4 Changes between 0.9.6 and 0.9.7 [xx XXX 2000]
6 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
7 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
8 when writing a 32767 byte record.
9 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
11 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
12 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
14 (RSA objects have a reference count access to which is protected
15 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
16 so they are meant to be shared between threads.)
17 [Bodo Moeller, Geoff Thorpe; original patch submitted by
18 "Reddie, Steven" <Steven.Reddie@ca.com>]
20 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
21 entries for variables.
23 *) Add functionality to apps/openssl.c for detecting locking
24 problems: As the program is single-threaded, all we have
25 to do is register a locking callback using an array for
26 storing which locks are currently held by the program.
28 Fix a deadlock in CRYPTO_mem_leaks() that was detected in
32 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
33 SSL_get_ex_data_X509_STORE_idx(), which is used in
34 ssl_verify_cert_chain() and thus can be called at any time
35 during TLS/SSL handshakes so that thread-safety is essential.
36 Unfortunately, the ex_data design is not at all suited
37 for multi-threaded use, so it probably should be abolished.
40 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
41 [Broadcom, tweaked and integrated by Geoff Thorpe]
43 *) Move common extension printing code to new function
44 X509V3_print_extensions(). Reorganise OCSP print routines and
45 implement some needed OCSP ASN1 functions. Add OCSP extensions.
48 *) New function X509_signature_print() to remove duplication in some
52 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
53 set (this was treated exactly the same as SET OF previously). This
54 is used to reorder the STACK representing the structure to match the
55 encoding. This will be used to get round a problem where a PKCS7
56 structure which was signed could not be verified because the STACK
57 order did not reflect the encoded order.
60 *) Reimplement the OCSP ASN1 module using the new code.
63 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
64 for its ASN1 operations. The old style function pointers still exist
65 for now but they will eventually go away.
68 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
69 completely replaces the old ASN1 functionality.
72 *) Change BN_mod_exp_recp so that negative moduli are tolerated
73 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
74 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
78 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
79 of not touching the result's sign bit.
82 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
86 *) Changed the LHASH code to use prototypes for callbacks, and created
87 macros to declare and implement thin (optionally static) functions
88 that provide type-safety and avoid function pointer casting for the
89 type-specific callbacks.
92 *) Use better test patterns in bntest.
95 *) Added Kerberos Cipher Suites to be used with TLS, as written in
97 [Veers Staats <staatsvr@asc.hpc.mil>,
98 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
100 *) rand_win.c fix for Borland C.
103 *) BN_rshift bugfix for n == 0.
106 *) Reformat the FAQ so the different questions and answers can be divided
107 in sections depending on the subject.
110 *) Have the zlib compression code load ZLIB.DLL dynamically under
114 *) New function BN_mod_sqrt for computing square roots modulo a prime
115 (using the probabilistic Tonelli-Shanks algorithm unless
116 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
117 be handled deterministically).
118 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
120 *) Store verify_result within SSL_SESSION also for client side to
121 avoid potential security hole. (Re-used sessions on the client side
122 always resulted in verify_result==X509_V_OK, not using the original
123 result of the server certificate verification.)
126 *) Make BN_mod_inverse faster by explicitly handling small quotients
127 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
128 512 bits], about 30% for larger ones [1024 or 2048 bits].)
131 *) Disable ssl2_peek and ssl3_peek (i.e., both implementations
132 of SSL_peek) because they both are completely broken.
133 For fixing this, the internal read functions now have an additional
134 'peek' parameter, but the actual peek functionality has not
135 yet been implemented.
138 *) New function BN_kronecker.
141 *) Fix BN_gcd so that it works on negative inputs; the result is
142 positive unless both parameters are zero.
143 Previously something reasonably close to an infinite loop was
144 possible because numbers could be growing instead of shrinking
145 in the implementation of Euclid's algorithm.
148 *) Fix BN_is_word() and BN_is_one() macros to take into account the
149 sign of the number in question.
151 Fix BN_is_word(a,w) to work correctly for w == 0.
153 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
154 because its test if the absolute value of 'a' equals 'w'.
155 Note that BN_abs_is_word does *not* handle w == 0 reliably;
156 it exists mostly for use in the implementations of BN_is_zero(),
157 BN_is_one(), and BN_is_word().
160 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
161 the method-specific "init()" handler. Also clean up ex_data after
162 calling the method-specific "finish()" handler. Previously, this was
163 happening the other way round.
166 *) New function BN_swap.
169 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
170 the exponentiation functions are more likely to produce reasonable
171 results on negative inputs.
174 *) Change BN_mod_mul so that the result is always non-negative.
175 Previously, it could be negative if one of the factors was negative;
176 I don't think anyone really wanted that behaviour.
179 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
180 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
181 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
182 and add new functions:
195 These functions always generate non-negative results.
197 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
198 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
200 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
201 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
203 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
205 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
206 was actually never needed) and in BN_mul(). The removal in BN_mul()
207 required a small change in bn_mul_part_recursive() and the addition
208 of the functions bn_cmp_part_words(), bn_sub_part_words() and
209 bn_add_part_words(), which do the same thing as bn_cmp_words(),
210 bn_sub_words() and bn_add_words() except they take arrays with
214 *) In 'openssl passwd', verify passwords read from the terminal
215 unless the '-salt' option is used (which usually means that
216 verification would just waste user's time since the resulting
217 hash is going to be compared with some given password hash)
218 or the new '-noverify' option is used.
220 This is an incompatible change, but it does not affect
221 non-interactive use of 'openssl passwd' (passwords on the command
222 line, '-stdin' option, '-in ...' option) and thus should not
226 *) Remove all references to RSAref, since there's no more need for it.
229 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
230 The previous value, 12, was not always sufficient for BN_mod_exp().
233 *) Make DSO load along a path given through an environment variable
234 (SHLIB_PATH) with shl_load().
237 *) Constify the ENGINE code as a result of BIGNUM constification.
238 Also constify the RSA code and most things related to it. In a
239 few places, most notable in the depth of the ASN.1 code, ugly
240 casts back to non-const were required (to be solved at a later
244 *) Make it so the openssl application has all engines loaded by default.
247 *) Constify the BIGNUM routines a little more.
250 *) Make sure that shared libraries get the internal name engine with
251 the full version number and not just 0. This should mark the
252 shared libraries as not backward compatible. Of course, this should
253 be changed again when we can guarantee backward binary compatibility.
256 *) Add the following functions:
262 ENGINE_load_builtin_engines()
264 That way, an application can itself choose if external engines that
265 are built-in in OpenSSL shall ever be used or not. The benefit is
266 that applications won't have to be linked with libdl or other dso
267 libraries unless it's really needed.
269 Changed 'openssl engine' to load all engines on demand.
270 Changed the engine header files to avoid the duplication of some
271 declarations (they differed!).
274 *) 'openssl engine' can now list capabilities.
277 *) Better error reporting in 'openssl engine'.
280 *) Never call load_dh_param(NULL) in s_server.
283 *) Add engine application. It can currently list engines by name and
284 identity, and test if they are actually available.
287 *) Add support for shared libraries under Irix.
288 [Albert Chin-A-Young <china@thewrittenword.com>]
290 *) Improve RPM specification file by forcing symbolic linking and making
291 sure the installed documentation is also owned by root.root.
292 [Damien Miller <djm@mindrot.org>]
294 *) Add configuration option to build on Linux on both big-endian and
296 [Ralf Baechle <ralf@uni-koblenz.de>]
298 *) Give the OpenSSL applications more possibilities to make use of
299 keys (public as well as private) handled by engines.
302 *) Add OCSP code that comes from CertCo.
305 *) Add VMS support for the Rijndael code.
308 *) Added untested support for Nuron crypto accelerator.
311 *) Add support for external cryptographic devices. This code was
312 previously distributed separately as the "engine" branch.
313 [Geoff Thorpe, Richard Levitte]
315 *) Rework the filename-translation in the DSO code. It is now possible to
316 have far greater control over how a "name" is turned into a filename
317 depending on the operating environment and any oddities about the
318 different shared library filenames on each system.
321 *) Support threads on FreeBSD-elf in Configure.
324 *) Add the possibility to create shared libraries on HP-UX
327 *) Fix for SHA1 assembly problem with MASM: it produces
328 warnings about corrupt line number information when assembling
329 with debugging information. This is caused by the overlapping
331 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
334 NCONF_get_number() has no error checking at all. As a replacement,
335 NCONF_get_number_e() is defined (_e for "error checking") and is
336 promoted strongly. The old NCONF_get_number is kept around for
337 binary backward compatibility.
338 Make it possible for methods to load from something other than a BIO,
339 by providing a function pointer that is given a name instead of a BIO.
340 For example, this could be used to load configuration data from an
344 *) Fix typo in get_cert_by_subject() in by_dir.c
345 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
347 *) Rework the system to generate shared libraries:
349 - Make note of the expected extension for the shared libraries and
350 if there is a need for symbolic links from for example libcrypto.so.0
351 to libcrypto.so.0.9.7. There is extended info in Configure for
354 - Make as few rebuilds of the shared libraries as possible.
356 - Still avoid linking the OpenSSL programs with the shared libraries.
358 - When installing, install the shared libraries separately from the
362 *) Fix for non blocking accept BIOs. Added new I/O special reason
363 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
364 with non blocking I/O was not possible because no retry code was
365 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
369 *) Added the beginnings of Rijndael support.
372 *) Fix for bug in DirectoryString mask setting. Add support for
373 X509_NAME_print_ex() in 'req' and X509_print_ex() function
374 to allow certificate printing to more controllable, additional
375 'certopt' option to 'x509' to allow new printing options to be
379 *) Clean old EAY MD5 hack from e_os.h.
382 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
384 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
385 and not in SSL_clear because the latter is also used by the
386 accept/connect functions; previously, the settings made by
387 SSL_set_read_ahead would be lost during the handshake.
388 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
390 *) Correct util/mkdef.pl to be selective about disabled algorithms.
391 Previously, it would create entries for disableed algorithms no
395 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
397 *) In ssl23_get_client_hello, generate an error message when faced
398 with an initial SSL 3.0/TLS record that is too small to contain the
399 first two bytes of the ClientHello message, i.e. client_version.
400 (Note that this is a pathologic case that probably has never happened
401 in real life.) The previous approach was to use the version number
402 from the record header as a substitute; but our protocol choice
403 should not depend on that one because it is not authenticated
404 by the Finished messages.
407 *) More robust randomness gathering functions for Windows.
408 [Jeffrey Altman <jaltman@columbia.edu>]
410 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
411 not set then we don't setup the error code for issuer check errors
412 to avoid possibly overwriting other errors which the callback does
413 handle. If an application does set the flag then we assume it knows
414 what it is doing and can handle the new informational codes
418 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
419 a general "ANY" type, as such it should be able to decode anything
420 including tagged types. However it didn't check the class so it would
421 wrongly interpret tagged types in the same way as their universal
422 counterpart and unknown types were just rejected. Changed so that the
423 tagged and unknown types are handled in the same way as a SEQUENCE:
424 that is the encoding is stored intact. There is also a new type
425 "V_ASN1_OTHER" which is used when the class is not universal, in this
426 case we have no idea what the actual type is so we just lump them all
430 *) On VMS, stdout may very well lead to a file that is written to
431 in a record-oriented fashion. That means that every write() will
432 write a separate record, which will be read separately by the
433 programs trying to read from it. This can be very confusing.
435 The solution is to put a BIO filter in the way that will buffer
436 text until a linefeed is reached, and then write everything a
437 line at a time, so every record written will be an actual line,
438 not chunks of lines and not (usually doesn't happen, but I've
439 seen it once) several lines in one record. BIO_f_linebuffer() is
442 Currently, it's a VMS-only method, because that's where it has
443 been tested well enough.
446 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
447 it can return incorrect results.
448 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
449 but it was in 0.9.6-beta[12].)
452 *) Disable the check for content being present when verifying detached
453 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
454 include zero length content when signing messages.
457 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
458 BIO_ctrl (for BIO pairs).
461 *) Add DSO method for VMS.
464 *) Bug fix: Montgomery multiplication could produce results with the
468 *) Add RPM specification openssl.spec and modify it to build three
469 packages. The default package contains applications, application
470 documentation and run-time libraries. The devel package contains
471 include files, static libraries and function documentation. The
472 doc package contains the contents of the doc directory. The original
473 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
476 *) Add a large number of documentation files for many SSL routines.
477 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
479 *) Add a configuration entry for Sony News 4.
480 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
482 *) Don't set the two most significant bits to one when generating a
483 random number < q in the DSA library.
486 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
487 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
488 the underlying transport is blocking) if a handshake took place.
489 (The default behaviour is needed by applications such as s_client
490 and s_server that use select() to determine when to use SSL_read;
491 but for applications that know in advance when to expect data, it
492 just makes things more complicated.)
495 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
499 *) Add a few more EBCDIC conditionals that make `req' and `x509'
500 work better on such systems.
501 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
503 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
504 Update PKCS12_parse() so it copies the friendlyName and the
505 keyid to the certificates aux info.
508 *) Fix bug in PKCS7_verify() which caused an infinite loop
509 if there was more than one signature.
510 [Sven Uszpelkat <su@celocom.de>]
512 *) Major change in util/mkdef.pl to include extra information
513 about each symbol, as well as presentig variables as well
514 as functions. This change means that there's n more need
515 to rebuild the .num files when some algorithms are excluded.
