Enhanced support for Alpha Linux is added. Now ./config checks if
8447 the host supports BWX extension and if Compaq C is present on the
8448 $PATH. Just exploiting of the BWX extension results in 20-30%
8449 performance kick for some algorithms, e.g. DES and RC4 to mention
8450 a couple. Compaq C in turn generates ~20% faster code for MD5 and
8454 *) Add support for MS "fast SGC". This is arguably a violation of the
8455 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
8456 weak crypto and after checking the certificate is SGC a second one
8457 with strong crypto. MS SGC stops the first handshake after receiving
8458 the server certificate message and sends a second client hello. Since
8459 a server will typically do all the time consuming operations before
8460 expecting any further messages from the client (server key exchange
8461 is the most expensive) there is little difference between the two.
8463 To get OpenSSL to support MS SGC we have to permit a second client
8464 hello message after we have sent server done. In addition we have to
8465 reset the MAC if we do get this second client hello.
8468 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
8469 if a DER encoded private key is RSA or DSA traditional format. Changed
8470 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
8471 format DER encoded private key. Newer code should use PKCS#8 format which
8472 has the key type encoded in the ASN1 structure. Added DER private key
8473 support to pkcs8 application.
8476 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
8477 ciphersuites has been selected (as required by the SSL 3/TLS 1
8478 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
8479 is set, we interpret this as a request to violate the specification
8480 (the worst that can happen is a handshake failure, and 'correct'
8481 behaviour would result in a handshake failure anyway).
8484 *) In SSL_CTX_add_session, take into account that there might be multiple
8485 SSL_SESSION structures with the same session ID (e.g. when two threads
8486 concurrently obtain them from an external cache).
8487 The internal cache can handle only one SSL_SESSION with a given ID,
8488 so if there's a conflict, we now throw out the old one to achieve
8492 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
8493 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
8494 some routines that use cipher OIDs: some ciphers do not have OIDs
8495 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
8499 *) Simplify the trust setting structure and code. Now we just have
8500 two sequences of OIDs for trusted and rejected settings. These will
8501 typically have values the same as the extended key usage extension
8502 and any application specific purposes.
8504 The trust checking code now has a default behaviour: it will just
8505 check for an object with the same NID as the passed id. Functions can
8506 be provided to override either the default behaviour or the behaviour
8507 for a given id. SSL client, server and email already have functions
8508 in place for compatibility: they check the NID and also return "trusted"
8509 if the certificate is self signed.
8512 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
8513 traditional format into an EVP_PKEY structure.
8516 *) Add a password callback function PEM_cb() which either prompts for
8517 a password if usr_data is NULL or otherwise assumes it is a null
8518 terminated password. Allow passwords to be passed on command line
8519 environment or config files in a few more utilities.
8522 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
8523 keys. Add some short names for PKCS#8 PBE algorithms and allow them
8524 to be specified on the command line for the pkcs8 and pkcs12 utilities.
8525 Update documentation.
8528 *) Support for ASN1 "NULL" type. This could be handled before by using
8529 ASN1_TYPE but there wasn't any function that would try to read a NULL
8530 and produce an error if it couldn't. For compatibility we also have
8531 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
8532 don't allocate anything because they don't need to.
8535 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
8537 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
8539 *) Rebuild of the memory allocation routines used by OpenSSL code and
8540 possibly others as well. The purpose is to make an interface that
8541 provide hooks so anyone can build a separate set of allocation and
8542 deallocation routines to be used by OpenSSL, for example memory
8543 pool implementations, or something else, which was previously hard
8544 since Malloc(), Realloc() and Free() were defined as macros having
8545 the values malloc, realloc and free, respectively (except for Win32
8546 compilations). The same is provided for memory debugging code.
8547 OpenSSL already comes with functionality to find memory leaks, but
8548 this gives people a chance to debug other memory problems.
8550 With these changes, a new set of functions and macros have appeared:
8552 CRYPTO_set_mem_debug_functions() [F]
8553 CRYPTO_get_mem_debug_functions() [F]
8554 CRYPTO_dbg_set_options() [F]
8555 CRYPTO_dbg_get_options() [F]
8556 CRYPTO_malloc_debug_init() [M]
8558 The memory debug functions are NULL by default, unless the library
8559 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
8560 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
8561 gives the standard debugging functions that come with OpenSSL) or
8562 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
8563 provided by the library user) must be used. When the standard
8564 debugging functions are used, CRYPTO_dbg_set_options can be used to
8565 request additional information:
8566 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
8567 the CRYPTO_MDEBUG_xxx macro when compiling the library.
8569 Also, things like CRYPTO_set_mem_functions will always give the
8570 expected result (the new set of functions is used for allocation
8571 and deallocation) at all times, regardless of platform and compiler
8574 To finish it up, some functions that were never use in any other
8575 way than through macros have a new API and new semantic:
8578 CRYPTO_dbg_realloc()
8581 All macros of value have retained their old syntax.
8582 [Richard Levitte and Bodo Moeller]
8584 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
8585 ordering of SMIMECapabilities wasn't in "strength order" and there
8586 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
8590 *) Some ASN1 types with illegal zero length encoding (INTEGER,
8591 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
8592 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
8594 *) Merge in my S/MIME library for OpenSSL. This provides a simple
8595 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
8596 functionality to handle multipart/signed properly) and a utility
8597 called 'smime' to call all this stuff. This is based on code I
8598 originally wrote for Celo who have kindly allowed it to be
8599 included in OpenSSL.
8602 *) Add variants des_set_key_checked and des_set_key_unchecked of
8603 des_set_key (aka des_key_sched). Global variable des_check_key
8604 decides which of these is called by des_set_key; this way
8605 des_check_key behaves as it always did, but applications and
8606 the library itself, which was buggy for des_check_key == 1,
8607 have a cleaner way to pick the version they need.
8610 *) New function PKCS12_newpass() which changes the password of a
8614 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
8615 dynamic mix. In both cases the ids can be used as an index into the
8616 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
8617 functions so they accept a list of the field values and the
8618 application doesn't need to directly manipulate the X509_TRUST
8622 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
8626 *) Modify the way the V3 extension code looks up extensions. This now
8627 works in a similar way to the object code: we have some "standard"
8628 extensions in a static table which is searched with OBJ_bsearch()
8629 and the application can add dynamic ones if needed. The file
8630 crypto/x509v3/ext_dat.h now has the info: this file needs to be
8631 updated whenever a new extension is added to the core code and kept
8632 in ext_nid order. There is a simple program 'tabtest.c' which checks
8633 this. New extensions are not added too often so this file can readily
8634 be maintained manually.
8636 There are two big advantages in doing things this way. The extensions
8637 can be looked up immediately and no longer need to be "added" using
8638 X509V3_add_standard_extensions(): this function now does nothing.
8639 [Side note: I get *lots* of email saying the extension code doesn't
8640 work because people forget to call this function]
8641 Also no dynamic allocation is done unless new extensions are added:
8642 so if we don't add custom extensions there is no need to call
8643 X509V3_EXT_cleanup().
8646 *) Modify enc utility's salting as follows: make salting the default. Add a
8647 magic header, so unsalted files fail gracefully instead of just decrypting
8648 to garbage. This is because not salting is a big security hole, so people
8649 should be discouraged from doing it.
8652 *) Fixes and enhancements to the 'x509' utility. It allowed a message
8653 digest to be passed on the command line but it only used this
8654 parameter when signing a certificate. Modified so all relevant
8655 operations are affected by the digest parameter including the
8656 -fingerprint and -x509toreq options. Also -x509toreq choked if a
8657 DSA key was used because it didn't fix the digest.
8660 *) Initial certificate chain verify code. Currently tests the untrusted
8661 certificates for consistency with the verify purpose (which is set
8662 when the X509_STORE_CTX structure is set up) and checks the pathlength.
8664 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
8665 this is because it will reject chains with invalid extensions whereas
8666 every previous version of OpenSSL and SSLeay made no checks at all.
8668 Trust code: checks the root CA for the relevant trust settings. Trust
8669 settings have an initial value consistent with the verify purpose: e.g.
8670 if the verify purpose is for SSL client use it expects the CA to be
8671 trusted for SSL client use. However the default value can be changed to
8672 permit custom trust settings: one example of this would be to only trust
8673 certificates from a specific "secure" set of CAs.
8675 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
8676 which should be used for version portability: especially since the
8677 verify structure is likely to change more often now.
