4 Changes between 0.9.6 and 0.9.7 [xx XXX 2000]
6 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
7 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
8 files to get correct declarations of the ASN.1 item variables.
11 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
12 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
14 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
15 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
16 ASN1_ITEM and no wrapper functions.
19 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
20 replace the old function pointer based I/O routines. Change most of
21 the *_d2i_bio() and *_d2i_fp() functions to use these.
24 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
25 lines, recognice more "algorithms" that can be deselected, and make
26 it complain about algorithm deselection that isn't recognised.
29 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
30 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
31 to use new functions. Add NO_ASN1_OLD which can be set to remove
32 some old style ASN1 functions: this can be used to determine if old
33 code will still work when these eventually go away.
36 *) New extension functions for OCSP structures, these follow the
37 same conventions as certificates and CRLs.
40 *) New function X509V3_add1_i2d(). This automatically encodes and
41 adds an extension. Its behaviour can be customised with various
42 flags to append, replace or delete. Various wrappers added for
46 *) Fix to avoid calling the underlying ASN1 print routine when
47 an extension cannot be parsed. Correct a typo in the
48 OCSP_SERVICELOC extension. Tidy up print OCSP format.
51 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
52 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
53 when writing a 32767 byte record.
54 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
56 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
57 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
59 (RSA objects have a reference count access to which is protected
60 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
61 so they are meant to be shared between threads.)
62 [Bodo Moeller, Geoff Thorpe; original patch submitted by
63 "Reddie, Steven" <Steven.Reddie@ca.com>]
65 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
66 entries for variables.
69 *) Fix a deadlock in CRYPTO_mem_leaks().
72 *) Add functionality to apps/openssl.c for detecting locking
73 problems: As the program is single-threaded, all we have
74 to do is register a locking callback using an array for
75 storing which locks are currently held by the program.
78 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
79 SSL_get_ex_data_X509_STORE_idx(), which is used in
80 ssl_verify_cert_chain() and thus can be called at any time
81 during TLS/SSL handshakes so that thread-safety is essential.
82 Unfortunately, the ex_data design is not at all suited
83 for multi-threaded use, so it probably should be abolished.
86 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
87 [Broadcom, tweaked and integrated by Geoff Thorpe]
89 *) Move common extension printing code to new function
90 X509V3_print_extensions(). Reorganise OCSP print routines and
91 implement some needed OCSP ASN1 functions. Add OCSP extensions.
94 *) New function X509_signature_print() to remove duplication in some
98 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
99 set (this was treated exactly the same as SET OF previously). This
100 is used to reorder the STACK representing the structure to match the
101 encoding. This will be used to get round a problem where a PKCS7
102 structure which was signed could not be verified because the STACK
103 order did not reflect the encoded order.
106 *) Reimplement the OCSP ASN1 module using the new code.
109 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
110 for its ASN1 operations. The old style function pointers still exist
111 for now but they will eventually go away.
114 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
115 completely replaces the old ASN1 functionality with a table driven
116 encoder and decoder which interprets an ASN1_ITEM structure describing
117 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
118 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
119 has also been converted to the new form.
122 *) Change BN_mod_exp_recp so that negative moduli are tolerated
123 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
124 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
128 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
129 of not touching the result's sign bit.
132 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
136 *) Changed the LHASH code to use prototypes for callbacks, and created
137 macros to declare and implement thin (optionally static) functions
138 that provide type-safety and avoid function pointer casting for the
139 type-specific callbacks.
142 *) Use better test patterns in bntest.
145 *) Added Kerberos Cipher Suites to be used with TLS, as written in
147 [Veers Staats <staatsvr@asc.hpc.mil>,
148 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
150 *) rand_win.c fix for Borland C.
153 *) BN_rshift bugfix for n == 0.
156 *) Reformat the FAQ so the different questions and answers can be divided
157 in sections depending on the subject.
160 *) Have the zlib compression code load ZLIB.DLL dynamically under
164 *) New function BN_mod_sqrt for computing square roots modulo a prime
165 (using the probabilistic Tonelli-Shanks algorithm unless
166 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
167 be handled deterministically).
168 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
170 *) Store verify_result within SSL_SESSION also for client side to
171 avoid potential security hole. (Re-used sessions on the client side
172 always resulted in verify_result==X509_V_OK, not using the original
173 result of the server certificate verification.)
176 *) Make BN_mod_inverse faster by explicitly handling small quotients
177 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
178 512 bits], about 30% for larger ones [1024 or 2048 bits].)