518 *) Allow the verify time to be set by an application,
519 rather than always using the current time.
522 *) Phase 2 verify code reorganisation. The certificate
523 verify code now looks up an issuer certificate by a
524 number of criteria: subject name, authority key id
525 and key usage. It also verifies self signed certificates
526 by the same criteria. The main comparison function is
527 X509_check_issued() which performs these checks.
529 Lot of changes were necessary in order to support this
530 without completely rewriting the lookup code.
532 Authority and subject key identifier are now cached.
534 The LHASH 'certs' is X509_STORE has now been replaced
535 by a STACK_OF(X509_OBJECT). This is mainly because an
536 LHASH can't store or retrieve multiple objects with
539 As a result various functions (which were all internal
540 use only) have changed to handle the new X509_STORE
541 structure. This will break anything that messed round
542 with X509_STORE internally.
544 The functions X509_STORE_add_cert() now checks for an
545 exact match, rather than just subject name.
547 The X509_STORE API doesn't directly support the retrieval
548 of multiple certificates matching a given criteria, however
549 this can be worked round by performing a lookup first
550 (which will fill the cache with candidate certificates)
551 and then examining the cache for matches. This is probably
552 the best we can do without throwing out X509_LOOKUP
553 entirely (maybe later...).
555 The X509_VERIFY_CTX structure has been enhanced considerably.
557 All certificate lookup operations now go via a get_issuer()
558 callback. Although this currently uses an X509_STORE it
559 can be replaced by custom lookups. This is a simple way
560 to bypass the X509_STORE hackery necessary to make this
561 work and makes it possible to use more efficient techniques
562 in future. A very simple version which uses a simple
563 STACK for its trusted certificate store is also provided
564 using X509_STORE_CTX_trusted_stack().
566 The verify_cb() and verify() callbacks now have equivalents
567 in the X509_STORE_CTX structure.
569 X509_STORE_CTX also has a 'flags' field which can be used
570 to customise the verify behaviour.
573 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
574 excludes S/MIME capabilities.
577 *) When a certificate request is read in keep a copy of the
578 original encoding of the signed data and use it when outputing
579 again. Signatures then use the original encoding rather than
580 a decoded, encoded version which may cause problems if the
581 request is improperly encoded.
584 *) For consistency with other BIO_puts implementations, call
585 buffer_write(b, ...) directly in buffer_puts instead of calling
588 In BIO_puts, increment b->num_write as in BIO_write.
589 [Peter.Sylvester@EdelWeb.fr]
591 *) Fix BN_mul_word for the case where the word is 0. (We have to use
592 BN_zero, we may not return a BIGNUM with an array consisting of
596 *) Avoid calling abort() from within the library when problems are
597 detected, except if preprocessor symbols have been defined
598 (such as REF_CHECK, BN_DEBUG etc.).
601 *) New openssl application 'rsautl'. This utility can be
602 used for low level RSA operations. DER public key
603 BIO/fp routines also added.
606 *) New Configure entry and patches for compiling on QNX 4.
607 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
609 *) A demo state-machine implementation was sponsored by
610 Nuron (http://www.nuron.com/) and is now available in
614 *) New options added to the 'dgst' utility for signature
615 generation and verification.
618 *) Unrecognized PKCS#7 content types are now handled via a
619 catch all ASN1_TYPE structure. This allows unsupported
620 types to be stored as a "blob" and an application can
621 encode and decode it manually.
624 *) Fix various signed/unsigned issues to make a_strex.c
626 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
628 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
629 length if passed a buffer. ASN1_INTEGER_to_BN failed
630 if passed a NULL BN and its argument was negative.
631 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
633 *) Modification to PKCS#7 encoding routines to output definite
634 length encoding. Since currently the whole structures are in
635 memory there's not real point in using indefinite length
636 constructed encoding. However if OpenSSL is compiled with
637 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
640 *) Added BIO_vprintf() and BIO_vsnprintf().
643 *) Added more prefixes to parse for in the the strings written
644 through a logging bio, to cover all the levels that are available
645 through syslog. The prefixes are now:
647 PANIC, EMERG, EMR => LOG_EMERG
648 ALERT, ALR => LOG_ALERT
649 CRIT, CRI => LOG_CRIT
650 ERROR, ERR => LOG_ERR
651 WARNING, WARN, WAR => LOG_WARNING
652 NOTICE, NOTE, NOT => LOG_NOTICE
653 INFO, INF => LOG_INFO
654 DEBUG, DBG => LOG_DEBUG
656 and as before, if none of those prefixes are present at the
657 beginning of the string, LOG_ERR is chosen.
659 On Win32, the LOG_* levels are mapped according to this:
661 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
662 LOG_WARNING => EVENTLOG_WARNING_TYPE
663 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
667 *) Made it possible to reconfigure with just the configuration
668 argument "reconf" or "reconfigure". The command line arguments
669 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
670 and are retrieved from there when reconfiguring.
674 [Assar Westerlund <assar@sics.se>, Richard Levitte]
676 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
679 *) The obj_dat.pl script was messing up the sorting of object
680 names. The reason was that it compared the quoted version
681 of strings as a result "OCSP" > "OCSP Signing" because
682 " > SPACE. Changed script to store unquoted versions of
683 names and add quotes on output. It was also omitting some
684 names from the lookup table if they were given a default
685 value (that is if SN is missing it is given the same
686 value as LN and vice versa), these are now added on the
687 grounds that if an object has a name we should be able to
688 look it up. Finally added warning output when duplicate
689 short or long names are found.
692 *) Changes needed for Tandem NSK.
693 [Scott Uroff <scott@xypro.com>]
695 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
696 RSA_padding_check_SSLv23(), special padding was never detected
697 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
698 version rollback attacks was not effective.
700 In s23_clnt.c, don't use special rollback-attack detection padding
701 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
702 client; similarly, in s23_srvr.c, don't do the rollback check if
703 SSL 2.0 is the only protocol enabled in the server.
706 *) Make it possible to get hexdumps of unprintable data with 'openssl
707 asn1parse'. By implication, the functions ASN1_parse_dump() and
708 BIO_dump_indent() are added.
711 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
712 these print out strings and name structures based on various
713 flags including RFC2253 support and proper handling of
714 multibyte characters. Added options to the 'x509' utility
715 to allow the various flags to be set.
718 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
719 Also change the functions X509_cmp_current_time() and
720 X509_gmtime_adj() work with an ASN1_TIME structure,
721 this will enable certificates using GeneralizedTime in validity
725 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
726 negative public key encodings) on by default,
727 NO_NEG_PUBKEY_BUG can be set to disable it.
730 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
731 content octets. An i2c_ASN1_OBJECT is unnecessary because
732 the encoding can be trivially obtained from the structure.
735 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
736 not read locks (CRYPTO_r_[un]lock).
739 *) A first attempt at creating official support for shared
740 libraries through configuration. I've kept it so the
741 default is static libraries only, and the OpenSSL programs
742 are always statically linked for now, but there are
743 preparations for dynamic linking in place.
744 This has been tested on Linux and True64.
747 *) Randomness polling function for Win9x, as described in:
748 Peter Gutmann, Software Generation of Practically Strong
752 *) Fix so PRNG is seeded in req if using an already existing
756 *) New options to smime application. -inform and -outform
757 allow alternative formats for the S/MIME message including
758 PEM and DER. The -content option allows the content to be
759 specified separately. This should allow things like Netscape
760 form signing output easier to verify.
763 *) Fix the ASN1 encoding of tags using the 'long form'.
766 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
767 STRING types. These convert content octets to and from the
768 underlying type. The actual tag and length octets are
769 already assumed to have been read in and checked. These
770 are needed because all other string types have virtually
771 identical handling apart from the tag. By having versions
772 of the ASN1 functions that just operate on content octets
773 IMPLICIT tagging can be handled properly. It also allows
774 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
775 and ASN1_INTEGER are identical apart from the tag.
778 *) Change the handling of OID objects as follows:
780 - New object identifiers are inserted in objects.txt, following
781 the syntax given in objects.README.
782 - objects.pl is used to process obj_mac.num and create a new
784 - obj_dat.pl is used to create a new obj_dat.h, using the data in
787 This is currently kind of a hack, and the perl code in objects.pl
788 isn't very elegant, but it works as I intended. The simplest way
789 to check that it worked correctly is to look in obj_dat.h and
790 check the array nid_objs and make sure the objects haven't moved
791 around (this is important!). Additions are OK, as well as
792 consistent name changes.
795 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
798 *) Addition of the command line parameter '-rand file' to 'openssl req'.
799 The given file adds to whatever has already been seeded into the
800 random pool through the RANDFILE configuration file option or
801 environment variable, or the default random state file.
804 *) mkstack.pl now sorts each macro group into lexical order.
805 Previously the output order depended on the order the files
806 appeared in the directory, resulting in needless rewriting
810 *) Patches to make OpenSSL compile under Win32 again. Mostly
811 work arounds for the VC++ problem that it treats func() as
812 func(void). Also stripped out the parts of mkdef.pl that
813 added extra typesafe functions: these no longer exist.
816 *) Reorganisation of the stack code. The macros are now all
817 collected in safestack.h . Each macro is defined in terms of
818 a "stack macro" of the form SKM_<name>(type, a, b). The
819 DEBUG_SAFESTACK is now handled in terms of function casts,
820 this has the advantage of retaining type safety without the
821 use of additional functions. If DEBUG_SAFESTACK is not defined
822 then the non typesafe macros are used instead. Also modified the
823 mkstack.pl script to handle the new form. Needs testing to see
824 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
825 the default if no major problems. Similar behaviour for ASN1_SET_OF
829 *) When some versions of IIS use the 'NET' form of private key the
830 key derivation algorithm is different. Normally MD5(password) is
831 used as a 128 bit RC4 key. In the modified case
832 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
833 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
834 as the old Netscape_RSA functions except they have an additional
835 'sgckey' parameter which uses the modified algorithm. Also added
836 an -sgckey command line option to the rsa utility. Thanks to
837 Adrian Peck <bertie@ncipher.com> for posting details of the modified
838 algorithm to openssl-dev.
841 *) The evp_local.h macros were using 'c.##kname' which resulted in
842 invalid expansion on some systems (SCO 5.0.5 for example).
843 Corrected to 'c.kname'.
844 [Phillip Porch <root@theporch.com>]
846 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
847 a STACK of email addresses from a certificate or request, these look
848 in the subject name and the subject alternative name extensions and
849 omit any duplicate addresses.
852 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
853 This makes DSA verification about 2 % faster.
856 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
857 (meaning that now 2^5 values will be precomputed, which is only 4 KB
858 plus overhead for 1024 bit moduli).
859 This makes exponentiations about 0.5 % faster for 1024 bit
860 exponents (as measured by "openssl speed rsa2048").
863 *) Rename memory handling macros to avoid conflicts with other
865 Malloc => OPENSSL_malloc
866 Malloc_locked => OPENSSL_malloc_locked
867 Realloc => OPENSSL_realloc
871 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
872 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
876 [John Jarvie <jjarvie@newsguy.com>]
878 *) The type-safe stack code has been rejigged. It is now only compiled
879 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
880 by default all type-specific stack functions are "#define"d back to
881 standard stack functions. This results in more streamlined output
882 but retains the type-safety checking possibilities of the original
886 *) The STACK code has been cleaned up, and certain type declarations
887 that didn't make a lot of sense have been brought in line. This has
888 also involved a cleanup of sorts in safestack.h to more correctly
889 map type-safe stack functions onto their plain stack counterparts.
890 This work has also resulted in a variety of "const"ifications of
891 lots of the code, especially "_cmp" operations which should normally
892 be prototyped with "const" parameters anyway.
895 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
896 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
897 (The PRNG state consists of two parts, the large pool 'state' and 'md',
898 where all of 'md' is used each time the PRNG is used, but 'state'
899 is used only indexed by a cyclic counter. As entropy may not be
900 well distributed from the beginning, 'md' is important as a
901 chaining variable. However, the output function chains only half
902 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
903 all of 'md', and seeding with STATE_SIZE dummy bytes will result
904 in all of 'state' being rewritten, with the new values depending
905 on virtually all of 'md'. This overcomes the 80 bit limitation.)
908 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
909 the handshake is continued after ssl_verify_cert_chain();
910 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
911 can lead to 'unexplainable' connection aborts later.
912 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
914 *) Major EVP API cipher revision.
915 Add hooks for extra EVP features. This allows various cipher
916 parameters to be set in the EVP interface. Support added for variable
917 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
918 setting of RC2 and RC5 parameters.
920 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
923 Remove lots of duplicated code from the EVP library. For example *every*
924 cipher init() function handles the 'iv' in the same way according to the
925 cipher mode. They also all do nothing if the 'key' parameter is NULL and
926 for CFB and OFB modes they zero ctx->num.
928 New functionality allows removal of S/MIME code RC2 hack.
930 Most of the routines have the same form and so can be declared in terms
933 By shifting this to the top level EVP_CipherInit() it can be removed from
934 all individual ciphers. If the cipher wants to handle IVs or keys
935 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
938 Change lots of functions like EVP_EncryptUpdate() to now return a
939 value: although software versions of the algorithms cannot fail
940 any installed hardware versions can.
943 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
944 this option is set, tolerate broken clients that send the negotiated
945 protocol version number instead of the requested protocol version
949 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
950 i.e. non-zero for export ciphersuites, zero otherwise.