8679 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
8680 to set them. If not set then assume SSL clients will verify SSL servers
8683 Two new options to the verify program: -untrusted allows a set of
8684 untrusted certificates to be passed in and -purpose which sets the
8685 intended purpose of the certificate. If a purpose is set then the
8686 new chain verify code is used to check extension consistency.
8689 *) Support for the authority information access extension.
8692 *) Modify RSA and DSA PEM read routines to transparently handle
8693 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
8694 public keys in a format compatible with certificate
8695 SubjectPublicKeyInfo structures. Unfortunately there were already
8696 functions called *_PublicKey_* which used various odd formats so
8697 these are retained for compatibility: however the DSA variants were
8698 never in a public release so they have been deleted. Changed dsa/rsa
8699 utilities to handle the new format: note no releases ever handled public
8700 keys so we should be OK.
8702 The primary motivation for this change is to avoid the same fiasco
8703 that dogs private keys: there are several incompatible private key
8704 formats some of which are standard and some OpenSSL specific and
8705 require various evil hacks to allow partial transparent handling and
8706 even then it doesn't work with DER formats. Given the option anything
8707 other than PKCS#8 should be dumped: but the other formats have to
8708 stay in the name of compatibility.
8710 With public keys and the benefit of hindsight one standard format
8711 is used which works with EVP_PKEY, RSA or DSA structures: though
8712 it clearly returns an error if you try to read the wrong kind of key.
8714 Added a -pubkey option to the 'x509' utility to output the public key.
8715 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
8716 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
8717 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
8718 that do the same as the EVP_PKEY_assign_*() except they up the
8719 reference count of the added key (they don't "swallow" the
8723 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
8724 CRLs would fail if the file contained no certificates or no CRLs:
8725 added a new function to read in both types and return the number
8726 read: this means that if none are read it will be an error. The
8727 DER versions of the certificate and CRL reader would always fail
8728 because it isn't possible to mix certificates and CRLs in DER format
8729 without choking one or the other routine. Changed this to just read
8730 a certificate: this is the best we can do. Also modified the code
8731 in apps/verify.c to take notice of return codes: it was previously
8732 attempting to read in certificates from NULL pointers and ignoring
8733 any errors: this is one reason why the cert and CRL reader seemed
8734 to work. It doesn't check return codes from the default certificate
8735 routines: these may well fail if the certificates aren't installed.
8738 *) Code to support otherName option in GeneralName.
8741 *) First update to verify code. Change the verify utility
8742 so it warns if it is passed a self signed certificate:
8743 for consistency with the normal behaviour. X509_verify
8744 has been modified to it will now verify a self signed
8745 certificate if *exactly* the same certificate appears
8746 in the store: it was previously impossible to trust a
8747 single self signed certificate. This means that:
8748 openssl verify ss.pem
8749 now gives a warning about a self signed certificate but
8750 openssl verify -CAfile ss.pem ss.pem
8754 *) For servers, store verify_result in SSL_SESSION data structure
8755 (and add it to external session representation).
8756 This is needed when client certificate verifications fails,
8757 but an application-provided verification callback (set by
8758 SSL_CTX_set_cert_verify_callback) allows accepting the session
8759 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
8760 but returns 1): When the session is reused, we have to set
8761 ssl->verify_result to the appropriate error code to avoid
8763 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
8765 *) Fix a bug in the new PKCS#7 code: it didn't consider the
8766 case in PKCS7_dataInit() where the signed PKCS7 structure
8767 didn't contain any existing data because it was being created.
8768 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
8770 *) Add a salt to the key derivation routines in enc.c. This
8771 forms the first 8 bytes of the encrypted file. Also add a
8772 -S option to allow a salt to be input on the command line.
8775 *) New function X509_cmp(). Oddly enough there wasn't a function
8776 to compare two certificates. We do this by working out the SHA1
8777 hash and comparing that. X509_cmp() will be needed by the trust
8781 *) SSL_get1_session() is like SSL_get_session(), but increments
8782 the reference count in the SSL_SESSION returned.
8783 [Geoff Thorpe <geoff@eu.c2.net>]
8785 *) Fix for 'req': it was adding a null to request attributes.
8786 Also change the X509_LOOKUP and X509_INFO code to handle
8787 certificate auxiliary information.
8790 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8794 *) Add the possibility to add extra information to the memory leak
8795 detecting output, to form tracebacks, showing from where each
8796 allocation was originated: CRYPTO_push_info("constant string") adds
8797 the string plus current file name and line number to a per-thread
8798 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8799 is like calling CYRPTO_pop_info() until the stack is empty.
8800 Also updated memory leak detection code to be multi-thread-safe.
8803 *) Add options -text and -noout to pkcs7 utility and delete the
8804 encryption options which never did anything. Update docs.
8807 *) Add options to some of the utilities to allow the pass phrase
8808 to be included on either the command line (not recommended on
8809 OSes like Unix) or read from the environment. Update the
8810 manpages and fix a few bugs.
8813 *) Add a few manpages for some of the openssl commands.
8816 *) Fix the -revoke option in ca. It was freeing up memory twice,
8817 leaking and not finding already revoked certificates.
8820 *) Extensive changes to support certificate auxiliary information.
8821 This involves the use of X509_CERT_AUX structure and X509_AUX
8822 functions. An X509_AUX function such as PEM_read_X509_AUX()
8823 can still read in a certificate file in the usual way but it
8824 will also read in any additional "auxiliary information". By
8825 doing things this way a fair degree of compatibility can be
8826 retained: existing certificates can have this information added
8827 using the new 'x509' options.
8829 Current auxiliary information includes an "alias" and some trust
8830 settings. The trust settings will ultimately be used in enhanced
8831 certificate chain verification routines: currently a certificate
8832 can only be trusted if it is self signed and then it is trusted
8836 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8837 The problem was that one of the replacement routines had not been working
8838 since SSLeay releases. For now the offending routine has been replaced
8839 with non-optimised assembler. Even so, this now gives around 95%
8840 performance improvement for 1024 bit RSA signs.
8843 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8844 handling. Most clients have the effective key size in bits equal to
8845 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8846 A few however don't do this and instead use the size of the decrypted key
8847 to determine the RC2 key length and the AlgorithmIdentifier to determine
8848 the effective key length. In this case the effective key length can still
8849 be 40 bits but the key length can be 168 bits for example. This is fixed
8850 by manually forcing an RC2 key into the EVP_PKEY structure because the
8851 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8852 the key length and effective key length are equal.
8855 *) Add a bunch of functions that should simplify the creation of
8856 X509_NAME structures. Now you should be able to do:
8857 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8858 and have it automatically work out the correct field type and fill in
8859 the structures. The more adventurous can try:
8860 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8861 and it will (hopefully) work out the correct multibyte encoding.
8864 *) Change the 'req' utility to use the new field handling and multibyte
8865 copy routines. Before the DN field creation was handled in an ad hoc
8866 way in req, ca, and x509 which was rather broken and didn't support
8867 BMPStrings or UTF8Strings. Since some software doesn't implement
8868 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8869 using the dirstring_type option. See the new comment in the default
8870 openssl.cnf for more info.
8873 *) Make crypto/rand/md_rand.c more robust:
8874 - Assure unique random numbers after fork().
8875 - Make sure that concurrent threads access the global counter and
8876 md serializably so that we never lose entropy in them
8877 or use exactly the same state in multiple threads.
8878 Access to the large state is not always serializable because
8879 the additional locking could be a performance killer, and
8880 md should be large enough anyway.
8883 *) New file apps/app_rand.c with commonly needed functionality
8884 for handling the random seed file.
8886 Use the random seed file in some applications that previously did not:
8888 dsaparam -genkey (which also ignored its '-rand' option),
8891 x509 (when signing).
8892 Except on systems with /dev/urandom, it is crucial to have a random
8893 seed file at least for key creation, DSA signing, and for DH exchanges;
8894 for RSA signatures we could do without one.
8896 gendh and gendsa (unlike genrsa) used to read only the first byte
8897 of each file listed in the '-rand' option. The function as previously
8898 found in genrsa is now in app_rand.c and is used by all programs
8899 that support '-rand'.
8902 *) In RAND_write_file, use mode 0600 for creating files;
8903 don't just chmod when it may be too late.
8906 *) Report an error from X509_STORE_load_locations
8907 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8910 *) New function ASN1_mbstring_copy() this copies a string in either
8911 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8912 into an ASN1_STRING type. A mask of permissible types is passed
8913 and it chooses the "minimal" type to use or an error if not type
8917 *) Add function equivalents to the various macros in asn1.h. The old
8918 macros are retained with an M_ prefix. Code inside the library can
8919 use the M_ macros. External code (including the openssl utility)
8920 should *NOT* in order to be "shared library friendly".