181 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
182 SSL3_RT_APPLICATION_DATA, return 0.
183 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
187 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
188 releases, have been re-implemented by renaming the previous
189 implementations of ssl2_read and ssl3_read to ssl2_read_internal
190 and ssl3_read_internal, respectively, and adding 'peek' parameters
191 to them. The new ssl[23]_{read,peek} functions are calls to
192 ssl[23]_read_internal with the 'peek' flag set appropriately.
193 A 'peek' parameter has also been added to ssl3_read_bytes, which
194 does the actual work for ssl3_read_internal.
197 *) New function BN_kronecker.
200 *) Fix BN_gcd so that it works on negative inputs; the result is
201 positive unless both parameters are zero.
202 Previously something reasonably close to an infinite loop was
203 possible because numbers could be growing instead of shrinking
204 in the implementation of Euclid's algorithm.
207 *) Fix BN_is_word() and BN_is_one() macros to take into account the
208 sign of the number in question.
210 Fix BN_is_word(a,w) to work correctly for w == 0.
212 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
213 because its test if the absolute value of 'a' equals 'w'.
214 Note that BN_abs_is_word does *not* handle w == 0 reliably;
215 it exists mostly for use in the implementations of BN_is_zero(),
216 BN_is_one(), and BN_is_word().
219 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
220 the method-specific "init()" handler. Also clean up ex_data after
221 calling the method-specific "finish()" handler. Previously, this was
222 happening the other way round.
225 *) New function BN_swap.
228 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
229 the exponentiation functions are more likely to produce reasonable
230 results on negative inputs.
233 *) Change BN_mod_mul so that the result is always non-negative.
234 Previously, it could be negative if one of the factors was negative;
235 I don't think anyone really wanted that behaviour.
238 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
239 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
240 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
241 and add new functions:
254 These functions always generate non-negative results.
256 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
257 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
259 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
260 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
262 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
264 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
265 was actually never needed) and in BN_mul(). The removal in BN_mul()
266 required a small change in bn_mul_part_recursive() and the addition
267 of the functions bn_cmp_part_words(), bn_sub_part_words() and
268 bn_add_part_words(), which do the same thing as bn_cmp_words(),
269 bn_sub_words() and bn_add_words() except they take arrays with
273 *) In 'openssl passwd', verify passwords read from the terminal
274 unless the '-salt' option is used (which usually means that
275 verification would just waste user's time since the resulting
276 hash is going to be compared with some given password hash)
277 or the new '-noverify' option is used.
279 This is an incompatible change, but it does not affect
280 non-interactive use of 'openssl passwd' (passwords on the command
281 line, '-stdin' option, '-in ...' option) and thus should not
285 *) Remove all references to RSAref, since there's no more need for it.
288 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
289 The previous value, 12, was not always sufficient for BN_mod_exp().
292 *) Make DSO load along a path given through an environment variable
293 (SHLIB_PATH) with shl_load().
296 *) Constify the ENGINE code as a result of BIGNUM constification.
297 Also constify the RSA code and most things related to it. In a
298 few places, most notable in the depth of the ASN.1 code, ugly
299 casts back to non-const were required (to be solved at a later
303 *) Make it so the openssl application has all engines loaded by default.
306 *) Constify the BIGNUM routines a little more.
309 *) Make sure that shared libraries get the internal name engine with
310 the full version number and not just 0. This should mark the
311 shared libraries as not backward compatible. Of course, this should
312 be changed again when we can guarantee backward binary compatibility.
315 *) Add the following functions:
321 ENGINE_load_builtin_engines()
323 That way, an application can itself choose if external engines that
324 are built-in in OpenSSL shall ever be used or not. The benefit is
325 that applications won't have to be linked with libdl or other dso
326 libraries unless it's really needed.
328 Changed 'openssl engine' to load all engines on demand.
329 Changed the engine header files to avoid the duplication of some
330 declarations (they differed!).
333 *) 'openssl engine' can now list capabilities.
336 *) Better error reporting in 'openssl engine'.
339 *) Never call load_dh_param(NULL) in s_server.
342 *) Add engine application. It can currently list engines by name and
343 identity, and test if they are actually available.
346 *) Add support for shared libraries under Irix.
347 [Albert Chin-A-Young <china@thewrittenword.com>]
349 *) Improve RPM specification file by forcing symbolic linking and making
350 sure the installed documentation is also owned by root.root.