951 Previous versions had this flag inverted, inconsistent with
952 rsa_tmp_cb (..._TMP_RSA_CB).
953 [Bodo Moeller; problem reported by Amit Chopra]
955 *) Add missing DSA library text string. Work around for some IIS
956 key files with invalid SEQUENCE encoding.
959 *) Add a document (doc/standards.txt) that list all kinds of standards
960 and so on that are implemented in OpenSSL.
963 *) Enhance c_rehash script. Old version would mishandle certificates
964 with the same subject name hash and wouldn't handle CRLs at all.
965 Added -fingerprint option to crl utility, to support new c_rehash
969 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
972 *) Fix for SSL server purpose checking. Server checking was
973 rejecting certificates which had extended key usage present
974 but no ssl client purpose.
975 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
977 *) Make PKCS#12 code work with no password. The PKCS#12 spec
978 is a little unclear about how a blank password is handled.
979 Since the password in encoded as a BMPString with terminating
980 double NULL a zero length password would end up as just the
981 double NULL. However no password at all is different and is
982 handled differently in the PKCS#12 key generation code. NS
983 treats a blank password as zero length. MSIE treats it as no
984 password on export: but it will try both on import. We now do
985 the same: PKCS12_parse() tries zero length and no password if
986 the password is set to "" or NULL (NULL is now a valid password:
987 it wasn't before) as does the pkcs12 application.
990 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
991 perror when PEM_read_bio_X509_REQ fails, the error message must
992 be obtained from the error queue.
995 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
996 it in ERR_remove_state if appropriate, and change ERR_get_state
997 accordingly to avoid race conditions (this is necessary because
998 thread_hash is no longer constant once set).
1001 *) Bugfix for linux-elf makefile.one.
1004 *) RSA_get_default_method() will now cause a default
1005 RSA_METHOD to be chosen if one doesn't exist already.
1006 Previously this was only set during a call to RSA_new()
1007 or RSA_new_method(NULL) meaning it was possible for
1008 RSA_get_default_method() to return NULL.
1011 *) Added native name translation to the existing DSO code
1012 that will convert (if the flag to do so is set) filenames
1013 that are sufficiently small and have no path information
1014 into a canonical native form. Eg. "blah" converted to
1015 "libblah.so" or "blah.dll" etc.
1018 *) New function ERR_error_string_n(e, buf, len) which is like
1019 ERR_error_string(e, buf), but writes at most 'len' bytes
1020 including the 0 terminator. For ERR_error_string_n, 'buf'
1022 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
1024 *) CONF library reworked to become more general. A new CONF
1025 configuration file reader "class" is implemented as well as a
1026 new functions (NCONF_*, for "New CONF") to handle it. The now
1027 old CONF_* functions are still there, but are reimplemented to
1028 work in terms of the new functions. Also, a set of functions
1029 to handle the internal storage of the configuration data is
1030 provided to make it easier to write new configuration file
1031 reader "classes" (I can definitely see something reading a
1032 configuration file in XML format, for example), called _CONF_*,
1033 or "the configuration storage API"...
1035 The new configuration file reading functions are:
1037 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
1038 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
1040 NCONF_default, NCONF_WIN32
1042 NCONF_dump_fp, NCONF_dump_bio
1044 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
1045 NCONF_new creates a new CONF object. This works in the same way
1046 as other interfaces in OpenSSL, like the BIO interface.
1047 NCONF_dump_* dump the internal storage of the configuration file,
1048 which is useful for debugging. All other functions take the same
1049 arguments as the old CONF_* functions wth the exception of the
1050 first that must be a `CONF *' instead of a `LHASH *'.
1052 To make it easer to use the new classes with the old CONF_* functions,
1053 the function CONF_set_default_method is provided.
1056 *) Add '-tls1' option to 'openssl ciphers', which was already
1057 mentioned in the documentation but had not been implemented.
1058 (This option is not yet really useful because even the additional
1059 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
1062 *) Initial DSO code added into libcrypto for letting OpenSSL (and
1063 OpenSSL-based applications) load shared libraries and bind to
1064 them in a portable way.
1065 [Geoff Thorpe, with contributions from Richard Levitte]
1067 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
1069 *) Make sure _lrotl and _lrotr are only used with MSVC.
1071 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
1072 (the default implementation of RAND_status).
1074 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
1075 to '-clrext' (= clear extensions), as intended and documented.
1076 [Bodo Moeller; inconsistency pointed out by Michael Attili
1079 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
1080 was larger than the MD block size.
1081 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
1083 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
1084 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
1085 using the passed key: if the passed key was a private key the result
1086 of X509_print(), for example, would be to print out all the private key
1090 *) des_quad_cksum() byte order bug fix.
1091 [Ulf Möller, using the problem description in krb4-0.9.7, where
1092 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
1094 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
1096 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
1098 *) For easily testing in shell scripts whether some command
1099 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
1100 returns with exit code 0 iff no command of the given name is available.
1101 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
1102 the output goes to stdout and nothing is printed to stderr.
1103 Additional arguments are always ignored.
1105 Since for each cipher there is a command of the same name,
1106 the 'no-cipher' compilation switches can be tested this way.
1108 ('openssl no-XXX' is not able to detect pseudo-commands such
1109 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
1112 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
1115 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
1116 is set; it will be thrown away anyway because each handshake creates
1118 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
1119 to parameters -- in previous versions (since OpenSSL 0.9.3) the
1120 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
1121 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
1124 *) New s_client option -ign_eof: EOF at stdin is ignored, and
1125 'Q' and 'R' lose their special meanings (quit/renegotiate).
1126 This is part of what -quiet does; unlike -quiet, -ign_eof
1127 does not suppress any output.
1130 *) Add compatibility options to the purpose and trust code. The
1131 purpose X509_PURPOSE_ANY is "any purpose" which automatically
1132 accepts a certificate or CA, this was the previous behaviour,
1133 with all the associated security issues.
1135 X509_TRUST_COMPAT is the old trust behaviour: only and
1136 automatically trust self signed roots in certificate store. A
1137 new trust setting X509_TRUST_DEFAULT is used to specify that
1138 a purpose has no associated trust setting and it should instead
1139 use the value in the default purpose.
1142 *) Fix the PKCS#8 DSA private key code so it decodes keys again
1143 and fix a memory leak.
1146 *) In util/mkerr.pl (which implements 'make errors'), preserve
1147 reason strings from the previous version of the .c file, as
1148 the default to have only downcase letters (and digits) in
1149 automatically generated reasons codes is not always appropriate.
1152 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
1153 using strerror. Previously, ERR_reason_error_string() returned
1154 library names as reason strings for SYSerr; but SYSerr is a special
1155 case where small numbers are errno values, not library numbers.
1158 *) Add '-dsaparam' option to 'openssl dhparam' application. This
1159 converts DSA parameters into DH parameters. (When creating parameters,
1160 DSA_generate_parameters is used.)
1163 *) Include 'length' (recommended exponent length) in C code generated
1164 by 'openssl dhparam -C'.
1167 *) The second argument to set_label in perlasm was already being used
1168 so couldn't be used as a "file scope" flag. Moved to third argument
1172 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
1173 instead of RAND_bytes for encryption IVs and salts.
1176 *) Include RAND_status() into RAND_METHOD instead of implementing
1177 it only for md_rand.c Otherwise replacing the PRNG by calling
1178 RAND_set_rand_method would be impossible.
1181 *) Don't let DSA_generate_key() enter an infinite loop if the random
1182 number generation fails.
1185 *) New 'rand' application for creating pseudo-random output.
1188 *) Added configuration support for Linux/IA64
1189 [Rolf Haberrecker <rolf@suse.de>]
1191 *) Assembler module support for Mingw32.
1194 *) Shared library support for HPUX (in shlib/).
1195 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
1197 *) Shared library support for Solaris gcc.
1198 [Lutz Behnke <behnke@trustcenter.de>]
1200 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
1202 *) PKCS7_encrypt() was adding text MIME headers twice because they
1203 were added manually and by SMIME_crlf_copy().
1206 *) In bntest.c don't call BN_rand with zero bits argument.
1207 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
1209 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
1210 case was implemented. This caused BN_div_recp() to fail occasionally.
1213 *) Add an optional second argument to the set_label() in the perl
1214 assembly language builder. If this argument exists and is set
1215 to 1 it signals that the assembler should use a symbol whose
1216 scope is the entire file, not just the current function. This
1217 is needed with MASM which uses the format label:: for this scope.
1218 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
1220 *) Change the ASN1 types so they are typedefs by default. Before
1221 almost all types were #define'd to ASN1_STRING which was causing
1222 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
1226 *) Change names of new functions to the new get1/get0 naming
1227 convention: After 'get1', the caller owns a reference count
1228 and has to call ..._free; 'get0' returns a pointer to some
1229 data structure without incrementing reference counters.
1230 (Some of the existing 'get' functions increment a reference
1231 counter, some don't.)
1232 Similarly, 'set1' and 'add1' functions increase reference
1233 counters or duplicate objects.
1236 *) Allow for the possibility of temp RSA key generation failure:
1237 the code used to assume it always worked and crashed on failure.
1240 *) Fix potential buffer overrun problem in BIO_printf().
1241 [Ulf Möller, using public domain code by Patrick Powell; problem
1242 pointed out by David Sacerdote <das33@cornell.edu>]
1244 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
1245 RAND_egd() and RAND_status(). In the command line application,
1246 the EGD socket can be specified like a seed file using RANDFILE
1250 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
1251 Some CAs (e.g. Verisign) distribute certificates in this form.
1254 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
1255 list to exclude them. This means that no special compilation option
1256 is needed to use anonymous DH: it just needs to be included in the
1260 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
1261 EVP_MD_type. The old functionality is available in a new macro called
1262 EVP_MD_md(). Change code that uses it and update docs.
1265 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
1266 where the 'void *' argument is replaced by a function pointer argument.
1267 Previously 'void *' was abused to point to functions, which works on
1268 many platforms, but is not correct. As these functions are usually
1269 called by macros defined in OpenSSL header files, most source code
1270 should work without changes.
1273 *) <openssl/opensslconf.h> (which is created by Configure) now contains
1274 sections with information on -D... compiler switches used for
1275 compiling the library so that applications can see them. To enable
1276 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
1277 must be defined. E.g.,
1278 #define OPENSSL_ALGORITHM_DEFINES
1279 #include <openssl/opensslconf.h>
1280 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
1281 [Richard Levitte, Ulf and Bodo Möller]
1283 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
1287 *) Change the 'other' type in certificate aux info to a STACK_OF
1288 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
1289 the required ASN1 format: arbitrary types determined by an OID.
1292 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
1293 argument to 'req'. This is not because the function is newer or
1294 better than others it just uses the work 'NEW' in the certificate
1295 request header lines. Some software needs this.
1298 *) Reorganise password command line arguments: now passwords can be
1299 obtained from various sources. Delete the PEM_cb function and make
1300 it the default behaviour: i.e. if the callback is NULL and the
1301 usrdata argument is not NULL interpret it as a null terminated pass
1302 phrase. If usrdata and the callback are NULL then the pass phrase
1303 is prompted for as usual.
1306 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
1307 the support is automatically enabled. The resulting binaries will
1308 autodetect the card and use it if present.
1309 [Ben Laurie and Compaq Inc.]
1311 *) Work around for Netscape hang bug. This sends certificate request
1312 and server done in one record. Since this is perfectly legal in the
1313 SSL/TLS protocol it isn't a "bug" option and is on by default. See
1314 the bugs/SSLv3 entry for more info.
1317 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
1320 *) Add -rand argument to smime and pkcs12 applications and read/write
1324 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
1327 *) Add command line password options to the remaining applications.
1330 *) Bug fix for BN_div_recp() for numerators with an even number of
1334 *) More tests in bntest.c, and changed test_bn output.
1337 *) ./config recognizes MacOS X now.
1340 *) Bug fix for BN_div() when the first words of num and divsor are
1341 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
1344 *) Add support for various broken PKCS#8 formats, and command line
1345 options to produce them.
1348 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
1349 get temporary BIGNUMs from a BN_CTX.
1352 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
1356 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
1357 include a #define from the old name to the new. The original intent
1358 was that statically linked binaries could for example just call
1359 SSLeay_add_all_ciphers() to just add ciphers to the table and not
1360 link with digests. This never worked becayse SSLeay_add_all_digests()
1361 and SSLeay_add_all_ciphers() were in the same source file so calling
1362 one would link with the other. They are now in separate source files.
1365 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
1368 *) Use a less unusual form of the Miller-Rabin primality test (it used
1369 a binary algorithm for exponentiation integrated into the Miller-Rabin
1370 loop, our standard modexp algorithms are faster).
1373 *) Support for the EBCDIC character set completed.
1374 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
1376 *) Source code cleanups: use const where appropriate, eliminate casts,
1377 use void * instead of char * in lhash.
1380 *) Bugfix: ssl3_send_server_key_exchange was not restartable
1381 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
1382 this the server could overwrite ephemeral keys that the client
1386 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
1387 using 50 iterations of the Rabin-Miller test.
1389 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
1390 iterations of the Rabin-Miller test as required by the appendix
1391 to FIPS PUB 186[-1]) instead of DSA_is_prime.
1392 As BN_is_prime_fasttest includes trial division, DSA parameter
1393 generation becomes much faster.