8923 *) Add various functions that can check a certificate's extensions
8924 to see if it usable for various purposes such as SSL client,
8925 server or S/MIME and CAs of these types. This is currently
8926 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8927 verification. Also added a -purpose flag to x509 utility to
8928 print out all the purposes.
8931 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8935 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8936 for, obtain and decode and extension and obtain its critical flag.
8937 This allows all the necessary extension code to be handled in a
8938 single function call.
8941 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8942 platforms. See crypto/rc4/rc4_enc.c for further details.
8945 *) New -noout option to asn1parse. This causes no output to be produced
8946 its main use is when combined with -strparse and -out to extract data
8947 from a file (which may not be in ASN.1 format).
8950 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8951 when producing the local key id.
8952 [Richard Levitte <levitte@stacken.kth.se>]
8954 *) New option -dhparam in s_server. This allows a DH parameter file to be
8955 stated explicitly. If it is not stated then it tries the first server
8956 certificate file. The previous behaviour hard coded the filename
8960 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8961 a public key to be input or output. For example:
8962 openssl rsa -in key.pem -pubout -out pubkey.pem
8963 Also added necessary DSA public key functions to handle this.
8966 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8967 in the message. This was handled by allowing
8968 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8969 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8971 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8972 to the end of the strings whereas this didn't. This would cause problems
8973 if strings read with d2i_ASN1_bytes() were later modified.
8974 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8976 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8977 data and it contains EOF it will end up returning an error. This is
8978 caused by input 46 bytes long. The cause is due to the way base64
8979 BIOs find the start of base64 encoded data. They do this by trying a
8980 trial decode on each line until they find one that works. When they
8981 do a flag is set and it starts again knowing it can pass all the
8982 data directly through the decoder. Unfortunately it doesn't reset
8983 the context it uses. This means that if EOF is reached an attempt
8984 is made to pass two EOFs through the context and this causes the
8985 resulting error. This can also cause other problems as well. As is
8986 usual with these problems it takes *ages* to find and the fix is
8987 trivial: move one line.
8988 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8990 *) Ugly workaround to get s_client and s_server working under Windows. The
8991 old code wouldn't work because it needed to select() on sockets and the
8992 tty (for keypresses and to see if data could be written). Win32 only
8993 supports select() on sockets so we select() with a 1s timeout on the
8994 sockets and then see if any characters are waiting to be read, if none
8995 are present then we retry, we also assume we can always write data to
8996 the tty. This isn't nice because the code then blocks until we've
8997 received a complete line of data and it is effectively polling the
8998 keyboard at 1s intervals: however it's quite a bit better than not
8999 working at all :-) A dedicated Windows application might handle this
9000 with an event loop for example.
9003 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
9004 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
9005 will be called when RSA_sign() and RSA_verify() are used. This is useful
9006 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
9007 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
9008 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
9009 This necessitated the support of an extra signature type NID_md5_sha1
9010 for SSL signatures and modifications to the SSL library to use it instead
9011 of calling RSA_public_decrypt() and RSA_private_encrypt().
9014 *) Add new -verify -CAfile and -CApath options to the crl program, these
9015 will lookup a CRL issuers certificate and verify the signature in a
9016 similar way to the verify program. Tidy up the crl program so it
9017 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
9018 less strict. It will now permit CRL extensions even if it is not
9019 a V2 CRL: this will allow it to tolerate some broken CRLs.
9022 *) Initialize all non-automatic variables each time one of the openssl
9023 sub-programs is started (this is necessary as they may be started
9024 multiple times from the "OpenSSL>" prompt).
9025 [Lennart Bang, Bodo Moeller]
9027 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
9028 removing all other RSA functionality (this is what NO_RSA does). This
9029 is so (for example) those in the US can disable those operations covered
9030 by the RSA patent while allowing storage and parsing of RSA keys and RSA
9034 *) Non-copying interface to BIO pairs.
9035 (still largely untested)
9038 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
9039 ASCII string. This was handled independently in various places before.
9042 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
9043 UTF8 strings a character at a time.
9046 *) Use client_version from client hello to select the protocol
9047 (s23_srvr.c) and for RSA client key exchange verification
9048 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
9051 *) Add various utility functions to handle SPKACs, these were previously
9052 handled by poking round in the structure internals. Added new function
9053 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
9054 print, verify and generate SPKACs. Based on an original idea from
9055 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
9058 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
9061 *) Allow the config file extension section to be overwritten on the
9062 command line. Based on an original idea from Massimiliano Pala
9063 <madwolf@comune.modena.it>. The new option is called -extensions
9064 and can be applied to ca, req and x509. Also -reqexts to override
9065 the request extensions in req and -crlexts to override the crl extensions
9069 *) Add new feature to the SPKAC handling in ca. Now you can include
9070 the same field multiple times by preceding it by "XXXX." for example:
9073 this is the same syntax as used in the req config file.
9076 *) Allow certificate extensions to be added to certificate requests. These
9077 are specified in a 'req_extensions' option of the req section of the
9078 config file. They can be printed out with the -text option to req but
9079 are otherwise ignored at present.
9082 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
9083 data read consists of only the final block it would not decrypted because
9084 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
9085 A misplaced 'break' also meant the decrypted final block might not be
9086 copied until the next read.
9089 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
9090 a few extra parameters to the DH structure: these will be useful if
9091 for example we want the value of 'q' or implement X9.42 DH.
9094 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
9095 provides hooks that allow the default DSA functions or functions on a
9096 "per key" basis to be replaced. This allows hardware acceleration and
9097 hardware key storage to be handled without major modification to the
9098 library. Also added low level modexp hooks and CRYPTO_EX structure and
9099 associated functions.
9102 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
9103 as "read only": it can't be written to and the buffer it points to will
9104 not be freed. Reading from a read only BIO is much more efficient than
9105 a normal memory BIO. This was added because there are several times when
9106 an area of memory needs to be read from a BIO. The previous method was
9107 to create a memory BIO and write the data to it, this results in two
9108 copies of the data and an O(n^2) reading algorithm. There is a new
9109 function BIO_new_mem_buf() which creates a read only memory BIO from
9110 an area of memory. Also modified the PKCS#7 routines to use read only
9114 *) Bugfix: ssl23_get_client_hello did not work properly when called in
9115 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
9116 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
9117 but a retry condition occured while trying to read the rest.
9120 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
9121 NID_pkcs7_encrypted by default: this was wrong since this should almost
9122 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
9123 the encrypted data type: this is a more sensible place to put it and it
9124 allows the PKCS#12 code to be tidied up that duplicated this
9128 *) Changed obj_dat.pl script so it takes its input and output files on
9129 the command line. This should avoid shell escape redirection problems
9133 *) Initial support for certificate extension requests, these are included
9134 in things like Xenroll certificate requests. Included functions to allow
9135 extensions to be obtained and added.
9138 *) -crlf option to s_client and s_server for sending newlines as
9139 CRLF (as required by many protocols).
9142 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
9144 *) Install libRSAglue.a when OpenSSL is built with RSAref.
9145 [Ralf S. Engelschall]
9147 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
9148 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
9150 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
9154 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
9155 DH parameters/keys (q is lost during that conversion, but the resulting
9156 DH parameters contain its length).
9158 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
9159 much faster than DH_generate_parameters (which creates parameters
9160 where p = 2*q + 1), and also the smaller q makes DH computations
9161 much more efficient (160-bit exponentiation instead of 1024-bit
9162 exponentiation); so this provides a convenient way to support DHE
9163 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
9164 utter importance to use
9165 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9167 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9168 when such DH parameters are used, because otherwise small subgroup
9169 attacks may become possible!
9172 *) Avoid memory leak in i2d_DHparams.
9175 *) Allow the -k option to be used more than once in the enc program:
9176 this allows the same encrypted message to be read by multiple recipients.
9179 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
9180 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
9181 it will always use the numerical form of the OID, even if it has a short
9185 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
9186 method only got called if p,q,dmp1,dmq1,iqmp components were present,
9187 otherwise bn_mod_exp was called. In the case of hardware keys for example
9188 no private key components need be present and it might store extra data
9189 in the RSA structure, which cannot be accessed from bn_mod_exp.
9190 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
9191 private key operations.
9194 *) Added support for SPARC Linux.
9197 *) pem_password_cb function type incompatibly changed from
9198 typedef int pem_password_cb(char *buf, int size, int rwflag);
9200 ....(char *buf, int size, int rwflag, void *userdata);
9201 so that applications can pass data to their callbacks:
9202 The PEM[_ASN1]_{read,write}... functions and macros now take an
9203 additional void * argument, which is just handed through whenever
9204 the password callback is called.