351 [Damien Miller <djm@mindrot.org>]
353 *) Add configuration option to build on Linux on both big-endian and
355 [Ralf Baechle <ralf@uni-koblenz.de>]
357 *) Give the OpenSSL applications more possibilities to make use of
358 keys (public as well as private) handled by engines.
361 *) Add OCSP code that comes from CertCo.
364 *) Add VMS support for the Rijndael code.
367 *) Added untested support for Nuron crypto accelerator.
370 *) Add support for external cryptographic devices. This code was
371 previously distributed separately as the "engine" branch.
372 [Geoff Thorpe, Richard Levitte]
374 *) Rework the filename-translation in the DSO code. It is now possible to
375 have far greater control over how a "name" is turned into a filename
376 depending on the operating environment and any oddities about the
377 different shared library filenames on each system.
380 *) Support threads on FreeBSD-elf in Configure.
383 *) Add the possibility to create shared libraries on HP-UX
386 *) Fix for SHA1 assembly problem with MASM: it produces
387 warnings about corrupt line number information when assembling
388 with debugging information. This is caused by the overlapping
390 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
393 NCONF_get_number() has no error checking at all. As a replacement,
394 NCONF_get_number_e() is defined (_e for "error checking") and is
395 promoted strongly. The old NCONF_get_number is kept around for
396 binary backward compatibility.
397 Make it possible for methods to load from something other than a BIO,
398 by providing a function pointer that is given a name instead of a BIO.
399 For example, this could be used to load configuration data from an
403 *) Fix typo in get_cert_by_subject() in by_dir.c
404 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
406 *) Rework the system to generate shared libraries:
408 - Make note of the expected extension for the shared libraries and
409 if there is a need for symbolic links from for example libcrypto.so.0
410 to libcrypto.so.0.9.7. There is extended info in Configure for
413 - Make as few rebuilds of the shared libraries as possible.
415 - Still avoid linking the OpenSSL programs with the shared libraries.
417 - When installing, install the shared libraries separately from the
421 *) Fix for non blocking accept BIOs. Added new I/O special reason
422 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
423 with non blocking I/O was not possible because no retry code was
424 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
428 *) Added the beginnings of Rijndael support.
431 *) Fix for bug in DirectoryString mask setting. Add support for
432 X509_NAME_print_ex() in 'req' and X509_print_ex() function
433 to allow certificate printing to more controllable, additional
434 'certopt' option to 'x509' to allow new printing options to be
438 *) Clean old EAY MD5 hack from e_os.h.
441 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
443 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
444 and not in SSL_clear because the latter is also used by the
445 accept/connect functions; previously, the settings made by
446 SSL_set_read_ahead would be lost during the handshake.
447 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
449 *) Correct util/mkdef.pl to be selective about disabled algorithms.
450 Previously, it would create entries for disableed algorithms no
454 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
456 *) In ssl23_get_client_hello, generate an error message when faced
457 with an initial SSL 3.0/TLS record that is too small to contain the
458 first two bytes of the ClientHello message, i.e. client_version.
459 (Note that this is a pathologic case that probably has never happened
460 in real life.) The previous approach was to use the version number
461 from the record header as a substitute; but our protocol choice
462 should not depend on that one because it is not authenticated
463 by the Finished messages.
466 *) More robust randomness gathering functions for Windows.
467 [Jeffrey Altman <jaltman@columbia.edu>]
469 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
470 not set then we don't setup the error code for issuer check errors
471 to avoid possibly overwriting other errors which the callback does
472 handle. If an application does set the flag then we assume it knows
473 what it is doing and can handle the new informational codes
477 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
478 a general "ANY" type, as such it should be able to decode anything
479 including tagged types. However it didn't check the class so it would
480 wrongly interpret tagged types in the same way as their universal
481 counterpart and unknown types were just rejected. Changed so that the
482 tagged and unknown types are handled in the same way as a SEQUENCE:
483 that is the encoding is stored intact. There is also a new type
484 "V_ASN1_OTHER" which is used when the class is not universal, in this
485 case we have no idea what the actual type is so we just lump them all
489 *) On VMS, stdout may very well lead to a file that is written to
490 in a record-oriented fashion. That means that every write() will
491 write a separate record, which will be read separately by the
492 programs trying to read from it. This can be very confusing.