1395 This implies a change for the callback functions in DSA_is_prime
1396 and DSA_generate_parameters: The callback function is called once
1397 for each positive witness in the Rabin-Miller test, not just
1398 occasionally in the inner loop; and the parameters to the
1399 callback function now provide an iteration count for the outer
1400 loop rather than for the current invocation of the inner loop.
1401 DSA_generate_parameters additionally can call the callback
1402 function with an 'iteration count' of -1, meaning that a
1403 candidate has passed the trial division test (when q is generated
1404 from an application-provided seed, trial division is skipped).
1407 *) New function BN_is_prime_fasttest that optionally does trial
1408 division before starting the Rabin-Miller test and has
1409 an additional BN_CTX * argument (whereas BN_is_prime always
1410 has to allocate at least one BN_CTX).
1411 'callback(1, -1, cb_arg)' is called when a number has passed the
1412 trial division stage.
1415 *) Fix for bug in CRL encoding. The validity dates weren't being handled
1419 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
1422 *) New function BN_pseudo_rand().
1425 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
1426 bignum version of BN_from_montgomery() with the working code from
1427 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
1431 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
1432 made it impossible to use the same SSL_SESSION data structure in
1433 SSL2 clients in multiple threads.
1436 *) The return value of RAND_load_file() no longer counts bytes obtained
1437 by stat(). RAND_load_file(..., -1) is new and uses the complete file
1438 to seed the PRNG (previously an explicit byte count was required).
1439 [Ulf Möller, Bodo Möller]
1441 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
1442 used (char *) instead of (void *) and had casts all over the place.
1445 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
1448 *) Retain source code compatibility for BN_prime_checks macro:
1449 BN_is_prime(..., BN_prime_checks, ...) now uses
1450 BN_prime_checks_for_size to determine the appropriate number of
1451 Rabin-Miller iterations.
1454 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
1455 DH_CHECK_P_NOT_SAFE_PRIME.
1456 (Check if this is true? OpenPGP calls them "strong".)
1459 *) Merge the functionality of "dh" and "gendh" programs into a new program
1460 "dhparam". The old programs are retained for now but will handle DH keys
1461 (instead of parameters) in future.
1464 *) Make the ciphers, s_server and s_client programs check the return values
1465 when a new cipher list is set.
1468 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
1469 ciphers. Before when the 56bit ciphers were enabled the sorting was
1472 The syntax for the cipher sorting has been extended to support sorting by
1473 cipher-strength (using the strength_bits hard coded in the tables).
1474 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
1476 Fix a bug in the cipher-command parser: when supplying a cipher command
1477 string with an "undefined" symbol (neither command nor alphanumeric
1478 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
1479 an error is flagged.
1481 Due to the strength-sorting extension, the code of the
1482 ssl_create_cipher_list() function was completely rearranged. I hope that
1483 the readability was also increased :-)
1484 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
1486 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
1487 for the first serial number and places 2 in the serial number file. This
1488 avoids problems when the root CA is created with serial number zero and
1489 the first user certificate has the same issuer name and serial number
1493 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
1494 the new code. Add documentation for this stuff.
1497 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
1498 X509_*() to X509at_*() on the grounds that they don't handle X509
1499 structures and behave in an analagous way to the X509v3 functions:
1500 they shouldn't be called directly but wrapper functions should be used
1503 So we also now have some wrapper functions that call the X509at functions
1504 when passed certificate requests. (TO DO: similar things can be done with
1505 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
1506 things. Some of these need some d2i or i2d and print functionality
1507 because they handle more complex structures.)
1510 *) Add missing #ifndefs that caused missing symbols when building libssl
1511 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
1512 NO_RSA in ssl/s2*.c.
1513 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
1515 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
1516 has a return value which indicates the quality of the random data
1517 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
1518 error queue. New function RAND_pseudo_bytes() generates output that is
1519 guaranteed to be unique but not unpredictable. RAND_add is like
1520 RAND_seed, but takes an extra argument for an entropy estimate
1521 (RAND_seed always assumes full entropy).
1524 *) Do more iterations of Rabin-Miller probable prime test (specifically,
1525 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
1526 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
1527 in crypto/bn/bn_prime.c for the complete table). This guarantees a
1528 false-positive rate of at most 2^-80 for random input.
1531 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
1534 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
1535 in the 0.9.5 release), this returns the chain
1536 from an X509_CTX structure with a dup of the stack and all
1537 the X509 reference counts upped: so the stack will exist
1538 after X509_CTX_cleanup() has been called. Modify pkcs12.c
1541 Also make SSL_SESSION_print() print out the verify return
1545 *) Add manpage for the pkcs12 command. Also change the default
1546 behaviour so MAC iteration counts are used unless the new
1547 -nomaciter option is used. This improves file security and
1548 only older versions of MSIE (4.0 for example) need it.
1551 *) Honor the no-xxx Configure options when creating .DEF files.
1554 *) Add PKCS#10 attributes to field table: challengePassword,
1555 unstructuredName and unstructuredAddress. These are taken from
1556 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
1557 international characters are used.
1559 More changes to X509_ATTRIBUTE code: allow the setting of types
1560 based on strings. Remove the 'loc' parameter when adding
1561 attributes because these will be a SET OF encoding which is sorted
1565 *) Initial changes to the 'req' utility to allow request generation
1566 automation. This will allow an application to just generate a template
1567 file containing all the field values and have req construct the
1570 Initial support for X509_ATTRIBUTE handling. Stacks of these are
1571 used all over the place including certificate requests and PKCS#7
1572 structures. They are currently handled manually where necessary with
1573 some primitive wrappers for PKCS#7. The new functions behave in a
1574 manner analogous to the X509 extension functions: they allow
1575 attributes to be looked up by NID and added.
1577 Later something similar to the X509V3 code would be desirable to
1578 automatically handle the encoding, decoding and printing of the
1579 more complex types. The string types like challengePassword can
1580 be handled by the string table functions.
1582 Also modified the multi byte string table handling. Now there is
1583 a 'global mask' which masks out certain types. The table itself
1584 can use the flag STABLE_NO_MASK to ignore the mask setting: this
1585 is useful when for example there is only one permissible type
1586 (as in countryName) and using the mask might result in no valid
1590 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
1591 SSL_get_peer_finished to allow applications to obtain the latest
1592 Finished messages sent to the peer or expected from the peer,
1593 respectively. (SSL_get_peer_finished is usually the Finished message
1594 actually received from the peer, otherwise the protocol will be aborted.)
1596 As the Finished message are message digests of the complete handshake
1597 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
1598 be used for external authentication procedures when the authentication
1599 provided by SSL/TLS is not desired or is not enough.
1602 *) Enhanced support for Alpha Linux is added. Now ./config checks if
1603 the host supports BWX extension and if Compaq C is present on the
1604 $PATH. Just exploiting of the BWX extension results in 20-30%
1605 performance kick for some algorithms, e.g. DES and RC4 to mention
1606 a couple. Compaq C in turn generates ~20% faster code for MD5 and
1610 *) Add support for MS "fast SGC". This is arguably a violation of the
1611 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
1612 weak crypto and after checking the certificate is SGC a second one
1613 with strong crypto. MS SGC stops the first handshake after receiving
1614 the server certificate message and sends a second client hello. Since
1615 a server will typically do all the time consuming operations before
1616 expecting any further messages from the client (server key exchange
1617 is the most expensive) there is little difference between the two.
1619 To get OpenSSL to support MS SGC we have to permit a second client
1620 hello message after we have sent server done. In addition we have to
1621 reset the MAC if we do get this second client hello.
1624 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
1625 if a DER encoded private key is RSA or DSA traditional format. Changed
1626 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
1627 format DER encoded private key. Newer code should use PKCS#8 format which
1628 has the key type encoded in the ASN1 structure. Added DER private key
1629 support to pkcs8 application.
1632 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
1633 ciphersuites has been selected (as required by the SSL 3/TLS 1
1634 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
1635 is set, we interpret this as a request to violate the specification
1636 (the worst that can happen is a handshake failure, and 'correct'
1637 behaviour would result in a handshake failure anyway).
1640 *) In SSL_CTX_add_session, take into account that there might be multiple
1641 SSL_SESSION structures with the same session ID (e.g. when two threads
1642 concurrently obtain them from an external cache).
1643 The internal cache can handle only one SSL_SESSION with a given ID,
1644 so if there's a conflict, we now throw out the old one to achieve
1648 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
1649 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
1650 some routines that use cipher OIDs: some ciphers do not have OIDs
1651 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
1655 *) Simplify the trust setting structure and code. Now we just have
1656 two sequences of OIDs for trusted and rejected settings. These will
1657 typically have values the same as the extended key usage extension
1658 and any application specific purposes.
1660 The trust checking code now has a default behaviour: it will just
1661 check for an object with the same NID as the passed id. Functions can
1662 be provided to override either the default behaviour or the behaviour
1663 for a given id. SSL client, server and email already have functions
1664 in place for compatibility: they check the NID and also return "trusted"
1665 if the certificate is self signed.
1668 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
1669 traditional format into an EVP_PKEY structure.
1672 *) Add a password callback function PEM_cb() which either prompts for
1673 a password if usr_data is NULL or otherwise assumes it is a null
1674 terminated password. Allow passwords to be passed on command line
1675 environment or config files in a few more utilities.
1678 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
1679 keys. Add some short names for PKCS#8 PBE algorithms and allow them
1680 to be specified on the command line for the pkcs8 and pkcs12 utilities.
1681 Update documentation.
1684 *) Support for ASN1 "NULL" type. This could be handled before by using
1685 ASN1_TYPE but there wasn't any function that would try to read a NULL
1686 and produce an error if it couldn't. For compatibility we also have
1687 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
1688 don't allocate anything because they don't need to.
1691 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
1693 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
1695 *) Rebuild of the memory allocation routines used by OpenSSL code and
1696 possibly others as well. The purpose is to make an interface that
1697 provide hooks so anyone can build a separate set of allocation and
1698 deallocation routines to be used by OpenSSL, for example memory
1699 pool implementations, or something else, which was previously hard
1700 since Malloc(), Realloc() and Free() were defined as macros having
1701 the values malloc, realloc and free, respectively (except for Win32
1702 compilations). The same is provided for memory debugging code.
1703 OpenSSL already comes with functionality to find memory leaks, but
1704 this gives people a chance to debug other memory problems.
1706 With these changes, a new set of functions and macros have appeared:
1708 CRYPTO_set_mem_debug_functions() [F]
1709 CRYPTO_get_mem_debug_functions() [F]
1710 CRYPTO_dbg_set_options() [F]
1711 CRYPTO_dbg_get_options() [F]
1712 CRYPTO_malloc_debug_init() [M]
1714 The memory debug functions are NULL by default, unless the library
1715 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
1716 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
1717 gives the standard debugging functions that come with OpenSSL) or
1718 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
1719 provided by the library user) must be used. When the standard
1720 debugging functions are used, CRYPTO_dbg_set_options can be used to
1721 request additional information:
1722 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
1723 the CRYPTO_MDEBUG_xxx macro when compiling the library.
1725 Also, things like CRYPTO_set_mem_functions will always give the
1726 expected result (the new set of functions is used for allocation
1727 and deallocation) at all times, regardless of platform and compiler
1730 To finish it up, some functions that were never use in any other
1731 way than through macros have a new API and new semantic:
1734 CRYPTO_dbg_realloc()
1737 All macros of value have retained their old syntax.
1738 [Richard Levitte and Bodo Moeller]
1740 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
1741 ordering of SMIMECapabilities wasn't in "strength order" and there
1742 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
1746 *) Some ASN1 types with illegal zero length encoding (INTEGER,
1747 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
1748 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
1750 *) Merge in my S/MIME library for OpenSSL. This provides a simple
1751 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
1752 functionality to handle multipart/signed properly) and a utility
1753 called 'smime' to call all this stuff. This is based on code I
1754 originally wrote for Celo who have kindly allowed it to be
1755 included in OpenSSL.
1758 *) Add variants des_set_key_checked and des_set_key_unchecked of
1759 des_set_key (aka des_key_sched). Global variable des_check_key
1760 decides which of these is called by des_set_key; this way
1761 des_check_key behaves as it always did, but applications and
1762 the library itself, which was buggy for des_check_key == 1,
1763 have a cleaner way to pick the version they need.
1766 *) New function PKCS12_newpass() which changes the password of a
1770 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
1771 dynamic mix. In both cases the ids can be used as an index into the
1772 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
1773 functions so they accept a list of the field values and the
1774 application doesn't need to directly manipulate the X509_TRUST
1778 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
1782 *) Modify the way the V3 extension code looks up extensions. This now
1783 works in a similar way to the object code: we have some "standard"
1784 extensions in a static table which is searched with OBJ_bsearch()
1785 and the application can add dynamic ones if needed. The file
1786 crypto/x509v3/ext_dat.h now has the info: this file needs to be
1787 updated whenever a new extension is added to the core code and kept
1788 in ext_nid order. There is a simple program 'tabtest.c' which checks
1789 this. New extensions are not added too often so this file can readily
1790 be maintained manually.
1792 There are two big advantages in doing things this way. The extensions
1793 can be looked up immediately and no longer need to be "added" using
1794 X509V3_add_standard_extensions(): this function now does nothing.
1795 [Side note: I get *lots* of email saying the extension code doesn't
1796 work because people forget to call this function]
1797 Also no dynamic allocation is done unless new extensions are added:
1798 so if we don't add custom extensions there is no need to call
1799 X509V3_EXT_cleanup().
1802 *) Modify enc utility's salting as follows: make salting the default. Add a
1803 magic header, so unsalted files fail gracefully instead of just decrypting
1804 to garbage. This is because not salting is a big security hole, so people
1805 should be discouraged from doing it.