9205 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
9207 New function SSL_CTX_set_default_passwd_cb_userdata.
9209 Compatibility note: As many C implementations push function arguments
9210 onto the stack in reverse order, the new library version is likely to
9211 interoperate with programs that have been compiled with the old
9212 pem_password_cb definition (PEM_whatever takes some data that
9213 happens to be on the stack as its last argument, and the callback
9214 just ignores this garbage); but there is no guarantee whatsoever that
9217 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
9218 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
9219 problems not only on Windows, but also on some Unix platforms.
9220 To avoid problematic command lines, these definitions are now in an
9221 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
9222 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
9225 *) MIPS III/IV assembler module is reimplemented.
9228 *) More DES library cleanups: remove references to srand/rand and
9229 delete an unused file.
9232 *) Add support for the the free Netwide assembler (NASM) under Win32,
9233 since not many people have MASM (ml) and it can be hard to obtain.
9234 This is currently experimental but it seems to work OK and pass all
9235 the tests. Check out INSTALL.W32 for info.
9238 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
9239 without temporary keys kept an extra copy of the server key,
9240 and connections with temporary keys did not free everything in case
9244 *) New function RSA_check_key and new openssl rsa option -check
9245 for verifying the consistency of RSA keys.
9246 [Ulf Moeller, Bodo Moeller]
9248 *) Various changes to make Win32 compile work:
9249 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
9250 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
9251 comparison" warnings.
9252 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
9255 *) Add a debugging option to PKCS#5 v2 key generation function: when
9256 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
9257 derived keys are printed to stderr.
9260 *) Copy the flags in ASN1_STRING_dup().
9261 [Roman E. Pavlov <pre@mo.msk.ru>]
9263 *) The x509 application mishandled signing requests containing DSA
9264 keys when the signing key was also DSA and the parameters didn't match.
9266 It was supposed to omit the parameters when they matched the signing key:
9267 the verifying software was then supposed to automatically use the CA's
9268 parameters if they were absent from the end user certificate.
9270 Omitting parameters is no longer recommended. The test was also
9271 the wrong way round! This was probably due to unusual behaviour in
9272 EVP_cmp_parameters() which returns 1 if the parameters match.
9273 This meant that parameters were omitted when they *didn't* match and
9274 the certificate was useless. Certificates signed with 'ca' didn't have
9276 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
9278 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
9279 The interface is as follows:
9280 Applications can use
9281 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
9282 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
9283 "off" is now the default.
9284 The library internally uses
9285 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
9286 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
9287 to disable memory-checking temporarily.
9289 Some inconsistent states that previously were possible (and were
9290 even the default) are now avoided.
9292 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
9293 with each memory chunk allocated; this is occasionally more helpful
9294 than just having a counter.
9296 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
9298 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
9302 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
9303 which largely parallels "options", but is for changing API behaviour,
9304 whereas "options" are about protocol behaviour.
9305 Initial "mode" flags are:
9307 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
9308 a single record has been written.
9309 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
9310 retries use the same buffer location.
9311 (But all of the contents must be
9315 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
9318 *) Fix problems with no-hmac etc.
9319 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
9321 *) New functions RSA_get_default_method(), RSA_set_method() and
9322 RSA_get_method(). These allows replacement of RSA_METHODs without having
9323 to mess around with the internals of an RSA structure.
9326 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
9327 Also really enable memory leak checks in openssl.c and in some
9329 [Chad C. Mulligan, Bodo Moeller]
9331 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
9332 up the length of negative integers. This has now been simplified to just
9333 store the length when it is first determined and use it later, rather
9334 than trying to keep track of where data is copied and updating it to
9336 [Steve Henson, reported by Brien Wheeler
9337 <bwheeler@authentica-security.com>]
9339 *) Add a new function PKCS7_signatureVerify. This allows the verification
9340 of a PKCS#7 signature but with the signing certificate passed to the
9341 function itself. This contrasts with PKCS7_dataVerify which assumes the
9342 certificate is present in the PKCS#7 structure. This isn't always the
9343 case: certificates can be omitted from a PKCS#7 structure and be
9344 distributed by "out of band" means (such as a certificate database).
9347 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
9348 function prototypes in pem.h, also change util/mkdef.pl to add the
9349 necessary function names.
9352 *) mk1mf.pl (used by Windows builds) did not properly read the
9353 options set by Configure in the top level Makefile, and Configure
9354 was not even able to write more than one option correctly.
9355 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
9358 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
9359 file to be loaded from a BIO or FILE pointer. The BIO version will
9360 for example allow memory BIOs to contain config info.
9363 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
9364 Whoever hopes to achieve shared-library compatibility across versions
9365 must use this, not the compile-time macro.
9366 (Exercise 0.9.4: Which is the minimum library version required by
9368 Note: All this applies only to multi-threaded programs, others don't
9372 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
9373 through a BIO pair triggered the default case, i.e.
9374 SSLerr(...,SSL_R_UNKNOWN_STATE).
9377 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
9378 can use the SSL library even if none of the specific BIOs is
9382 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
9383 for the encoded length.
9384 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
9386 *) Add initial documentation of the X509V3 functions.
9389 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
9390 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
9391 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
9392 secure PKCS#8 private key format with a high iteration count.
9395 *) Fix determination of Perl interpreter: A perl or perl5
9396 _directory_ in $PATH was also accepted as the interpreter.
9397 [Ralf S. Engelschall]
9399 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
9400 wrong with it but it was very old and did things like calling
9401 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
9405 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
9406 to use the new extension code.
9409 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
9410 with macros. This should make it easier to change their form, add extra
9411 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
9415 *) Add to configuration table a new entry that can specify an alternative
9416 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
9417 according to Mark Crispin <MRC@Panda.COM>.
9421 *) DES CBC did not update the IV. Weird.
9424 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
9425 Changing the behaviour of the former might break existing programs --
9426 where IV updating is needed, des_ncbc_encrypt can be used.
9429 *) When bntest is run from "make test" it drives bc to check its
9430 calculations, as well as internally checking them. If an internal check
9431 fails, it needs to cause bc to give a non-zero result or make test carries
9432 on without noticing the failure. Fixed.
9435 *) DES library cleanups.
9438 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
9439 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
9440 ciphers. NOTE: although the key derivation function has been verified
9441 against some published test vectors it has not been extensively tested
9442 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
9446 *) Instead of "mkdir -p", which is not fully portable, use new
9447 Perl script "util/mkdir-p.pl".
9450 *) Rewrite the way password based encryption (PBE) is handled. It used to
9451 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
9452 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
9453 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
9454 the 'parameter' field of the AlgorithmIdentifier is passed to the
9455 underlying key generation function so it must do its own ASN1 parsing.
9456 This has also changed the EVP_PBE_CipherInit() function which now has a
9457 'parameter' argument instead of literal salt and iteration count values
9458 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
9461 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
9462 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
9463 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
9464 KEY" because this clashed with PKCS#8 unencrypted string. Since this
9465 value was just used as a "magic string" and not used directly its
9466 value doesn't matter.
9469 *) Introduce some semblance of const correctness to BN. Shame C doesn't
9473 *) "linux-sparc64" configuration (ultrapenguin).
9474 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
9475 "linux-sparc" configuration.
9476 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
9478 *) config now generates no-xxx options for missing ciphers.
9481 *) Support the EBCDIC character set (work in progress).
9482 File ebcdic.c not yet included because it has a different license.
9483 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9485 *) Support BS2000/OSD-POSIX.
9486 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9488 *) Make callbacks for key generation use void * instead of char *.
9491 *) Make S/MIME samples compile (not yet tested).
9494 *) Additional typesafe stacks.
9497 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
9501 Changes between 0.9.3 and 0.9.3a [29 May 1999]
9503 *) New configuration variant "sco5-gcc".
9505 *) Updated some demos.
9506 [Sean O Riordain, Wade Scholine]
9508 *) Add missing BIO_free at exit of pkcs12 application.
9511 *) Fix memory leak in conf.c.
9514 *) Updates for Win32 to assembler version of MD5.
9517 *) Set #! path to perl in apps/der_chop to where we found it
9518 instead of using a fixed path.
9521 *) SHA library changes for irix64-mips4-cc.
9524 *) Improvements for VMS support.
9528 Changes between 0.9.2b and 0.9.3 [24 May 1999]
9530 *) Bignum library bug fix. IRIX 6 passes "make test" now!