494 The solution is to put a BIO filter in the way that will buffer
495 text until a linefeed is reached, and then write everything a
496 line at a time, so every record written will be an actual line,
497 not chunks of lines and not (usually doesn't happen, but I've
498 seen it once) several lines in one record. BIO_f_linebuffer() is
501 Currently, it's a VMS-only method, because that's where it has
502 been tested well enough.
505 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
506 it can return incorrect results.
507 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
508 but it was in 0.9.6-beta[12].)
511 *) Disable the check for content being present when verifying detached
512 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
513 include zero length content when signing messages.
516 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
517 BIO_ctrl (for BIO pairs).
520 *) Add DSO method for VMS.
523 *) Bug fix: Montgomery multiplication could produce results with the
527 *) Add RPM specification openssl.spec and modify it to build three
528 packages. The default package contains applications, application
529 documentation and run-time libraries. The devel package contains
530 include files, static libraries and function documentation. The
531 doc package contains the contents of the doc directory. The original
532 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
535 *) Add a large number of documentation files for many SSL routines.
536 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
538 *) Add a configuration entry for Sony News 4.
539 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
541 *) Don't set the two most significant bits to one when generating a
542 random number < q in the DSA library.
545 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
546 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
547 the underlying transport is blocking) if a handshake took place.
548 (The default behaviour is needed by applications such as s_client
549 and s_server that use select() to determine when to use SSL_read;
550 but for applications that know in advance when to expect data, it
551 just makes things more complicated.)
554 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
558 *) Add a few more EBCDIC conditionals that make `req' and `x509'
559 work better on such systems.
560 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
562 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
563 Update PKCS12_parse() so it copies the friendlyName and the
564 keyid to the certificates aux info.
567 *) Fix bug in PKCS7_verify() which caused an infinite loop
568 if there was more than one signature.
569 [Sven Uszpelkat <su@celocom.de>]
571 *) Major change in util/mkdef.pl to include extra information
572 about each symbol, as well as presentig variables as well
573 as functions. This change means that there's n more need
574 to rebuild the .num files when some algorithms are excluded.
577 *) Allow the verify time to be set by an application,
578 rather than always using the current time.
581 *) Phase 2 verify code reorganisation. The certificate
582 verify code now looks up an issuer certificate by a
583 number of criteria: subject name, authority key id
584 and key usage. It also verifies self signed certificates
585 by the same criteria. The main comparison function is
586 X509_check_issued() which performs these checks.
588 Lot of changes were necessary in order to support this
589 without completely rewriting the lookup code.
591 Authority and subject key identifier are now cached.
593 The LHASH 'certs' is X509_STORE has now been replaced
594 by a STACK_OF(X509_OBJECT). This is mainly because an
595 LHASH can't store or retrieve multiple objects with
598 As a result various functions (which were all internal
599 use only) have changed to handle the new X509_STORE
600 structure. This will break anything that messed round
601 with X509_STORE internally.
603 The functions X509_STORE_add_cert() now checks for an
604 exact match, rather than just subject name.
606 The X509_STORE API doesn't directly support the retrieval
607 of multiple certificates matching a given criteria, however
608 this can be worked round by performing a lookup first
609 (which will fill the cache with candidate certificates)
610 and then examining the cache for matches. This is probably
611 the best we can do without throwing out X509_LOOKUP
612 entirely (maybe later...).
614 The X509_VERIFY_CTX structure has been enhanced considerably.
616 All certificate lookup operations now go via a get_issuer()
617 callback. Although this currently uses an X509_STORE it
618 can be replaced by custom lookups. This is a simple way
619 to bypass the X509_STORE hackery necessary to make this
620 work and makes it possible to use more efficient techniques
621 in future. A very simple version which uses a simple
622 STACK for its trusted certificate store is also provided
623 using X509_STORE_CTX_trusted_stack().
625 The verify_cb() and verify() callbacks now have equivalents
626 in the X509_STORE_CTX structure.
628 X509_STORE_CTX also has a 'flags' field which can be used
629 to customise the verify behaviour.
632 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
633 excludes S/MIME capabilities.
636 *) When a certificate request is read in keep a copy of the
637 original encoding of the signed data and use it when outputing
638 again. Signatures then use the original encoding rather than
639 a decoded, encoded version which may cause problems if the
640 request is improperly encoded.