1808 *) Fixes and enhancements to the 'x509' utility. It allowed a message
1809 digest to be passed on the command line but it only used this
1810 parameter when signing a certificate. Modified so all relevant
1811 operations are affected by the digest parameter including the
1812 -fingerprint and -x509toreq options. Also -x509toreq choked if a
1813 DSA key was used because it didn't fix the digest.
1816 *) Initial certificate chain verify code. Currently tests the untrusted
1817 certificates for consistency with the verify purpose (which is set
1818 when the X509_STORE_CTX structure is set up) and checks the pathlength.
1820 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
1821 this is because it will reject chains with invalid extensions whereas
1822 every previous version of OpenSSL and SSLeay made no checks at all.
1824 Trust code: checks the root CA for the relevant trust settings. Trust
1825 settings have an initial value consistent with the verify purpose: e.g.
1826 if the verify purpose is for SSL client use it expects the CA to be
1827 trusted for SSL client use. However the default value can be changed to
1828 permit custom trust settings: one example of this would be to only trust
1829 certificates from a specific "secure" set of CAs.
1831 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
1832 which should be used for version portability: especially since the
1833 verify structure is likely to change more often now.
1835 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
1836 to set them. If not set then assume SSL clients will verify SSL servers
1839 Two new options to the verify program: -untrusted allows a set of
1840 untrusted certificates to be passed in and -purpose which sets the
1841 intended purpose of the certificate. If a purpose is set then the
1842 new chain verify code is used to check extension consistency.
1845 *) Support for the authority information access extension.
1848 *) Modify RSA and DSA PEM read routines to transparently handle
1849 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
1850 public keys in a format compatible with certificate
1851 SubjectPublicKeyInfo structures. Unfortunately there were already
1852 functions called *_PublicKey_* which used various odd formats so
1853 these are retained for compatibility: however the DSA variants were
1854 never in a public release so they have been deleted. Changed dsa/rsa
1855 utilities to handle the new format: note no releases ever handled public
1856 keys so we should be OK.
1858 The primary motivation for this change is to avoid the same fiasco
1859 that dogs private keys: there are several incompatible private key
1860 formats some of which are standard and some OpenSSL specific and
1861 require various evil hacks to allow partial transparent handling and
1862 even then it doesn't work with DER formats. Given the option anything
1863 other than PKCS#8 should be dumped: but the other formats have to
1864 stay in the name of compatibility.
1866 With public keys and the benefit of hindsight one standard format
1867 is used which works with EVP_PKEY, RSA or DSA structures: though
1868 it clearly returns an error if you try to read the wrong kind of key.
1870 Added a -pubkey option to the 'x509' utility to output the public key.
1871 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
1872 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
1873 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
1874 that do the same as the EVP_PKEY_assign_*() except they up the
1875 reference count of the added key (they don't "swallow" the
1879 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
1880 CRLs would fail if the file contained no certificates or no CRLs:
1881 added a new function to read in both types and return the number
1882 read: this means that if none are read it will be an error. The
1883 DER versions of the certificate and CRL reader would always fail
1884 because it isn't possible to mix certificates and CRLs in DER format
1885 without choking one or the other routine. Changed this to just read
1886 a certificate: this is the best we can do. Also modified the code
1887 in apps/verify.c to take notice of return codes: it was previously
1888 attempting to read in certificates from NULL pointers and ignoring
1889 any errors: this is one reason why the cert and CRL reader seemed
1890 to work. It doesn't check return codes from the default certificate
1891 routines: these may well fail if the certificates aren't installed.
1894 *) Code to support otherName option in GeneralName.
1897 *) First update to verify code. Change the verify utility
1898 so it warns if it is passed a self signed certificate:
1899 for consistency with the normal behaviour. X509_verify
1900 has been modified to it will now verify a self signed
1901 certificate if *exactly* the same certificate appears
1902 in the store: it was previously impossible to trust a
1903 single self signed certificate. This means that:
1904 openssl verify ss.pem
1905 now gives a warning about a self signed certificate but
1906 openssl verify -CAfile ss.pem ss.pem
1910 *) For servers, store verify_result in SSL_SESSION data structure
1911 (and add it to external session representation).
1912 This is needed when client certificate verifications fails,
1913 but an application-provided verification callback (set by
1914 SSL_CTX_set_cert_verify_callback) allows accepting the session
1915 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
1916 but returns 1): When the session is reused, we have to set
1917 ssl->verify_result to the appropriate error code to avoid
1919 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
1921 *) Fix a bug in the new PKCS#7 code: it didn't consider the
1922 case in PKCS7_dataInit() where the signed PKCS7 structure
1923 didn't contain any existing data because it was being created.
1924 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
1926 *) Add a salt to the key derivation routines in enc.c. This
1927 forms the first 8 bytes of the encrypted file. Also add a
1928 -S option to allow a salt to be input on the command line.
1931 *) New function X509_cmp(). Oddly enough there wasn't a function
1932 to compare two certificates. We do this by working out the SHA1
1933 hash and comparing that. X509_cmp() will be needed by the trust
1937 *) SSL_get1_session() is like SSL_get_session(), but increments
1938 the reference count in the SSL_SESSION returned.
1939 [Geoff Thorpe <geoff@eu.c2.net>]
1941 *) Fix for 'req': it was adding a null to request attributes.
1942 Also change the X509_LOOKUP and X509_INFO code to handle
1943 certificate auxiliary information.
1946 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
1950 *) Add the possibility to add extra information to the memory leak
1951 detecting output, to form tracebacks, showing from where each
1952 allocation was originated: CRYPTO_push_info("constant string") adds
1953 the string plus current file name and line number to a per-thread
1954 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
1955 is like calling CYRPTO_pop_info() until the stack is empty.
1956 Also updated memory leak detection code to be multi-thread-safe.
1959 *) Add options -text and -noout to pkcs7 utility and delete the
1960 encryption options which never did anything. Update docs.
1963 *) Add options to some of the utilities to allow the pass phrase
1964 to be included on either the command line (not recommended on
1965 OSes like Unix) or read from the environment. Update the
1966 manpages and fix a few bugs.
1969 *) Add a few manpages for some of the openssl commands.
1972 *) Fix the -revoke option in ca. It was freeing up memory twice,
1973 leaking and not finding already revoked certificates.
1976 *) Extensive changes to support certificate auxiliary information.
1977 This involves the use of X509_CERT_AUX structure and X509_AUX
1978 functions. An X509_AUX function such as PEM_read_X509_AUX()
1979 can still read in a certificate file in the usual way but it
1980 will also read in any additional "auxiliary information". By
1981 doing things this way a fair degree of compatibility can be
1982 retained: existing certificates can have this information added
1983 using the new 'x509' options.
1985 Current auxiliary information includes an "alias" and some trust
1986 settings. The trust settings will ultimately be used in enhanced
1987 certificate chain verification routines: currently a certificate
1988 can only be trusted if it is self signed and then it is trusted
1992 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
1993 The problem was that one of the replacement routines had not been working
1994 since SSLeay releases. For now the offending routine has been replaced
1995 with non-optimised assembler. Even so, this now gives around 95%
1996 performance improvement for 1024 bit RSA signs.
1999 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
2000 handling. Most clients have the effective key size in bits equal to
2001 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
2002 A few however don't do this and instead use the size of the decrypted key
2003 to determine the RC2 key length and the AlgorithmIdentifier to determine
2004 the effective key length. In this case the effective key length can still
2005 be 40 bits but the key length can be 168 bits for example. This is fixed
2006 by manually forcing an RC2 key into the EVP_PKEY structure because the
2007 EVP code can't currently handle unusual RC2 key sizes: it always assumes
2008 the key length and effective key length are equal.
2011 *) Add a bunch of functions that should simplify the creation of
2012 X509_NAME structures. Now you should be able to do:
2013 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
2014 and have it automatically work out the correct field type and fill in
2015 the structures. The more adventurous can try:
2016 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
2017 and it will (hopefully) work out the correct multibyte encoding.
2020 *) Change the 'req' utility to use the new field handling and multibyte
2021 copy routines. Before the DN field creation was handled in an ad hoc
2022 way in req, ca, and x509 which was rather broken and didn't support
2023 BMPStrings or UTF8Strings. Since some software doesn't implement
2024 BMPStrings or UTF8Strings yet, they can be enabled using the config file
2025 using the dirstring_type option. See the new comment in the default
2026 openssl.cnf for more info.
2029 *) Make crypto/rand/md_rand.c more robust:
2030 - Assure unique random numbers after fork().
2031 - Make sure that concurrent threads access the global counter and
2032 md serializably so that we never lose entropy in them
2033 or use exactly the same state in multiple threads.
2034 Access to the large state is not always serializable because
2035 the additional locking could be a performance killer, and
2036 md should be large enough anyway.
2039 *) New file apps/app_rand.c with commonly needed functionality
2040 for handling the random seed file.
2042 Use the random seed file in some applications that previously did not:
2044 dsaparam -genkey (which also ignored its '-rand' option),
2047 x509 (when signing).
2048 Except on systems with /dev/urandom, it is crucial to have a random
2049 seed file at least for key creation, DSA signing, and for DH exchanges;
2050 for RSA signatures we could do without one.
2052 gendh and gendsa (unlike genrsa) used to read only the first byte
2053 of each file listed in the '-rand' option. The function as previously
2054 found in genrsa is now in app_rand.c and is used by all programs
2055 that support '-rand'.
2058 *) In RAND_write_file, use mode 0600 for creating files;
2059 don't just chmod when it may be too late.
2062 *) Report an error from X509_STORE_load_locations
2063 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
2066 *) New function ASN1_mbstring_copy() this copies a string in either
2067 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
2068 into an ASN1_STRING type. A mask of permissible types is passed
2069 and it chooses the "minimal" type to use or an error if not type
2073 *) Add function equivalents to the various macros in asn1.h. The old
2074 macros are retained with an M_ prefix. Code inside the library can
2075 use the M_ macros. External code (including the openssl utility)
2076 should *NOT* in order to be "shared library friendly".
2079 *) Add various functions that can check a certificate's extensions
2080 to see if it usable for various purposes such as SSL client,
2081 server or S/MIME and CAs of these types. This is currently
2082 VERY EXPERIMENTAL but will ultimately be used for certificate chain
2083 verification. Also added a -purpose flag to x509 utility to
2084 print out all the purposes.
2087 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
2091 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
2092 for, obtain and decode and extension and obtain its critical flag.
2093 This allows all the necessary extension code to be handled in a
2094 single function call.
2097 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
2098 platforms. See crypto/rc4/rc4_enc.c for further details.
2101 *) New -noout option to asn1parse. This causes no output to be produced
2102 its main use is when combined with -strparse and -out to extract data
2103 from a file (which may not be in ASN.1 format).
2106 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
2107 when producing the local key id.
2108 [Richard Levitte <levitte@stacken.kth.se>]
2110 *) New option -dhparam in s_server. This allows a DH parameter file to be
2111 stated explicitly. If it is not stated then it tries the first server
2112 certificate file. The previous behaviour hard coded the filename
2116 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
2117 a public key to be input or output. For example:
2118 openssl rsa -in key.pem -pubout -out pubkey.pem
2119 Also added necessary DSA public key functions to handle this.
2122 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
2123 in the message. This was handled by allowing
2124 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
2125 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
2127 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
2128 to the end of the strings whereas this didn't. This would cause problems
2129 if strings read with d2i_ASN1_bytes() were later modified.
2130 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
2132 *) Fix for base64 decode bug. When a base64 bio reads only one line of
2133 data and it contains EOF it will end up returning an error. This is
2134 caused by input 46 bytes long. The cause is due to the way base64
2135 BIOs find the start of base64 encoded data. They do this by trying a
2136 trial decode on each line until they find one that works. When they
2137 do a flag is set and it starts again knowing it can pass all the
2138 data directly through the decoder. Unfortunately it doesn't reset
2139 the context it uses. This means that if EOF is reached an attempt
2140 is made to pass two EOFs through the context and this causes the
2141 resulting error. This can also cause other problems as well. As is
2142 usual with these problems it takes *ages* to find and the fix is
2143 trivial: move one line.
2144 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
2146 *) Ugly workaround to get s_client and s_server working under Windows. The
2147 old code wouldn't work because it needed to select() on sockets and the
2148 tty (for keypresses and to see if data could be written). Win32 only
2149 supports select() on sockets so we select() with a 1s timeout on the
2150 sockets and then see if any characters are waiting to be read, if none
2151 are present then we retry, we also assume we can always write data to
2152 the tty. This isn't nice because the code then blocks until we've
2153 received a complete line of data and it is effectively polling the
2154 keyboard at 1s intervals: however it's quite a bit better than not
2155 working at all :-) A dedicated Windows application might handle this
2156 with an event loop for example.
2159 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
2160 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
2161 will be called when RSA_sign() and RSA_verify() are used. This is useful
2162 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
2163 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
2164 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
2165 This necessitated the support of an extra signature type NID_md5_sha1
2166 for SSL signatures and modifications to the SSL library to use it instead
2167 of calling RSA_public_decrypt() and RSA_private_encrypt().