9531 This also avoids the problems with SC4.2 and unpatched SC5.
9532 [Andy Polyakov <appro@fy.chalmers.se>]
9534 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
9535 These are required because of the typesafe stack would otherwise break
9536 existing code. If old code used a structure member which used to be STACK
9537 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
9538 sk_num or sk_value it would produce an error because the num, data members
9539 are not present in STACK_OF. Now it just produces a warning. sk_set
9540 replaces the old method of assigning a value to sk_value
9541 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
9542 that does this will no longer work (and should use sk_set instead) but
9543 this could be regarded as a "questionable" behaviour anyway.
9546 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
9547 correctly handle encrypted S/MIME data.
9550 *) Change type of various DES function arguments from des_cblock
9551 (which means, in function argument declarations, pointer to char)
9552 to des_cblock * (meaning pointer to array with 8 char elements),
9553 which allows the compiler to do more typechecking; it was like
9554 that back in SSLeay, but with lots of ugly casts.
9556 Introduce new type const_des_cblock.
9559 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
9560 problems: find RecipientInfo structure that matches recipient certificate
9561 and initialise the ASN1 structures properly based on passed cipher.
9564 *) Belatedly make the BN tests actually check the results.
9567 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
9568 to and from BNs: it was completely broken. New compilation option
9569 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
9570 key elements as negative integers.
9573 *) Reorganize and speed up MD5.
9574 [Andy Polyakov <appro@fy.chalmers.se>]
9577 [Richard Levitte <richard@levitte.org>]
9579 *) New option -out to asn1parse to allow the parsed structure to be
9580 output to a file. This is most useful when combined with the -strparse
9581 option to examine the output of things like OCTET STRINGS.
9584 *) Make SSL library a little more fool-proof by not requiring any longer
9585 that SSL_set_{accept,connect}_state be called before
9586 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
9587 in many applications because usually everything *appeared* to work as
9588 intended anyway -- now it really works as intended).
9591 *) Move openssl.cnf out of lib/.
9594 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
9595 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
9596 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
9597 [Ralf S. Engelschall]
9599 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
9600 handle PKCS#7 enveloped data properly.
9601 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
9603 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
9604 copying pointers. The cert_st handling is changed by this in
9605 various ways (and thus what used to be known as ctx->default_cert
9606 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
9607 any longer when s->cert does not give us what we need).
9608 ssl_cert_instantiate becomes obsolete by this change.
9609 As soon as we've got the new code right (possibly it already is?),
9610 we have solved a couple of bugs of the earlier code where s->cert
9611 was used as if it could not have been shared with other SSL structures.
9613 Note that using the SSL API in certain dirty ways now will result
9614 in different behaviour than observed with earlier library versions:
9615 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
9616 does not influence s as it used to.
9618 In order to clean up things more thoroughly, inside SSL_SESSION
9619 we don't use CERT any longer, but a new structure SESS_CERT
9620 that holds per-session data (if available); currently, this is
9621 the peer's certificate chain and, for clients, the server's certificate
9622 and temporary key. CERT holds only those values that can have
9623 meaningful defaults in an SSL_CTX.
9626 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
9627 from the internal representation. Various PKCS#7 fixes: remove some
9628 evil casts and set the enc_dig_alg field properly based on the signing
9632 *) Allow PKCS#12 password to be set from the command line or the
9633 environment. Let 'ca' get its config file name from the environment
9634 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
9638 *) Allow certificate policies extension to use an IA5STRING for the
9639 organization field. This is contrary to the PKIX definition but
9640 VeriSign uses it and IE5 only recognises this form. Document 'x509'
9644 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
9645 without disallowing inline assembler and the like for non-pedantic builds.
9648 *) Support Borland C++ builder.
9649 [Janez Jere <jj@void.si>, modified by Ulf Möller]
9654 *) SHA-1 cleanups and performance enhancements.
9655 [Andy Polyakov <appro@fy.chalmers.se>]
9657 *) Sparc v8plus assembler for the bignum library.
9658 [Andy Polyakov <appro@fy.chalmers.se>]
9660 *) Accept any -xxx and +xxx compiler options in Configure.
9663 *) Update HPUX configuration.
9666 *) Add missing sk_<type>_unshift() function to safestack.h
9667 [Ralf S. Engelschall]
9669 *) New function SSL_CTX_use_certificate_chain_file that sets the
9670 "extra_cert"s in addition to the certificate. (This makes sense
9671 only for "PEM" format files, as chains as a whole are not
9675 *) Support verify_depth from the SSL API.
9676 x509_vfy.c had what can be considered an off-by-one-error:
9677 Its depth (which was not part of the external interface)
9678 was actually counting the number of certificates in a chain;
9679 now it really counts the depth.
9682 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
9683 instead of X509err, which often resulted in confusing error
9684 messages since the error codes are not globally unique
9685 (e.g. an alleged error in ssl3_accept when a certificate
9686 didn't match the private key).
9688 *) New function SSL_CTX_set_session_id_context that allows to set a default
9689 value (so that you don't need SSL_set_session_id_context for each
9690 connection using the SSL_CTX).
9693 *) OAEP decoding bug fix.
9696 *) Support INSTALL_PREFIX for package builders, as proposed by
9700 *) New Configure options "threads" and "no-threads". For systems
9701 where the proper compiler options are known (currently Solaris
9702 and Linux), "threads" is the default.
9705 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
9708 *) Install various scripts to $(OPENSSLDIR)/misc, not to
9709 $(INSTALLTOP)/bin -- they shouldn't clutter directories
9710 such as /usr/local/bin.
9713 *) "make linux-shared" to build shared libraries.
9714 [Niels Poppe <niels@netbox.org>]
9716 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
9719 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
9720 extension adding in x509 utility.
9723 *) Remove NOPROTO sections and error code comments.
9726 *) Partial rewrite of the DEF file generator to now parse the ANSI
9730 *) New Configure options --prefix=DIR and --openssldir=DIR.
9733 *) Complete rewrite of the error code script(s). It is all now handled
9734 by one script at the top level which handles error code gathering,
9735 header rewriting and C source file generation. It should be much better
9736 than the old method: it now uses a modified version of Ulf's parser to
9737 read the ANSI prototypes in all header files (thus the old K&R definitions
9738 aren't needed for error creation any more) and do a better job of
9739 translating function codes into names. The old 'ASN1 error code imbedded
9740 in a comment' is no longer necessary and it doesn't use .err files which
9741 have now been deleted. Also the error code call doesn't have to appear all
9742 on one line (which resulted in some large lines...).
9745 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
9748 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
9749 0 (which usually indicates a closed connection), but continue reading.
9752 *) Fix some race conditions.
9755 *) Add support for CRL distribution points extension. Add Certificate
9756 Policies and CRL distribution points documentation.
9759 *) Move the autogenerated header file parts to crypto/opensslconf.h.
9762 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
9763 8 of keying material. Merlin has also confirmed interop with this fix
9764 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
9765 [Merlin Hughes <merlin@baltimore.ie>]
9767 *) Fix lots of warnings.
9768 [Richard Levitte <levitte@stacken.kth.se>]
9770 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
9771 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
9772 [Richard Levitte <levitte@stacken.kth.se>]
9774 *) Fix problems with sizeof(long) == 8.
9775 [Andy Polyakov <appro@fy.chalmers.se>]
9777 *) Change functions to ANSI C.
9780 *) Fix typos in error codes.
9781 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
9783 *) Remove defunct assembler files from Configure.
9786 *) SPARC v8 assembler BIGNUM implementation.
9787 [Andy Polyakov <appro@fy.chalmers.se>]
9789 *) Support for Certificate Policies extension: both print and set.
9790 Various additions to support the r2i method this uses.
9793 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9794 return a const string when you are expecting an allocated buffer.
9797 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9798 types DirectoryString and DisplayText.
9801 *) Add code to allow r2i extensions to access the configuration database,
9802 add an LHASH database driver and add several ctx helper functions.
9805 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9806 fail when they extended the size of a BIGNUM.
9809 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9810 support typesafe stack.
9813 *) Fix typo in SSL_[gs]et_options().
9814 [Nils Frostberg <nils@medcom.se>]
9816 *) Delete various functions and files that belonged to the (now obsolete)
9817 old X509V3 handling code.
9820 *) New Configure option "rsaref".
9823 *) Don't auto-generate pem.h.
9826 *) Introduce type-safe ASN.1 SETs.