643 *) For consistency with other BIO_puts implementations, call
644 buffer_write(b, ...) directly in buffer_puts instead of calling
647 In BIO_puts, increment b->num_write as in BIO_write.
648 [Peter.Sylvester@EdelWeb.fr]
650 *) Fix BN_mul_word for the case where the word is 0. (We have to use
651 BN_zero, we may not return a BIGNUM with an array consisting of
655 *) Avoid calling abort() from within the library when problems are
656 detected, except if preprocessor symbols have been defined
657 (such as REF_CHECK, BN_DEBUG etc.).
660 *) New openssl application 'rsautl'. This utility can be
661 used for low level RSA operations. DER public key
662 BIO/fp routines also added.
665 *) New Configure entry and patches for compiling on QNX 4.
666 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
668 *) A demo state-machine implementation was sponsored by
669 Nuron (http://www.nuron.com/) and is now available in
673 *) New options added to the 'dgst' utility for signature
674 generation and verification.
677 *) Unrecognized PKCS#7 content types are now handled via a
678 catch all ASN1_TYPE structure. This allows unsupported
679 types to be stored as a "blob" and an application can
680 encode and decode it manually.
683 *) Fix various signed/unsigned issues to make a_strex.c
685 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
687 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
688 length if passed a buffer. ASN1_INTEGER_to_BN failed
689 if passed a NULL BN and its argument was negative.
690 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
692 *) Modification to PKCS#7 encoding routines to output definite
693 length encoding. Since currently the whole structures are in
694 memory there's not real point in using indefinite length
695 constructed encoding. However if OpenSSL is compiled with
696 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
699 *) Added BIO_vprintf() and BIO_vsnprintf().
702 *) Added more prefixes to parse for in the the strings written
703 through a logging bio, to cover all the levels that are available
704 through syslog. The prefixes are now:
706 PANIC, EMERG, EMR => LOG_EMERG
707 ALERT, ALR => LOG_ALERT
708 CRIT, CRI => LOG_CRIT
709 ERROR, ERR => LOG_ERR
710 WARNING, WARN, WAR => LOG_WARNING
711 NOTICE, NOTE, NOT => LOG_NOTICE
712 INFO, INF => LOG_INFO
713 DEBUG, DBG => LOG_DEBUG
715 and as before, if none of those prefixes are present at the
716 beginning of the string, LOG_ERR is chosen.
718 On Win32, the LOG_* levels are mapped according to this:
720 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
721 LOG_WARNING => EVENTLOG_WARNING_TYPE
722 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
726 *) Made it possible to reconfigure with just the configuration
727 argument "reconf" or "reconfigure". The command line arguments
728 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
729 and are retrieved from there when reconfiguring.
733 [Assar Westerlund <assar@sics.se>, Richard Levitte]
735 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
738 *) The obj_dat.pl script was messing up the sorting of object
739 names. The reason was that it compared the quoted version
740 of strings as a result "OCSP" > "OCSP Signing" because
741 " > SPACE. Changed script to store unquoted versions of
742 names and add quotes on output. It was also omitting some
743 names from the lookup table if they were given a default
744 value (that is if SN is missing it is given the same
745 value as LN and vice versa), these are now added on the
746 grounds that if an object has a name we should be able to
747 look it up. Finally added warning output when duplicate
748 short or long names are found.
751 *) Changes needed for Tandem NSK.
752 [Scott Uroff <scott@xypro.com>]
754 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
755 RSA_padding_check_SSLv23(), special padding was never detected
756 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
757 version rollback attacks was not effective.
759 In s23_clnt.c, don't use special rollback-attack detection padding
760 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
761 client; similarly, in s23_srvr.c, don't do the rollback check if
762 SSL 2.0 is the only protocol enabled in the server.
765 *) Make it possible to get hexdumps of unprintable data with 'openssl
766 asn1parse'. By implication, the functions ASN1_parse_dump() and
767 BIO_dump_indent() are added.
770 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
771 these print out strings and name structures based on various
772 flags including RFC2253 support and proper handling of
773 multibyte characters. Added options to the 'x509' utility
774 to allow the various flags to be set.
777 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
778 Also change the functions X509_cmp_current_time() and
779 X509_gmtime_adj() work with an ASN1_TIME structure,
780 this will enable certificates using GeneralizedTime in validity
784 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
785 negative public key encodings) on by default,
786 NO_NEG_PUBKEY_BUG can be set to disable it.
789 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
790 content octets. An i2c_ASN1_OBJECT is unnecessary because
791 the encoding can be trivially obtained from the structure.
794 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
795 not read locks (CRYPTO_r_[un]lock).