2170 *) Add new -verify -CAfile and -CApath options to the crl program, these
2171 will lookup a CRL issuers certificate and verify the signature in a
2172 similar way to the verify program. Tidy up the crl program so it
2173 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
2174 less strict. It will now permit CRL extensions even if it is not
2175 a V2 CRL: this will allow it to tolerate some broken CRLs.
2178 *) Initialize all non-automatic variables each time one of the openssl
2179 sub-programs is started (this is necessary as they may be started
2180 multiple times from the "OpenSSL>" prompt).
2181 [Lennart Bang, Bodo Moeller]
2183 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
2184 removing all other RSA functionality (this is what NO_RSA does). This
2185 is so (for example) those in the US can disable those operations covered
2186 by the RSA patent while allowing storage and parsing of RSA keys and RSA
2190 *) Non-copying interface to BIO pairs.
2191 (still largely untested)
2194 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
2195 ASCII string. This was handled independently in various places before.
2198 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
2199 UTF8 strings a character at a time.
2202 *) Use client_version from client hello to select the protocol
2203 (s23_srvr.c) and for RSA client key exchange verification
2204 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
2207 *) Add various utility functions to handle SPKACs, these were previously
2208 handled by poking round in the structure internals. Added new function
2209 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
2210 print, verify and generate SPKACs. Based on an original idea from
2211 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
2214 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
2217 *) Allow the config file extension section to be overwritten on the
2218 command line. Based on an original idea from Massimiliano Pala
2219 <madwolf@comune.modena.it>. The new option is called -extensions
2220 and can be applied to ca, req and x509. Also -reqexts to override
2221 the request extensions in req and -crlexts to override the crl extensions
2225 *) Add new feature to the SPKAC handling in ca. Now you can include
2226 the same field multiple times by preceding it by "XXXX." for example:
2229 this is the same syntax as used in the req config file.
2232 *) Allow certificate extensions to be added to certificate requests. These
2233 are specified in a 'req_extensions' option of the req section of the
2234 config file. They can be printed out with the -text option to req but
2235 are otherwise ignored at present.
2238 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
2239 data read consists of only the final block it would not decrypted because
2240 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
2241 A misplaced 'break' also meant the decrypted final block might not be
2242 copied until the next read.
2245 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
2246 a few extra parameters to the DH structure: these will be useful if
2247 for example we want the value of 'q' or implement X9.42 DH.
2250 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
2251 provides hooks that allow the default DSA functions or functions on a
2252 "per key" basis to be replaced. This allows hardware acceleration and
2253 hardware key storage to be handled without major modification to the
2254 library. Also added low level modexp hooks and CRYPTO_EX structure and
2255 associated functions.
2258 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
2259 as "read only": it can't be written to and the buffer it points to will
2260 not be freed. Reading from a read only BIO is much more efficient than
2261 a normal memory BIO. This was added because there are several times when
2262 an area of memory needs to be read from a BIO. The previous method was
2263 to create a memory BIO and write the data to it, this results in two
2264 copies of the data and an O(n^2) reading algorithm. There is a new
2265 function BIO_new_mem_buf() which creates a read only memory BIO from
2266 an area of memory. Also modified the PKCS#7 routines to use read only
2270 *) Bugfix: ssl23_get_client_hello did not work properly when called in
2271 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
2272 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
2273 but a retry condition occured while trying to read the rest.
2276 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
2277 NID_pkcs7_encrypted by default: this was wrong since this should almost
2278 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
2279 the encrypted data type: this is a more sensible place to put it and it
2280 allows the PKCS#12 code to be tidied up that duplicated this
2284 *) Changed obj_dat.pl script so it takes its input and output files on
2285 the command line. This should avoid shell escape redirection problems
2289 *) Initial support for certificate extension requests, these are included
2290 in things like Xenroll certificate requests. Included functions to allow
2291 extensions to be obtained and added.
2294 *) -crlf option to s_client and s_server for sending newlines as
2295 CRLF (as required by many protocols).
2298 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
2300 *) Install libRSAglue.a when OpenSSL is built with RSAref.
2301 [Ralf S. Engelschall]
2303 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
2304 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
2306 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
2310 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
2311 DH parameters/keys (q is lost during that conversion, but the resulting
2312 DH parameters contain its length).
2314 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
2315 much faster than DH_generate_parameters (which creates parameters
2316 where p = 2*q + 1), and also the smaller q makes DH computations
2317 much more efficient (160-bit exponentiation instead of 1024-bit
2318 exponentiation); so this provides a convenient way to support DHE
2319 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
2320 utter importance to use
2321 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
2323 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
2324 when such DH parameters are used, because otherwise small subgroup
2325 attacks may become possible!
2328 *) Avoid memory leak in i2d_DHparams.
2331 *) Allow the -k option to be used more than once in the enc program:
2332 this allows the same encrypted message to be read by multiple recipients.
2335 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
2336 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
2337 it will always use the numerical form of the OID, even if it has a short
2341 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
2342 method only got called if p,q,dmp1,dmq1,iqmp components were present,
2343 otherwise bn_mod_exp was called. In the case of hardware keys for example
2344 no private key components need be present and it might store extra data
2345 in the RSA structure, which cannot be accessed from bn_mod_exp.
2346 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
2347 private key operations.
2350 *) Added support for SPARC Linux.
2353 *) pem_password_cb function type incompatibly changed from
2354 typedef int pem_password_cb(char *buf, int size, int rwflag);
2356 ....(char *buf, int size, int rwflag, void *userdata);
2357 so that applications can pass data to their callbacks:
2358 The PEM[_ASN1]_{read,write}... functions and macros now take an
2359 additional void * argument, which is just handed through whenever
2360 the password callback is called.
2361 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
2363 New function SSL_CTX_set_default_passwd_cb_userdata.
2365 Compatibility note: As many C implementations push function arguments
2366 onto the stack in reverse order, the new library version is likely to
2367 interoperate with programs that have been compiled with the old
2368 pem_password_cb definition (PEM_whatever takes some data that
2369 happens to be on the stack as its last argument, and the callback
2370 just ignores this garbage); but there is no guarantee whatsoever that
2373 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
2374 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
2375 problems not only on Windows, but also on some Unix platforms.
2376 To avoid problematic command lines, these definitions are now in an
2377 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
2378 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
2381 *) MIPS III/IV assembler module is reimplemented.
2384 *) More DES library cleanups: remove references to srand/rand and
2385 delete an unused file.
2388 *) Add support for the the free Netwide assembler (NASM) under Win32,
2389 since not many people have MASM (ml) and it can be hard to obtain.
2390 This is currently experimental but it seems to work OK and pass all
2391 the tests. Check out INSTALL.W32 for info.
2394 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
2395 without temporary keys kept an extra copy of the server key,
2396 and connections with temporary keys did not free everything in case
2400 *) New function RSA_check_key and new openssl rsa option -check
2401 for verifying the consistency of RSA keys.
2402 [Ulf Moeller, Bodo Moeller]
2404 *) Various changes to make Win32 compile work:
2405 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
2406 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
2407 comparison" warnings.
2408 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
2411 *) Add a debugging option to PKCS#5 v2 key generation function: when
2412 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
2413 derived keys are printed to stderr.
2416 *) Copy the flags in ASN1_STRING_dup().
2417 [Roman E. Pavlov <pre@mo.msk.ru>]
2419 *) The x509 application mishandled signing requests containing DSA
2420 keys when the signing key was also DSA and the parameters didn't match.
2422 It was supposed to omit the parameters when they matched the signing key:
2423 the verifying software was then supposed to automatically use the CA's
2424 parameters if they were absent from the end user certificate.
2426 Omitting parameters is no longer recommended. The test was also
2427 the wrong way round! This was probably due to unusual behaviour in
2428 EVP_cmp_parameters() which returns 1 if the parameters match.
2429 This meant that parameters were omitted when they *didn't* match and
2430 the certificate was useless. Certificates signed with 'ca' didn't have
2432 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
2434 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
2435 The interface is as follows:
2436 Applications can use
2437 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
2438 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
2439 "off" is now the default.
2440 The library internally uses
2441 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
2442 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
2443 to disable memory-checking temporarily.
2445 Some inconsistent states that previously were possible (and were
2446 even the default) are now avoided.
2448 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
2449 with each memory chunk allocated; this is occasionally more helpful
2450 than just having a counter.
2452 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
2454 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
2458 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
2459 which largely parallels "options", but is for changing API behaviour,
2460 whereas "options" are about protocol behaviour.
2461 Initial "mode" flags are:
2463 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
2464 a single record has been written.
2465 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
2466 retries use the same buffer location.
2467 (But all of the contents must be
2471 *) Bugfix: SSL_set_mode ignored its parameter, only SSL_CTX_set_mode
2474 *) Fix problems with no-hmac etc.
2475 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
2477 *) New functions RSA_get_default_method(), RSA_set_method() and
2478 RSA_get_method(). These allows replacement of RSA_METHODs without having
2479 to mess around with the internals of an RSA structure.
2482 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
2483 Also really enable memory leak checks in openssl.c and in some
2485 [Chad C. Mulligan, Bodo Moeller]
2487 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
2488 up the length of negative integers. This has now been simplified to just
2489 store the length when it is first determined and use it later, rather
2490 than trying to keep track of where data is copied and updating it to
2492 [Steve Henson, reported by Brien Wheeler
2493 <bwheeler@authentica-security.com>]
2495 *) Add a new function PKCS7_signatureVerify. This allows the verification
2496 of a PKCS#7 signature but with the signing certificate passed to the
2497 function itself. This contrasts with PKCS7_dataVerify which assumes the
2498 certificate is present in the PKCS#7 structure. This isn't always the
2499 case: certificates can be omitted from a PKCS#7 structure and be
2500 distributed by "out of band" means (such as a certificate database).
2503 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
2504 function prototypes in pem.h, also change util/mkdef.pl to add the
2505 necessary function names.
2508 *) mk1mf.pl (used by Windows builds) did not properly read the
2509 options set by Configure in the top level Makefile, and Configure
2510 was not even able to write more than one option correctly.
2511 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
2514 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
2515 file to be loaded from a BIO or FILE pointer. The BIO version will
2516 for example allow memory BIOs to contain config info.
2519 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
2520 Whoever hopes to achieve shared-library compatibility across versions
2521 must use this, not the compile-time macro.
2522 (Exercise 0.9.4: Which is the minimum library version required by
2524 Note: All this applies only to multi-threaded programs, others don't
2528 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
2529 through a BIO pair triggered the default case, i.e.
2530 SSLerr(...,SSL_R_UNKNOWN_STATE).
2533 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
2534 can use the SSL library even if none of the specific BIOs is
2538 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
2539 for the encoded length.
2540 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
2542 *) Add initial documentation of the X509V3 functions.
2545 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
2546 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
2547 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
2548 secure PKCS#8 private key format with a high iteration count.
2551 *) Fix determination of Perl interpreter: A perl or perl5
2552 _directory_ in $PATH was also accepted as the interpreter.
2553 [Ralf S. Engelschall]
2555 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
2556 wrong with it but it was very old and did things like calling
2557 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
2561 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
2562 to use the new extension code.
2565 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
2566 with macros. This should make it easier to change their form, add extra
2567 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
2571 *) Add to configuration table a new entry that can specify an alternative
2572 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
2573 according to Mark Crispin <MRC@Panda.COM>.
2577 *) DES CBC did not update the IV. Weird.
2580 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
2581 Changing the behaviour of the former might break existing programs --
2582 where IV updating is needed, des_ncbc_encrypt can be used.
2585 *) When bntest is run from "make test" it drives bc to check its
2586 calculations, as well as internally checking them. If an internal check
2587 fails, it needs to cause bc to give a non-zero result or make test carries
2588 on without noticing the failure. Fixed.
2591 *) DES library cleanups.
2594 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
2595 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
2596 ciphers. NOTE: although the key derivation function has been verified
2597 against some published test vectors it has not been extensively tested
2598 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
2602 *) Instead of "mkdir -p", which is not fully portable, use new
2603 Perl script "util/mkdir-p.pl".
2606 *) Rewrite the way password based encryption (PBE) is handled. It used to
2607 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
2608 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
2609 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
2610 the 'parameter' field of the AlgorithmIdentifier is passed to the
2611 underlying key generation function so it must do its own ASN1 parsing.
2612 This has also changed the EVP_PBE_CipherInit() function which now has a
2613 'parameter' argument instead of literal salt and iteration count values
2614 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
2617 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
2618 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
2619 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
2620 KEY" because this clashed with PKCS#8 unencrypted string. Since this
2621 value was just used as a "magic string" and not used directly its
2622 value doesn't matter.
2625 *) Introduce some semblance of const correctness to BN. Shame C doesn't
2629 *) "linux-sparc64" configuration (ultrapenguin).
2630 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
2631 "linux-sparc" configuration.
2632 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
2634 *) config now generates no-xxx options for missing ciphers.
2637 *) Support the EBCDIC character set (work in progress).
2638 File ebcdic.c not yet included because it has a different license.
2639 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
2641 *) Support BS2000/OSD-POSIX.
2642 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
2644 *) Make callbacks for key generation use void * instead of char *.
2647 *) Make S/MIME samples compile (not yet tested).
2650 *) Additional typesafe stacks.
2653 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
2657 Changes between 0.9.3 and 0.9.3a [29 May 1999]
2659 *) New configuration variant "sco5-gcc".
2661 *) Updated some demos.
2662 [Sean O Riordain, Wade Scholine]
2664 *) Add missing BIO_free at exit of pkcs12 application.