9829 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9830 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9832 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9833 that links with OpenSSL (well at least cause lots of warnings), but fear
9834 not: the conversion is trivial, and it eliminates loads of evil casts. A
9835 few STACKed things have been converted already. Feel free to convert more.
9836 In the fullness of time, I'll do away with the STACK type altogether.
9839 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9840 specified in <certfile> by updating the entry in the index.txt file.
9841 This way one no longer has to edit the index.txt file manually for
9842 revoking a certificate. The -revoke option does the gory details now.
9843 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9845 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9846 `-text' option at all and this way the `-noout -text' combination was
9847 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9848 [Ralf S. Engelschall]
9850 *) Make sure a corresponding plain text error message exists for the
9851 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9852 verify callback function determined that a certificate was revoked.
9853 [Ralf S. Engelschall]
9855 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9856 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9857 all available cipers including rc5, which was forgotten until now.
9858 In order to let the testing shell script know which algorithms
9859 are available, a new (up to now undocumented) command
9860 "openssl list-cipher-commands" is used.
9863 *) Bugfix: s_client occasionally would sleep in select() when
9864 it should have checked SSL_pending() first.
9867 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9868 the raw DSA values prior to ASN.1 encoding.
9871 *) Tweaks to Configure
9872 [Niels Poppe <niels@netbox.org>]
9874 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9878 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9881 *) New config option to avoid instructions that are illegal on the 80386.
9882 The default code is faster, but requires at least a 486.
9885 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9886 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9887 same as SSL2_VERSION anyway.
9890 *) New "-showcerts" option for s_client.
9893 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9894 application. Various cleanups and fixes.
9897 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9898 modify error routines to work internally. Add error codes and PBE init
9899 to library startup routines.
9902 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9903 packing functions to asn1 and evp. Changed function names and error
9904 codes along the way.
9907 *) PKCS12 integration: and so it begins... First of several patches to
9908 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9909 objects to objects.h
9912 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9913 and display support for Thawte strong extranet extension.
9916 *) Add LinuxPPC support.
9917 [Jeff Dubrule <igor@pobox.org>]
9919 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9920 bn_div_words in alpha.s.
9921 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9923 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9924 OAEP isn't supported when OpenSSL is built with RSAref.
9925 [Ulf Moeller <ulf@fitug.de>]
9927 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9928 so they no longer are missing under -DNOPROTO.
9929 [Soren S. Jorvang <soren@t.dk>]
9932 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9934 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9935 doesn't work when the session is reused. Coming soon!
9938 *) Fix a security hole, that allows sessions to be reused in the wrong
9939 context thus bypassing client cert protection! All software that uses
9940 client certs and session caches in multiple contexts NEEDS PATCHING to
9941 allow session reuse! A fuller solution is in the works.
9942 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9944 *) Some more source tree cleanups (removed obsolete files
9945 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9946 permission on "config" script to be executable) and a fix for the INSTALL
9948 [Ulf Moeller <ulf@fitug.de>]
9950 *) Remove some legacy and erroneous uses of malloc, free instead of
9952 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9954 *) Make rsa_oaep_test return non-zero on error.
9955 [Ulf Moeller <ulf@fitug.de>]
9957 *) Add support for native Solaris shared libraries. Configure
9958 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9959 if someone would make that last step automatic.
9960 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9962 *) ctx_size was not built with the right compiler during "make links". Fixed.
9965 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9966 except NULL ciphers". This means the default cipher list will no longer
9967 enable NULL ciphers. They need to be specifically enabled e.g. with
9968 the string "DEFAULT:eNULL".
9971 *) Fix to RSA private encryption routines: if p < q then it would
9972 occasionally produce an invalid result. This will only happen with
9973 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9976 *) Be less restrictive and allow also `perl util/perlpath.pl
9977 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9978 because this way one can also use an interpreter named `perl5' (which is
9979 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9980 installed as `perl').
9981 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9983 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9984 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9986 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9987 advapi32.lib to Win32 build and change the pem test comparision
9988 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9989 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9990 and crypto/des/ede_cbcm_enc.c.
9993 *) DES quad checksum was broken on big-endian architectures. Fixed.
9996 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9997 Win32 test batch file so it (might) work again. The Win32 test batch file
9998 is horrible: I feel ill....
10001 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
10002 in e_os.h. Audit of header files to check ANSI and non ANSI
10003 sections: 10 functions were absent from non ANSI section and not exported
10004 from Windows DLLs. Fixed up libeay.num for new functions.
10007 *) Make `openssl version' output lines consistent.
10008 [Ralf S. Engelschall]
10010 *) Fix Win32 symbol export lists for BIO functions: Added
10011 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
10012 to ms/libeay{16,32}.def.
10013 [Ralf S. Engelschall]
10015 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
10016 fine under Unix and passes some trivial tests I've now added. But the
10017 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
10018 added to make sure no one expects that this stuff really works in the
10019 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
10020 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
10022 [Ralf S. Engelschall]
10024 *) Fix the generation of two part addresses in perl.
10025 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
10027 *) Add config entry for Linux on MIPS.
10028 [John Tobey <jtobey@channel1.com>]
10030 *) Make links whenever Configure is run, unless we are on Windoze.
10033 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
10034 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
10038 *) Add a useful kludge to allow package maintainers to specify compiler and
10039 other platforms details on the command line without having to patch the
10040 Configure script everytime: One now can use ``perl Configure
10041 <id>:<details>'', i.e. platform ids are allowed to have details appended
10042 to them (seperated by colons). This is treated as there would be a static
10043 pre-configured entry in Configure's %table under key <id> with value
10044 <details> and ``perl Configure <id>'' is called. So, when you want to
10045 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
10046 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
10047 now, which overrides the FreeBSD-elf entry on-the-fly.
10048 [Ralf S. Engelschall]
10050 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
10053 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
10054 on the `perl Configure ...' command line. This way one can compile
10055 OpenSSL libraries with Position Independent Code (PIC) which is needed
10056 for linking it into DSOs.
10057 [Ralf S. Engelschall]
10059 *) Remarkably, export ciphers were totally broken and no-one had noticed!
10063 *) Cleaned up the LICENSE document: The official contact for any license
10064 questions now is the OpenSSL core team under openssl-core@openssl.org.
10065 And add a paragraph about the dual-license situation to make sure people
10066 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
10067 to the OpenSSL toolkit.
10068 [Ralf S. Engelschall]
10070 *) General source tree makefile cleanups: Made `making xxx in yyy...'
10071 display consistent in the source tree and replaced `/bin/rm' by `rm'.
10072 Additonally cleaned up the `make links' target: Remove unnecessary
10073 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
10074 to speed processing and no longer clutter the display with confusing
10075 stuff. Instead only the actually done links are displayed.
10076 [Ralf S. Engelschall]
10078 *) Permit null encryption ciphersuites, used for authentication only. It used
10079 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
10080 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
10084 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
10085 signed attributes when verifying signatures (this would break them),
10086 the detached data encoding was wrong and public keys obtained using
10087 X509_get_pubkey() weren't freed.
10090 *) Add text documentation for the BUFFER functions. Also added a work around
10091 to a Win95 console bug. This was triggered by the password read stuff: the
10092 last character typed gets carried over to the next fread(). If you were
10093 generating a new cert request using 'req' for example then the last
10094 character of the passphrase would be CR which would then enter the first
10098 *) Added the new `Includes OpenSSL Cryptography Software' button as
10099 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
10100 button and can be used by applications based on OpenSSL to show the
10101 relationship to the OpenSSL project.
10102 [Ralf S. Engelschall]
10104 *) Remove confusing variables in function signatures in files
10105 ssl/ssl_lib.c and ssl/ssl.h.
10106 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10108 *) Don't install bss_file.c under PREFIX/include/
10109 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10111 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
10112 functions that return function pointers and has support for NT specific
10113 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
10114 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
10115 unsigned to signed types: this was killing the Win32 compile.
10118 *) Add new certificate file to stack functions,
10119 SSL_add_dir_cert_subjects_to_stack() and
10120 SSL_add_file_cert_subjects_to_stack(). These largely supplant
10121 SSL_load_client_CA_file(), and can be used to add multiple certs easily
10122 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
10123 This means that Apache-SSL and similar packages don't have to mess around
10124 to add as many CAs as they want to the preferred list.
10127 *) Experiment with doxygen documentation. Currently only partially applied to
10129 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
10130 openssl.doxy as the configuration file.
10133 *) Get rid of remaining C++-style comments which strict C compilers hate.
10134 [Ralf S. Engelschall, pointed out by Carlos Amengual]
10136 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
10137 compiled in by default: it has problems with large keys.