798 *) A first attempt at creating official support for shared
799 libraries through configuration. I've kept it so the
800 default is static libraries only, and the OpenSSL programs
801 are always statically linked for now, but there are
802 preparations for dynamic linking in place.
803 This has been tested on Linux and True64.
806 *) Randomness polling function for Win9x, as described in:
807 Peter Gutmann, Software Generation of Practically Strong
811 *) Fix so PRNG is seeded in req if using an already existing
815 *) New options to smime application. -inform and -outform
816 allow alternative formats for the S/MIME message including
817 PEM and DER. The -content option allows the content to be
818 specified separately. This should allow things like Netscape
819 form signing output easier to verify.
822 *) Fix the ASN1 encoding of tags using the 'long form'.
825 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
826 STRING types. These convert content octets to and from the
827 underlying type. The actual tag and length octets are
828 already assumed to have been read in and checked. These
829 are needed because all other string types have virtually
830 identical handling apart from the tag. By having versions
831 of the ASN1 functions that just operate on content octets
832 IMPLICIT tagging can be handled properly. It also allows
833 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
834 and ASN1_INTEGER are identical apart from the tag.
837 *) Change the handling of OID objects as follows:
839 - New object identifiers are inserted in objects.txt, following
840 the syntax given in objects.README.
841 - objects.pl is used to process obj_mac.num and create a new
843 - obj_dat.pl is used to create a new obj_dat.h, using the data in
846 This is currently kind of a hack, and the perl code in objects.pl
847 isn't very elegant, but it works as I intended. The simplest way
848 to check that it worked correctly is to look in obj_dat.h and
849 check the array nid_objs and make sure the objects haven't moved
850 around (this is important!). Additions are OK, as well as
851 consistent name changes.
854 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
857 *) Addition of the command line parameter '-rand file' to 'openssl req'.
858 The given file adds to whatever has already been seeded into the
859 random pool through the RANDFILE configuration file option or
860 environment variable, or the default random state file.
863 *) mkstack.pl now sorts each macro group into lexical order.
864 Previously the output order depended on the order the files
865 appeared in the directory, resulting in needless rewriting
869 *) Patches to make OpenSSL compile under Win32 again. Mostly
870 work arounds for the VC++ problem that it treats func() as
871 func(void). Also stripped out the parts of mkdef.pl that
872 added extra typesafe functions: these no longer exist.
875 *) Reorganisation of the stack code. The macros are now all
876 collected in safestack.h . Each macro is defined in terms of
877 a "stack macro" of the form SKM_<name>(type, a, b). The
878 DEBUG_SAFESTACK is now handled in terms of function casts,
879 this has the advantage of retaining type safety without the
880 use of additional functions. If DEBUG_SAFESTACK is not defined
881 then the non typesafe macros are used instead. Also modified the
882 mkstack.pl script to handle the new form. Needs testing to see
883 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
884 the default if no major problems. Similar behaviour for ASN1_SET_OF
888 *) When some versions of IIS use the 'NET' form of private key the
889 key derivation algorithm is different. Normally MD5(password) is
890 used as a 128 bit RC4 key. In the modified case
891 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
892 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
893 as the old Netscape_RSA functions except they have an additional
894 'sgckey' parameter which uses the modified algorithm. Also added
895 an -sgckey command line option to the rsa utility. Thanks to
896 Adrian Peck <bertie@ncipher.com> for posting details of the modified
897 algorithm to openssl-dev.
900 *) The evp_local.h macros were using 'c.##kname' which resulted in
901 invalid expansion on some systems (SCO 5.0.5 for example).
902 Corrected to 'c.kname'.
903 [Phillip Porch <root@theporch.com>]
905 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
906 a STACK of email addresses from a certificate or request, these look
907 in the subject name and the subject alternative name extensions and
908 omit any duplicate addresses.
911 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
912 This makes DSA verification about 2 % faster.
915 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
916 (meaning that now 2^5 values will be precomputed, which is only 4 KB
917 plus overhead for 1024 bit moduli).
918 This makes exponentiations about 0.5 % faster for 1024 bit
919 exponents (as measured by "openssl speed rsa2048").
922 *) Rename memory handling macros to avoid conflicts with other
924 Malloc => OPENSSL_malloc
925 Malloc_locked => OPENSSL_malloc_locked
926 Realloc => OPENSSL_realloc
930 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
931 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
935 [John Jarvie <jjarvie@newsguy.com>]
937 *) The type-safe stack code has been rejigged. It is now only compiled
938 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
939 by default all type-specific stack functions are "#define"d back to
940 standard stack functions. This results in more streamlined output
941 but retains the type-safety checking possibilities of the original
945 *) The STACK code has been cleaned up, and certain type declarations
946 that didn't make a lot of sense have been brought in line. This has
947 also involved a cleanup of sorts in safestack.h to more correctly
948 map type-safe stack functions onto their plain stack counterparts.