2667 *) Fix memory leak in conf.c.
2670 *) Updates for Win32 to assembler version of MD5.
2673 *) Set #! path to perl in apps/der_chop to where we found it
2674 instead of using a fixed path.
2677 *) SHA library changes for irix64-mips4-cc.
2680 *) Improvements for VMS support.
2684 Changes between 0.9.2b and 0.9.3 [24 May 1999]
2686 *) Bignum library bug fix. IRIX 6 passes "make test" now!
2687 This also avoids the problems with SC4.2 and unpatched SC5.
2688 [Andy Polyakov <appro@fy.chalmers.se>]
2690 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
2691 These are required because of the typesafe stack would otherwise break
2692 existing code. If old code used a structure member which used to be STACK
2693 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
2694 sk_num or sk_value it would produce an error because the num, data members
2695 are not present in STACK_OF. Now it just produces a warning. sk_set
2696 replaces the old method of assigning a value to sk_value
2697 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
2698 that does this will no longer work (and should use sk_set instead) but
2699 this could be regarded as a "questionable" behaviour anyway.
2702 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
2703 correctly handle encrypted S/MIME data.
2706 *) Change type of various DES function arguments from des_cblock
2707 (which means, in function argument declarations, pointer to char)
2708 to des_cblock * (meaning pointer to array with 8 char elements),
2709 which allows the compiler to do more typechecking; it was like
2710 that back in SSLeay, but with lots of ugly casts.
2712 Introduce new type const_des_cblock.
2715 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
2716 problems: find RecipientInfo structure that matches recipient certificate
2717 and initialise the ASN1 structures properly based on passed cipher.
2720 *) Belatedly make the BN tests actually check the results.
2723 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
2724 to and from BNs: it was completely broken. New compilation option
2725 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
2726 key elements as negative integers.
2729 *) Reorganize and speed up MD5.
2730 [Andy Polyakov <appro@fy.chalmers.se>]
2733 [Richard Levitte <richard@levitte.org>]
2735 *) New option -out to asn1parse to allow the parsed structure to be
2736 output to a file. This is most useful when combined with the -strparse
2737 option to examine the output of things like OCTET STRINGS.
2740 *) Make SSL library a little more fool-proof by not requiring any longer
2741 that SSL_set_{accept,connect}_state be called before
2742 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
2743 in many applications because usually everything *appeared* to work as
2744 intended anyway -- now it really works as intended).
2747 *) Move openssl.cnf out of lib/.
2750 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
2751 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
2752 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
2753 [Ralf S. Engelschall]
2755 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
2756 handle PKCS#7 enveloped data properly.
2757 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
2759 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
2760 copying pointers. The cert_st handling is changed by this in
2761 various ways (and thus what used to be known as ctx->default_cert
2762 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
2763 any longer when s->cert does not give us what we need).
2764 ssl_cert_instantiate becomes obsolete by this change.
2765 As soon as we've got the new code right (possibly it already is?),
2766 we have solved a couple of bugs of the earlier code where s->cert
2767 was used as if it could not have been shared with other SSL structures.
2769 Note that using the SSL API in certain dirty ways now will result
2770 in different behaviour than observed with earlier library versions:
2771 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
2772 does not influence s as it used to.
2774 In order to clean up things more thoroughly, inside SSL_SESSION
2775 we don't use CERT any longer, but a new structure SESS_CERT
2776 that holds per-session data (if available); currently, this is
2777 the peer's certificate chain and, for clients, the server's certificate
2778 and temporary key. CERT holds only those values that can have
2779 meaningful defaults in an SSL_CTX.
2782 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
2783 from the internal representation. Various PKCS#7 fixes: remove some
2784 evil casts and set the enc_dig_alg field properly based on the signing
2788 *) Allow PKCS#12 password to be set from the command line or the
2789 environment. Let 'ca' get its config file name from the environment
2790 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
2794 *) Allow certificate policies extension to use an IA5STRING for the
2795 organization field. This is contrary to the PKIX definition but
2796 VeriSign uses it and IE5 only recognises this form. Document 'x509'
2800 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
2801 without disallowing inline assembler and the like for non-pedantic builds.
2804 *) Support Borland C++ builder.
2805 [Janez Jere <jj@void.si>, modified by Ulf Möller]
2810 *) SHA-1 cleanups and performance enhancements.
2811 [Andy Polyakov <appro@fy.chalmers.se>]
2813 *) Sparc v8plus assembler for the bignum library.
2814 [Andy Polyakov <appro@fy.chalmers.se>]
2816 *) Accept any -xxx and +xxx compiler options in Configure.
2819 *) Update HPUX configuration.
2822 *) Add missing sk_<type>_unshift() function to safestack.h
2823 [Ralf S. Engelschall]
2825 *) New function SSL_CTX_use_certificate_chain_file that sets the
2826 "extra_cert"s in addition to the certificate. (This makes sense
2827 only for "PEM" format files, as chains as a whole are not
2831 *) Support verify_depth from the SSL API.
2832 x509_vfy.c had what can be considered an off-by-one-error:
2833 Its depth (which was not part of the external interface)
2834 was actually counting the number of certificates in a chain;
2835 now it really counts the depth.
2838 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
2839 instead of X509err, which often resulted in confusing error
2840 messages since the error codes are not globally unique
2841 (e.g. an alleged error in ssl3_accept when a certificate
2842 didn't match the private key).
2844 *) New function SSL_CTX_set_session_id_context that allows to set a default
2845 value (so that you don't need SSL_set_session_id_context for each
2846 connection using the SSL_CTX).
2849 *) OAEP decoding bug fix.
2852 *) Support INSTALL_PREFIX for package builders, as proposed by
2856 *) New Configure options "threads" and "no-threads". For systems
2857 where the proper compiler options are known (currently Solaris
2858 and Linux), "threads" is the default.
2861 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
2864 *) Install various scripts to $(OPENSSLDIR)/misc, not to
2865 $(INSTALLTOP)/bin -- they shouldn't clutter directories
2866 such as /usr/local/bin.
2869 *) "make linux-shared" to build shared libraries.
2870 [Niels Poppe <niels@netbox.org>]
2872 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
2875 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
2876 extension adding in x509 utility.
2879 *) Remove NOPROTO sections and error code comments.
2882 *) Partial rewrite of the DEF file generator to now parse the ANSI
2886 *) New Configure options --prefix=DIR and --openssldir=DIR.
2889 *) Complete rewrite of the error code script(s). It is all now handled
2890 by one script at the top level which handles error code gathering,
2891 header rewriting and C source file generation. It should be much better
2892 than the old method: it now uses a modified version of Ulf's parser to
2893 read the ANSI prototypes in all header files (thus the old K&R definitions
2894 aren't needed for error creation any more) and do a better job of
2895 translating function codes into names. The old 'ASN1 error code imbedded
2896 in a comment' is no longer necessary and it doesn't use .err files which
2897 have now been deleted. Also the error code call doesn't have to appear all
2898 on one line (which resulted in some large lines...).
2901 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
2904 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
2905 0 (which usually indicates a closed connection), but continue reading.
2908 *) Fix some race conditions.
2911 *) Add support for CRL distribution points extension. Add Certificate
2912 Policies and CRL distribution points documentation.
2915 *) Move the autogenerated header file parts to crypto/opensslconf.h.
2918 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
2919 8 of keying material. Merlin has also confirmed interop with this fix
2920 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
2921 [Merlin Hughes <merlin@baltimore.ie>]
2923 *) Fix lots of warnings.
2924 [Richard Levitte <levitte@stacken.kth.se>]
2926 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
2927 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
2928 [Richard Levitte <levitte@stacken.kth.se>]
2930 *) Fix problems with sizeof(long) == 8.
2931 [Andy Polyakov <appro@fy.chalmers.se>]
2933 *) Change functions to ANSI C.
2936 *) Fix typos in error codes.
2937 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
2939 *) Remove defunct assembler files from Configure.
2942 *) SPARC v8 assembler BIGNUM implementation.
2943 [Andy Polyakov <appro@fy.chalmers.se>]
2945 *) Support for Certificate Policies extension: both print and set.
2946 Various additions to support the r2i method this uses.
2949 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
2950 return a const string when you are expecting an allocated buffer.
2953 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
2954 types DirectoryString and DisplayText.
2957 *) Add code to allow r2i extensions to access the configuration database,
2958 add an LHASH database driver and add several ctx helper functions.
2961 *) Fix an evil bug in bn_expand2() which caused various BN functions to
2962 fail when they extended the size of a BIGNUM.
2965 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
2966 support typesafe stack.
2969 *) Fix typo in SSL_[gs]et_options().
2970 [Nils Frostberg <nils@medcom.se>]
2972 *) Delete various functions and files that belonged to the (now obsolete)
2973 old X509V3 handling code.
2976 *) New Configure option "rsaref".
2979 *) Don't auto-generate pem.h.
2982 *) Introduce type-safe ASN.1 SETs.
2985 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
2986 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
2988 *) Introduce type-safe STACKs. This will almost certainly break lots of code
2989 that links with OpenSSL (well at least cause lots of warnings), but fear
2990 not: the conversion is trivial, and it eliminates loads of evil casts. A
2991 few STACKed things have been converted already. Feel free to convert more.
2992 In the fullness of time, I'll do away with the STACK type altogether.
2995 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
2996 specified in <certfile> by updating the entry in the index.txt file.
2997 This way one no longer has to edit the index.txt file manually for
2998 revoking a certificate. The -revoke option does the gory details now.
2999 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
3001 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
3002 `-text' option at all and this way the `-noout -text' combination was
3003 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
3004 [Ralf S. Engelschall]
3006 *) Make sure a corresponding plain text error message exists for the
3007 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
3008 verify callback function determined that a certificate was revoked.
3009 [Ralf S. Engelschall]
3011 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
3012 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
3013 all available cipers including rc5, which was forgotten until now.
3014 In order to let the testing shell script know which algorithms
3015 are available, a new (up to now undocumented) command
3016 "openssl list-cipher-commands" is used.
3019 *) Bugfix: s_client occasionally would sleep in select() when
3020 it should have checked SSL_pending() first.
3023 *) New functions DSA_do_sign and DSA_do_verify to provide access to
3024 the raw DSA values prior to ASN.1 encoding.
3027 *) Tweaks to Configure
3028 [Niels Poppe <niels@netbox.org>]
3030 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
3034 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
3037 *) New config option to avoid instructions that are illegal on the 80386.
3038 The default code is faster, but requires at least a 486.
3041 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
3042 SSL2_SERVER_VERSION (not used at all) macros, which are now the
3043 same as SSL2_VERSION anyway.
3046 *) New "-showcerts" option for s_client.
3049 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
3050 application. Various cleanups and fixes.
3053 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
3054 modify error routines to work internally. Add error codes and PBE init
3055 to library startup routines.
3058 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
3059 packing functions to asn1 and evp. Changed function names and error
3060 codes along the way.
3063 *) PKCS12 integration: and so it begins... First of several patches to
3064 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
3065 objects to objects.h
3068 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
3069 and display support for Thawte strong extranet extension.
3072 *) Add LinuxPPC support.
3073 [Jeff Dubrule <igor@pobox.org>]
3075 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
3076 bn_div_words in alpha.s.
3077 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
3079 *) Make sure the RSA OAEP test is skipped under -DRSAref because
3080 OAEP isn't supported when OpenSSL is built with RSAref.
3081 [Ulf Moeller <ulf@fitug.de>]
3083 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
3084 so they no longer are missing under -DNOPROTO.
3085 [Soren S. Jorvang <soren@t.dk>]
3088 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
3090 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
3091 doesn't work when the session is reused. Coming soon!
3094 *) Fix a security hole, that allows sessions to be reused in the wrong
3095 context thus bypassing client cert protection! All software that uses
3096 client certs and session caches in multiple contexts NEEDS PATCHING to
3097 allow session reuse! A fuller solution is in the works.
3098 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
3100 *) Some more source tree cleanups (removed obsolete files
3101 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
3102 permission on "config" script to be executable) and a fix for the INSTALL
3104 [Ulf Moeller <ulf@fitug.de>]
3106 *) Remove some legacy and erroneous uses of malloc, free instead of
3108 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
3110 *) Make rsa_oaep_test return non-zero on error.
3111 [Ulf Moeller <ulf@fitug.de>]
3113 *) Add support for native Solaris shared libraries. Configure
3114 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
3115 if someone would make that last step automatic.
3116 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
3118 *) ctx_size was not built with the right compiler during "make links". Fixed.
3121 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
3122 except NULL ciphers". This means the default cipher list will no longer
3123 enable NULL ciphers. They need to be specifically enabled e.g. with
3124 the string "DEFAULT:eNULL".
3127 *) Fix to RSA private encryption routines: if p < q then it would
3128 occasionally produce an invalid result. This will only happen with
3129 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
3132 *) Be less restrictive and allow also `perl util/perlpath.pl
3133 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
3134 because this way one can also use an interpreter named `perl5' (which is
3135 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
3136 installed as `perl').
3137 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
3139 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
3140 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
3142 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
3143 advapi32.lib to Win32 build and change the pem test comparision
3144 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
3145 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
3146 and crypto/des/ede_cbcm_enc.c.
3149 *) DES quad checksum was broken on big-endian architectures. Fixed.
3152 *) Comment out two functions in bio.h that aren't implemented. Fix up the
3153 Win32 test batch file so it (might) work again. The Win32 test batch file
3154 is horrible: I feel ill....