10140 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
10141 DH private keys and/or callback functions which directly correspond to
10142 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
10143 is needed for applications which have to configure certificates on a
10144 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
10146 For the RSA certificate situation is makes no difference, but
10147 for the DSA certificate situation this fixes the "no shared cipher"
10148 problem where the OpenSSL cipher selection procedure failed because the
10149 temporary keys were not overtaken from the context and the API provided
10150 no way to reconfigure them.
10151 The new functions now let applications reconfigure the stuff and they
10152 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
10153 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
10154 non-public-API function ssl_cert_instantiate() is used as a helper
10155 function and also to reduce code redundancy inside ssl_rsa.c.
10156 [Ralf S. Engelschall]
10158 *) Move s_server -dcert and -dkey options out of the undocumented feature
10159 area because they are useful for the DSA situation and should be
10160 recognized by the users.
10161 [Ralf S. Engelschall]
10163 *) Fix the cipher decision scheme for export ciphers: the export bits are
10164 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
10165 SSL_EXP_MASK. So, the original variable has to be used instead of the
10166 already masked variable.
10167 [Richard Levitte <levitte@stacken.kth.se>]
10169 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
10170 [Richard Levitte <levitte@stacken.kth.se>]
10172 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
10173 from `int' to `unsigned int' because it's a length and initialized by
10174 EVP_DigestFinal() which expects an `unsigned int *'.
10175 [Richard Levitte <levitte@stacken.kth.se>]
10177 *) Don't hard-code path to Perl interpreter on shebang line of Configure
10178 script. Instead use the usual Shell->Perl transition trick.
10179 [Ralf S. Engelschall]
10181 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
10182 (in addition to RSA certificates) to match the behaviour of `openssl dsa
10183 -noout -modulus' as it's already the case for `openssl rsa -noout
10184 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
10185 currently the public key is printed (a decision which was already done by
10186 `openssl dsa -modulus' in the past) which serves a similar purpose.
10187 Additionally the NO_RSA no longer completely removes the whole -modulus
10188 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
10190 [Ralf S. Engelschall]
10192 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
10193 BIO. See the source (crypto/evp/bio_ok.c) for more info.
10194 [Arne Ansper <arne@ats.cyber.ee>]
10196 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
10197 to be added. Now both 'req' and 'ca' can use new objects defined in the
10201 *) Add cool BIO that does syslog (or event log on NT).
10202 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
10204 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
10205 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
10206 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
10207 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
10210 *) Add preliminary config info for new extension code.
10213 *) Make RSA_NO_PADDING really use no padding.
10214 [Ulf Moeller <ulf@fitug.de>]
10216 *) Generate errors when private/public key check is done.
10219 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
10220 for some CRL extensions and new objects added.
10223 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
10224 key usage extension and fuller support for authority key id.
10227 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
10228 padding method for RSA, which is recommended for new applications in PKCS
10229 #1 v2.0 (RFC 2437, October 1998).
10230 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
10231 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
10232 against Bleichbacher's attack on RSA.
10233 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
10236 *) Updates to the new SSL compression code
10237 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10239 *) Fix so that the version number in the master secret, when passed
10240 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
10241 (because the server will not accept higher), that the version number
10242 is 0x03,0x01, not 0x03,0x00
10243 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10245 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
10246 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
10247 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
10250 *) Support for RAW extensions where an arbitrary extension can be
10251 created by including its DER encoding. See apps/openssl.cnf for
10255 *) Make sure latest Perl versions don't interpret some generated C array
10256 code as Perl array code in the crypto/err/err_genc.pl script.
10257 [Lars Weber <3weber@informatik.uni-hamburg.de>]
10259 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
10260 not many people have the assembler. Various Win32 compilation fixes and
10261 update to the INSTALL.W32 file with (hopefully) more accurate Win32
10262 build instructions.
10265 *) Modify configure script 'Configure' to automatically create crypto/date.h
10266 file under Win32 and also build pem.h from pem.org. New script
10267 util/mkfiles.pl to create the MINFO file on environments that can't do a
10268 'make files': perl util/mkfiles.pl >MINFO should work.
10271 *) Major rework of DES function declarations, in the pursuit of correctness
10272 and purity. As a result, many evil casts evaporated, and some weirdness,
10273 too. You may find this causes warnings in your code. Zapping your evil
10274 casts will probably fix them. Mostly.
10277 *) Fix for a typo in asn1.h. Bug fix to object creation script
10278 obj_dat.pl. It considered a zero in an object definition to mean
10279 "end of object": none of the objects in objects.h have any zeros
10280 so it wasn't spotted.
10281 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
10283 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
10284 Masking (CBCM). In the absence of test vectors, the best I have been able
10285 to do is check that the decrypt undoes the encrypt, so far. Send me test
10286 vectors if you have them.
10289 *) Correct calculation of key length for export ciphers (too much space was
10290 allocated for null ciphers). This has not been tested!
10293 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
10294 message is now correct (it understands "crypto" and "ssl" on its
10295 command line). There is also now an "update" option. This will update
10296 the util/ssleay.num and util/libeay.num files with any new functions.
10298 perl util/mkdef.pl crypto ssl update
10299 it will update them.
10302 *) Overhauled the Perl interface (perl/*):
10303 - ported BN stuff to OpenSSL's different BN library
10304 - made the perl/ source tree CVS-aware
10305 - renamed the package from SSLeay to OpenSSL (the files still contain
10306 their history because I've copied them in the repository)
10307 - removed obsolete files (the test scripts will be replaced
10308 by better Test::Harness variants in the future)
10309 [Ralf S. Engelschall]
10311 *) First cut for a very conservative source tree cleanup:
10312 1. merge various obsolete readme texts into doc/ssleay.txt
10313 where we collect the old documents and readme texts.
10314 2. remove the first part of files where I'm already sure that we no
10315 longer need them because of three reasons: either they are just temporary
10316 files which were left by Eric or they are preserved original files where
10317 I've verified that the diff is also available in the CVS via "cvs diff
10318 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
10319 the crypto/md/ stuff).
10320 [Ralf S. Engelschall]
10322 *) More extension code. Incomplete support for subject and issuer alt
10323 name, issuer and authority key id. Change the i2v function parameters
10324 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
10325 what that's for :-) Fix to ASN1 macro which messed up
10326 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
10329 *) Preliminary support for ENUMERATED type. This is largely copied from the
10333 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
10334 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10336 *) Make sure `make rehash' target really finds the `openssl' program.
10337 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10339 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
10340 like to hear about it if this slows down other processors.
10343 *) Add CygWin32 platform information to Configure script.
10344 [Alan Batie <batie@aahz.jf.intel.com>]
10346 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
10347 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
10349 *) New program nseq to manipulate netscape certificate sequences
10352 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
10356 *) Fixes to BN code. Previously the default was to define BN_RECURSION
10357 but the BN code had some problems that would cause failures when
10358 doing certificate verification and some other functions.
10359 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10361 *) Add ASN1 and PEM code to support netscape certificate sequences.
10364 *) Add ASN1 and PEM code to support netscape certificate sequences.
10367 *) Add several PKIX and private extended key usage OIDs.
10370 *) Modify the 'ca' program to handle the new extension code. Modify
10371 openssl.cnf for new extension format, add comments.
10374 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
10375 and add a sample to openssl.cnf so req -x509 now adds appropriate
10379 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
10380 error code, add initial support to X509_print() and x509 application.
10383 *) Takes a deep breath and start addding X509 V3 extension support code. Add
10384 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
10385 stuff is currently isolated and isn't even compiled yet.
10388 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
10389 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
10390 Removed the versions check from X509 routines when loading extensions:
10391 this allows certain broken certificates that don't set the version
10392 properly to be processed.
10395 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
10396 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
10397 can still be regenerated with "make depend".
10400 *) Spelling mistake in C version of CAST-128.
10401 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
10403 *) Changes to the error generation code. The perl script err-code.pl
10404 now reads in the old error codes and retains the old numbers, only
10405 adding new ones if necessary. It also only changes the .err files if new
10406 codes are added. The makefiles have been modified to only insert errors
10407 when needed (to avoid needlessly modifying header files). This is done
10408 by only inserting errors if the .err file is newer than the auto generated
10409 C file. To rebuild all the error codes from scratch (the old behaviour)
10410 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
10411 or delete all the .err files.
10414 *) CAST-128 was incorrectly implemented for short keys. The C version has
10415 been fixed, but is untested. The assembler versions are also fixed, but
10416 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
10417 to regenerate it if needed.