949 This work has also resulted in a variety of "const"ifications of
950 lots of the code, especially "_cmp" operations which should normally
951 be prototyped with "const" parameters anyway.
954 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
955 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
956 (The PRNG state consists of two parts, the large pool 'state' and 'md',
957 where all of 'md' is used each time the PRNG is used, but 'state'
958 is used only indexed by a cyclic counter. As entropy may not be
959 well distributed from the beginning, 'md' is important as a
960 chaining variable. However, the output function chains only half
961 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
962 all of 'md', and seeding with STATE_SIZE dummy bytes will result
963 in all of 'state' being rewritten, with the new values depending
964 on virtually all of 'md'. This overcomes the 80 bit limitation.)
967 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
968 the handshake is continued after ssl_verify_cert_chain();
969 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
970 can lead to 'unexplainable' connection aborts later.
971 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
973 *) Major EVP API cipher revision.
974 Add hooks for extra EVP features. This allows various cipher
975 parameters to be set in the EVP interface. Support added for variable
976 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
977 setting of RC2 and RC5 parameters.
979 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
982 Remove lots of duplicated code from the EVP library. For example *every*
983 cipher init() function handles the 'iv' in the same way according to the
984 cipher mode. They also all do nothing if the 'key' parameter is NULL and
985 for CFB and OFB modes they zero ctx->num.
987 New functionality allows removal of S/MIME code RC2 hack.
989 Most of the routines have the same form and so can be declared in terms
992 By shifting this to the top level EVP_CipherInit() it can be removed from
993 all individual ciphers. If the cipher wants to handle IVs or keys
994 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
997 Change lots of functions like EVP_EncryptUpdate() to now return a
998 value: although software versions of the algorithms cannot fail
999 any installed hardware versions can.
1002 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
1003 this option is set, tolerate broken clients that send the negotiated
1004 protocol version number instead of the requested protocol version
1008 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
1009 i.e. non-zero for export ciphersuites, zero otherwise.
1010 Previous versions had this flag inverted, inconsistent with
1011 rsa_tmp_cb (..._TMP_RSA_CB).
1012 [Bodo Moeller; problem reported by Amit Chopra]
1014 *) Add missing DSA library text string. Work around for some IIS
1015 key files with invalid SEQUENCE encoding.
1018 *) Add a document (doc/standards.txt) that list all kinds of standards
1019 and so on that are implemented in OpenSSL.
1022 *) Enhance c_rehash script. Old version would mishandle certificates
1023 with the same subject name hash and wouldn't handle CRLs at all.
1024 Added -fingerprint option to crl utility, to support new c_rehash
1028 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
1031 *) Fix for SSL server purpose checking. Server checking was
1032 rejecting certificates which had extended key usage present
1033 but no ssl client purpose.
1034 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
1036 *) Make PKCS#12 code work with no password. The PKCS#12 spec
1037 is a little unclear about how a blank password is handled.
1038 Since the password in encoded as a BMPString with terminating
1039 double NULL a zero length password would end up as just the
1040 double NULL. However no password at all is different and is
1041 handled differently in the PKCS#12 key generation code. NS
1042 treats a blank password as zero length. MSIE treats it as no
1043 password on export: but it will try both on import. We now do
1044 the same: PKCS12_parse() tries zero length and no password if
1045 the password is set to "" or NULL (NULL is now a valid password:
1046 it wasn't before) as does the pkcs12 application.
1049 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
1050 perror when PEM_read_bio_X509_REQ fails, the error message must
1051 be obtained from the error queue.
1054 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
1055 it in ERR_remove_state if appropriate, and change ERR_get_state
1056 accordingly to avoid race conditions (this is necessary because
1057 thread_hash is no longer constant once set).
1060 *) Bugfix for linux-elf makefile.one.
1063 *) RSA_get_default_method() will now cause a default
1064 RSA_METHOD to be chosen if one doesn't exist already.
1065 Previously this was only set during a call to RSA_new()
1066 or RSA_new_method(NULL) meaning it was possible for
1067 RSA_get_default_method() to return NULL.
1070 *) Added native name translation to the existing DSO code
1071 that will convert (if the flag to do so is set) filenames
1072 that are sufficiently small and have no path information
1073 into a canonical native form. Eg. "blah" converted to
1074 "libblah.so" or "blah.dll" etc.