3157 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
3158 in e_os.h. Audit of header files to check ANSI and non ANSI
3159 sections: 10 functions were absent from non ANSI section and not exported
3160 from Windows DLLs. Fixed up libeay.num for new functions.
3163 *) Make `openssl version' output lines consistent.
3164 [Ralf S. Engelschall]
3166 *) Fix Win32 symbol export lists for BIO functions: Added
3167 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
3168 to ms/libeay{16,32}.def.
3169 [Ralf S. Engelschall]
3171 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
3172 fine under Unix and passes some trivial tests I've now added. But the
3173 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
3174 added to make sure no one expects that this stuff really works in the
3175 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
3176 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
3178 [Ralf S. Engelschall]
3180 *) Fix the generation of two part addresses in perl.
3181 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
3183 *) Add config entry for Linux on MIPS.
3184 [John Tobey <jtobey@channel1.com>]
3186 *) Make links whenever Configure is run, unless we are on Windoze.
3189 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
3190 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
3194 *) Add a useful kludge to allow package maintainers to specify compiler and
3195 other platforms details on the command line without having to patch the
3196 Configure script everytime: One now can use ``perl Configure
3197 <id>:<details>'', i.e. platform ids are allowed to have details appended
3198 to them (seperated by colons). This is treated as there would be a static
3199 pre-configured entry in Configure's %table under key <id> with value
3200 <details> and ``perl Configure <id>'' is called. So, when you want to
3201 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
3202 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
3203 now, which overrides the FreeBSD-elf entry on-the-fly.
3204 [Ralf S. Engelschall]
3206 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
3209 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
3210 on the `perl Configure ...' command line. This way one can compile
3211 OpenSSL libraries with Position Independent Code (PIC) which is needed
3212 for linking it into DSOs.
3213 [Ralf S. Engelschall]
3215 *) Remarkably, export ciphers were totally broken and no-one had noticed!
3219 *) Cleaned up the LICENSE document: The official contact for any license
3220 questions now is the OpenSSL core team under openssl-core@openssl.org.
3221 And add a paragraph about the dual-license situation to make sure people
3222 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
3223 to the OpenSSL toolkit.
3224 [Ralf S. Engelschall]
3226 *) General source tree makefile cleanups: Made `making xxx in yyy...'
3227 display consistent in the source tree and replaced `/bin/rm' by `rm'.
3228 Additonally cleaned up the `make links' target: Remove unnecessary
3229 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
3230 to speed processing and no longer clutter the display with confusing
3231 stuff. Instead only the actually done links are displayed.
3232 [Ralf S. Engelschall]
3234 *) Permit null encryption ciphersuites, used for authentication only. It used
3235 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
3236 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
3240 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
3241 signed attributes when verifying signatures (this would break them),
3242 the detached data encoding was wrong and public keys obtained using
3243 X509_get_pubkey() weren't freed.
3246 *) Add text documentation for the BUFFER functions. Also added a work around
3247 to a Win95 console bug. This was triggered by the password read stuff: the
3248 last character typed gets carried over to the next fread(). If you were
3249 generating a new cert request using 'req' for example then the last
3250 character of the passphrase would be CR which would then enter the first
3254 *) Added the new `Includes OpenSSL Cryptography Software' button as
3255 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
3256 button and can be used by applications based on OpenSSL to show the
3257 relationship to the OpenSSL project.
3258 [Ralf S. Engelschall]
3260 *) Remove confusing variables in function signatures in files
3261 ssl/ssl_lib.c and ssl/ssl.h.
3262 [Lennart Bong <lob@kulthea.stacken.kth.se>]
3264 *) Don't install bss_file.c under PREFIX/include/
3265 [Lennart Bong <lob@kulthea.stacken.kth.se>]
3267 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
3268 functions that return function pointers and has support for NT specific
3269 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
3270 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
3271 unsigned to signed types: this was killing the Win32 compile.
3274 *) Add new certificate file to stack functions,
3275 SSL_add_dir_cert_subjects_to_stack() and
3276 SSL_add_file_cert_subjects_to_stack(). These largely supplant
3277 SSL_load_client_CA_file(), and can be used to add multiple certs easily
3278 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
3279 This means that Apache-SSL and similar packages don't have to mess around
3280 to add as many CAs as they want to the preferred list.
3283 *) Experiment with doxygen documentation. Currently only partially applied to
3285 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
3286 openssl.doxy as the configuration file.
3289 *) Get rid of remaining C++-style comments which strict C compilers hate.
3290 [Ralf S. Engelschall, pointed out by Carlos Amengual]
3292 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
3293 compiled in by default: it has problems with large keys.
3296 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
3297 DH private keys and/or callback functions which directly correspond to
3298 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
3299 is needed for applications which have to configure certificates on a
3300 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
3302 For the RSA certificate situation is makes no difference, but
3303 for the DSA certificate situation this fixes the "no shared cipher"
3304 problem where the OpenSSL cipher selection procedure failed because the
3305 temporary keys were not overtaken from the context and the API provided
3306 no way to reconfigure them.
3307 The new functions now let applications reconfigure the stuff and they
3308 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
3309 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
3310 non-public-API function ssl_cert_instantiate() is used as a helper
3311 function and also to reduce code redundancy inside ssl_rsa.c.
3312 [Ralf S. Engelschall]
3314 *) Move s_server -dcert and -dkey options out of the undocumented feature
3315 area because they are useful for the DSA situation and should be
3316 recognized by the users.
3317 [Ralf S. Engelschall]
3319 *) Fix the cipher decision scheme for export ciphers: the export bits are
3320 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
3321 SSL_EXP_MASK. So, the original variable has to be used instead of the
3322 already masked variable.
3323 [Richard Levitte <levitte@stacken.kth.se>]
3325 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
3326 [Richard Levitte <levitte@stacken.kth.se>]
3328 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
3329 from `int' to `unsigned int' because it's a length and initialized by
3330 EVP_DigestFinal() which expects an `unsigned int *'.
3331 [Richard Levitte <levitte@stacken.kth.se>]
3333 *) Don't hard-code path to Perl interpreter on shebang line of Configure
3334 script. Instead use the usual Shell->Perl transition trick.
3335 [Ralf S. Engelschall]
3337 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
3338 (in addition to RSA certificates) to match the behaviour of `openssl dsa
3339 -noout -modulus' as it's already the case for `openssl rsa -noout
3340 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
3341 currently the public key is printed (a decision which was already done by
3342 `openssl dsa -modulus' in the past) which serves a similar purpose.
3343 Additionally the NO_RSA no longer completely removes the whole -modulus
3344 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
3346 [Ralf S. Engelschall]
3348 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
3349 BIO. See the source (crypto/evp/bio_ok.c) for more info.
3350 [Arne Ansper <arne@ats.cyber.ee>]
3352 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
3353 to be added. Now both 'req' and 'ca' can use new objects defined in the
3357 *) Add cool BIO that does syslog (or event log on NT).
3358 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
3360 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
3361 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
3362 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
3363 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
3366 *) Add preliminary config info for new extension code.
3369 *) Make RSA_NO_PADDING really use no padding.
3370 [Ulf Moeller <ulf@fitug.de>]
3372 *) Generate errors when private/public key check is done.
3375 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
3376 for some CRL extensions and new objects added.
3379 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
3380 key usage extension and fuller support for authority key id.
3383 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
3384 padding method for RSA, which is recommended for new applications in PKCS
3385 #1 v2.0 (RFC 2437, October 1998).
3386 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
3387 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
3388 against Bleichbacher's attack on RSA.
3389 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
3392 *) Updates to the new SSL compression code
3393 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3395 *) Fix so that the version number in the master secret, when passed
3396 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
3397 (because the server will not accept higher), that the version number
3398 is 0x03,0x01, not 0x03,0x00
3399 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3401 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
3402 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
3403 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
3406 *) Support for RAW extensions where an arbitrary extension can be
3407 created by including its DER encoding. See apps/openssl.cnf for
3411 *) Make sure latest Perl versions don't interpret some generated C array
3412 code as Perl array code in the crypto/err/err_genc.pl script.
3413 [Lars Weber <3weber@informatik.uni-hamburg.de>]
3415 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
3416 not many people have the assembler. Various Win32 compilation fixes and
3417 update to the INSTALL.W32 file with (hopefully) more accurate Win32
3421 *) Modify configure script 'Configure' to automatically create crypto/date.h
3422 file under Win32 and also build pem.h from pem.org. New script
3423 util/mkfiles.pl to create the MINFO file on environments that can't do a
3424 'make files': perl util/mkfiles.pl >MINFO should work.
3427 *) Major rework of DES function declarations, in the pursuit of correctness
3428 and purity. As a result, many evil casts evaporated, and some weirdness,
3429 too. You may find this causes warnings in your code. Zapping your evil
3430 casts will probably fix them. Mostly.
3433 *) Fix for a typo in asn1.h. Bug fix to object creation script
3434 obj_dat.pl. It considered a zero in an object definition to mean
3435 "end of object": none of the objects in objects.h have any zeros
3436 so it wasn't spotted.
3437 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
3439 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
3440 Masking (CBCM). In the absence of test vectors, the best I have been able
3441 to do is check that the decrypt undoes the encrypt, so far. Send me test
3442 vectors if you have them.
3445 *) Correct calculation of key length for export ciphers (too much space was
3446 allocated for null ciphers). This has not been tested!
3449 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
3450 message is now correct (it understands "crypto" and "ssl" on its
3451 command line). There is also now an "update" option. This will update
3452 the util/ssleay.num and util/libeay.num files with any new functions.
3454 perl util/mkdef.pl crypto ssl update
3455 it will update them.
3458 *) Overhauled the Perl interface (perl/*):
3459 - ported BN stuff to OpenSSL's different BN library
3460 - made the perl/ source tree CVS-aware
3461 - renamed the package from SSLeay to OpenSSL (the files still contain
3462 their history because I've copied them in the repository)
3463 - removed obsolete files (the test scripts will be replaced
3464 by better Test::Harness variants in the future)
3465 [Ralf S. Engelschall]
3467 *) First cut for a very conservative source tree cleanup:
3468 1. merge various obsolete readme texts into doc/ssleay.txt
3469 where we collect the old documents and readme texts.
3470 2. remove the first part of files where I'm already sure that we no
3471 longer need them because of three reasons: either they are just temporary
3472 files which were left by Eric or they are preserved original files where
3473 I've verified that the diff is also available in the CVS via "cvs diff
3474 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
3475 the crypto/md/ stuff).
3476 [Ralf S. Engelschall]
3478 *) More extension code. Incomplete support for subject and issuer alt
3479 name, issuer and authority key id. Change the i2v function parameters
3480 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
3481 what that's for :-) Fix to ASN1 macro which messed up
3482 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
3485 *) Preliminary support for ENUMERATED type. This is largely copied from the
3489 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
3490 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3492 *) Make sure `make rehash' target really finds the `openssl' program.
3493 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
3495 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
3496 like to hear about it if this slows down other processors.
3499 *) Add CygWin32 platform information to Configure script.
3500 [Alan Batie <batie@aahz.jf.intel.com>]
3502 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
3503 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
3505 *) New program nseq to manipulate netscape certificate sequences
3508 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
3512 *) Fixes to BN code. Previously the default was to define BN_RECURSION
3513 but the BN code had some problems that would cause failures when
3514 doing certificate verification and some other functions.
3515 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3517 *) Add ASN1 and PEM code to support netscape certificate sequences.
3520 *) Add ASN1 and PEM code to support netscape certificate sequences.
3523 *) Add several PKIX and private extended key usage OIDs.
3526 *) Modify the 'ca' program to handle the new extension code. Modify
3527 openssl.cnf for new extension format, add comments.
3530 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
3531 and add a sample to openssl.cnf so req -x509 now adds appropriate
3535 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
3536 error code, add initial support to X509_print() and x509 application.
3539 *) Takes a deep breath and start addding X509 V3 extension support code. Add
3540 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
3541 stuff is currently isolated and isn't even compiled yet.
3544 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
3545 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
3546 Removed the versions check from X509 routines when loading extensions:
3547 this allows certain broken certificates that don't set the version
3548 properly to be processed.
3551 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
3552 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
3553 can still be regenerated with "make depend".
3556 *) Spelling mistake in C version of CAST-128.
3557 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
3559 *) Changes to the error generation code. The perl script err-code.pl
3560 now reads in the old error codes and retains the old numbers, only
3561 adding new ones if necessary. It also only changes the .err files if new
3562 codes are added. The makefiles have been modified to only insert errors
3563 when needed (to avoid needlessly modifying header files). This is done
3564 by only inserting errors if the .err file is newer than the auto generated
3565 C file. To rebuild all the error codes from scratch (the old behaviour)
3566 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
3567 or delete all the .err files.
3570 *) CAST-128 was incorrectly implemented for short keys. The C version has
3571 been fixed, but is untested. The assembler versions are also fixed, but
3572 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
3573 to regenerate it if needed.
3574 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
3575 Hagino <itojun@kame.net>]
3577 *) File was opened incorrectly in randfile.c.
3578 [Ulf Möller <ulf@fitug.de>]
3580 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
3581 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
3582 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
3583 al: it's just almost always a UTCTime. Note this patch adds new error
3584 codes so do a "make errors" if there are problems.
3587 *) Correct Linux 1 recognition in config.
3588 [Ulf Möller <ulf@fitug.de>]
projects / openssl.git / blob