10418 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
10419 Hagino <itojun@kame.net>]
10421 *) File was opened incorrectly in randfile.c.
10422 [Ulf Möller <ulf@fitug.de>]
10424 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
10425 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
10426 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
10427 al: it's just almost always a UTCTime. Note this patch adds new error
10428 codes so do a "make errors" if there are problems.
10431 *) Correct Linux 1 recognition in config.
10432 [Ulf Möller <ulf@fitug.de>]
10434 *) Remove pointless MD5 hash when using DSA keys in ca.
10435 [Anonymous <nobody@replay.com>]
10437 *) Generate an error if given an empty string as a cert directory. Also
10438 generate an error if handed NULL (previously returned 0 to indicate an
10439 error, but didn't set one).
10440 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
10442 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
10445 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
10446 parameters. This was causing a warning which killed off the Win32 compile.
10449 *) Remove C++ style comments from crypto/bn/bn_local.h.
10450 [Neil Costigan <neil.costigan@celocom.com>]
10452 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
10453 based on a text string, looking up short and long names and finally
10454 "dot" format. The "dot" format stuff didn't work. Added new function
10455 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
10456 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
10457 OID is not part of the table.
10460 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
10461 X509_LOOKUP_by_alias().
10464 *) Sort openssl functions by name.
10467 *) Get the gendsa program working (hopefully) and add it to app list. Remove
10468 encryption from sample DSA keys (in case anyone is interested the password
10472 *) Make _all_ *_free functions accept a NULL pointer.
10473 [Frans Heymans <fheymans@isaserver.be>]
10475 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
10477 [Anonymous <nobody@replay.com>]
10479 *) s_server should send the CAfile as acceptable CAs, not its own cert.
10480 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10482 *) Don't blow it for numeric -newkey arguments to apps/req.
10483 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10485 *) Temp key "for export" tests were wrong in s3_srvr.c.
10486 [Anonymous <nobody@replay.com>]
10488 *) Add prototype for temp key callback functions
10489 SSL_CTX_set_tmp_{rsa,dh}_callback().
10492 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
10493 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
10496 *) X509_name_add_entry() freed the wrong thing after an error.
10497 [Arne Ansper <arne@ats.cyber.ee>]
10499 *) rsa_eay.c would attempt to free a NULL context.
10500 [Arne Ansper <arne@ats.cyber.ee>]
10502 *) BIO_s_socket() had a broken should_retry() on Windoze.
10503 [Arne Ansper <arne@ats.cyber.ee>]
10505 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
10506 [Arne Ansper <arne@ats.cyber.ee>]
10508 *) Make sure the already existing X509_STORE->depth variable is initialized
10509 in X509_STORE_new(), but document the fact that this variable is still
10510 unused in the certificate verification process.
10511 [Ralf S. Engelschall]
10513 *) Fix the various library and apps files to free up pkeys obtained from
10514 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
10517 *) Fix reference counting in X509_PUBKEY_get(). This makes
10518 demos/maurice/example2.c work, amongst others, probably.
10519 [Steve Henson and Ben Laurie]
10521 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
10522 `openssl' and second, the shortcut symlinks for the `openssl <command>'
10523 are no longer created. This way we have a single and consistent command
10524 line interface `openssl <command>', similar to `cvs <command>'.
10525 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
10527 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
10528 BIT STRING wrapper always have zero unused bits.
10531 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
10534 *) Make the top-level INSTALL documentation easier to understand.
10537 *) Makefiles updated to exit if an error occurs in a sub-directory
10538 make (including if user presses ^C) [Paul Sutton]
10540 *) Make Montgomery context stuff explicit in RSA data structure.
10543 *) Fix build order of pem and err to allow for generated pem.h.
10546 *) Fix renumbering bug in X509_NAME_delete_entry().
10549 *) Enhanced the err-ins.pl script so it makes the error library number
10550 global and can add a library name. This is needed for external ASN1 and
10551 other error libraries.
10554 *) Fixed sk_insert which never worked properly.
10557 *) Fix ASN1 macros so they can handle indefinite length construted
10558 EXPLICIT tags. Some non standard certificates use these: they can now
10562 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
10563 into a single doc/ssleay.txt bundle. This way the information is still
10564 preserved but no longer messes up this directory. Now it's new room for
10565 the new set of documenation files.
10566 [Ralf S. Engelschall]
10568 *) SETs were incorrectly DER encoded. This was a major pain, because they
10569 shared code with SEQUENCEs, which aren't coded the same. This means that
10570 almost everything to do with SETs or SEQUENCEs has either changed name or
10571 number of arguments.
10572 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
10574 *) Fix test data to work with the above.
10577 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
10578 was already fixed by Eric for 0.9.1 it seems.
10579 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
10581 *) Autodetect FreeBSD3.
10584 *) Fix various bugs in Configure. This affects the following platforms:
10588 unixware-2.0-pentium
10592 *) Eliminate generated files from CVS. Reorder tests to regenerate files
10593 before they are needed.
10596 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
10600 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
10602 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
10603 changed SSLeay to OpenSSL in version strings.
10604 [Ralf S. Engelschall]
10606 *) Some fixups to the top-level documents.
10609 *) Fixed the nasty bug where rsaref.h was not found under compile-time
10610 because the symlink to include/ was missing.
10611 [Ralf S. Engelschall]
10613 *) Incorporated the popular no-RSA/DSA-only patches
10614 which allow to compile a RSA-free SSLeay.
10615 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
10617 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
10618 when "ssleay" is still not found.
10619 [Ralf S. Engelschall]
10621 *) Added more platforms to Configure: Cray T3E, HPUX 11,
10622 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
10624 *) Updated the README file.
10625 [Ralf S. Engelschall]
10627 *) Added various .cvsignore files in the CVS repository subdirs
10628 to make a "cvs update" really silent.
10629 [Ralf S. Engelschall]
10631 *) Recompiled the error-definition header files and added
10632 missing symbols to the Win32 linker tables.
10633 [Ralf S. Engelschall]
10635 *) Cleaned up the top-level documents;
10636 o new files: CHANGES and LICENSE
10637 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
10638 o merged COPYRIGHT into LICENSE
10639 o removed obsolete TODO file
10640 o renamed MICROSOFT to INSTALL.W32
10641 [Ralf S. Engelschall]
10643 *) Removed dummy files from the 0.9.1b source tree:
10644 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
10645 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
10646 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
10647 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
10648 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
10649 [Ralf S. Engelschall]
10651 *) Added various platform portability fixes.
10654 *) The Genesis of the OpenSSL rpject:
10655 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
10656 Young and Tim J. Hudson created while they were working for C2Net until
10658 [The OpenSSL Project]
10661 Changes between 0.9.0b and 0.9.1b [not released]
10663 *) Updated a few CA certificates under certs/
10666 *) Changed some BIGNUM api stuff.
10669 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
10670 DGUX x86, Linux Alpha, etc.
10673 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
10674 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
10678 *) Add -strparse option to asn1pars program which parses nested
10680 [Dr Stephen Henson <shenson@bigfoot.com>]
10682 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
10685 *) DSA fix for "ca" program.
10688 *) Added "-genkey" option to "dsaparam" program.
10691 *) Added RIPE MD160 (rmd160) message digest.
10694 *) Added -a (all) option to "ssleay version" command.
10697 *) Added PLATFORM define which is the id given to Configure.
10700 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
10703 *) Extended the ASN.1 parser routines.
10706 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
10709 *) Added a BN_CTX to the BN library.
10712 *) Fixed the weak key values in DES library
10715 *) Changed API in EVP library for cipher aliases.
10718 *) Added support for RC2/64bit cipher.
10721 *) Converted the lhash library to the crypto/mem.c functions.
10724 *) Added more recognized ASN.1 object ids.
10727 *) Added more RSA padding checks for SSL/TLS.
10730 *) Added BIO proxy/filter functionality.
10733 *) Added extra_certs to SSL_CTX which can be used
10734 send extra CA certificates to the client in the CA cert chain sending
10735 process. It can be configured with SSL_CTX_add_extra_chain_cert().
10738 *) Now Fortezza is denied in the authentication phase because
10739 this is key exchange mechanism is not supported by SSLeay at all.
10742 *) Additional PKCS1 checks.
10745 *) Support the string "TLSv1" for all TLS v1 ciphers.
10748 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
10749 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
10752 *) Fixed a few memory leaks.
10755 *) Fixed various code and comment typos.
10758 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
10759 bytes sent in the client random.
10760 [Edward Bishop <ebishop@spyglass.com>]
projects / thirdparty / openssl.git / blob