1077 *) New function ERR_error_string_n(e, buf, len) which is like
1078 ERR_error_string(e, buf), but writes at most 'len' bytes
1079 including the 0 terminator. For ERR_error_string_n, 'buf'
1081 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
1083 *) CONF library reworked to become more general. A new CONF
1084 configuration file reader "class" is implemented as well as a
1085 new functions (NCONF_*, for "New CONF") to handle it. The now
1086 old CONF_* functions are still there, but are reimplemented to
1087 work in terms of the new functions. Also, a set of functions
1088 to handle the internal storage of the configuration data is
1089 provided to make it easier to write new configuration file
1090 reader "classes" (I can definitely see something reading a
1091 configuration file in XML format, for example), called _CONF_*,
1092 or "the configuration storage API"...
1094 The new configuration file reading functions are:
1096 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
1097 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
1099 NCONF_default, NCONF_WIN32
1101 NCONF_dump_fp, NCONF_dump_bio
1103 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
1104 NCONF_new creates a new CONF object. This works in the same way
1105 as other interfaces in OpenSSL, like the BIO interface.
1106 NCONF_dump_* dump the internal storage of the configuration file,
1107 which is useful for debugging. All other functions take the same
1108 arguments as the old CONF_* functions wth the exception of the
1109 first that must be a `CONF *' instead of a `LHASH *'.
1111 To make it easer to use the new classes with the old CONF_* functions,
1112 the function CONF_set_default_method is provided.
1115 *) Add '-tls1' option to 'openssl ciphers', which was already
1116 mentioned in the documentation but had not been implemented.
1117 (This option is not yet really useful because even the additional
1118 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
1121 *) Initial DSO code added into libcrypto for letting OpenSSL (and
1122 OpenSSL-based applications) load shared libraries and bind to
1123 them in a portable way.
1124 [Geoff Thorpe, with contributions from Richard Levitte]
1126 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
1128 *) Make sure _lrotl and _lrotr are only used with MSVC.
1130 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
1131 (the default implementation of RAND_status).
1133 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
1134 to '-clrext' (= clear extensions), as intended and documented.
1135 [Bodo Moeller; inconsistency pointed out by Michael Attili
1138 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
1139 was larger than the MD block size.
1140 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
1142 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
1143 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
1144 using the passed key: if the passed key was a private key the result
1145 of X509_print(), for example, would be to print out all the private key
1149 *) des_quad_cksum() byte order bug fix.
1150 [Ulf Möller, using the problem description in krb4-0.9.7, where
1151 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
1153 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
1155 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
1157 *) For easily testing in shell scripts whether some command
1158 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
1159 returns with exit code 0 iff no command of the given name is available.
1160 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
1161 the output goes to stdout and nothing is printed to stderr.
1162 Additional arguments are always ignored.
1164 Since for each cipher there is a command of the same name,
1165 the 'no-cipher' compilation switches can be tested this way.
1167 ('openssl no-XXX' is not able to detect pseudo-commands such
1168 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
1171 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
1174 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
1175 is set; it will be thrown away anyway because each handshake creates
1177 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
1178 to parameters -- in previous versions (since OpenSSL 0.9.3) the
1179 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
1180 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
1183 *) New s_client option -ign_eof: EOF at stdin is ignored, and
1184 'Q' and 'R' lose their special meanings (quit/renegotiate).
1185 This is part of what -quiet does; unlike -quiet, -ign_eof
1186 does not suppress any output.
1189 *) Add compatibility options to the purpose and trust code. The
1190 purpose X509_PURPOSE_ANY is "any purpose" which automatically
1191 accepts a certificate or CA, this was the previous behaviour,
1192 with all the associated security issues.
1194 X509_TRUST_COMPAT is the old trust behaviour: only and
1195 automatically trust self signed roots in certificate store. A
1196 new trust setting X509_TRUST_DEFAULT is used to specify that
1197 a purpose has no associated trust setting and it should instead
1198 use the value in the default purpose.
1201 *) Fix the PKCS#8 DSA private key code so it decodes keys again
1202 and fix a memory leak.
1205 *) In util/mkerr.pl (which implements 'make errors'), preserve
1206 reason strings from the previous version of the .c file, as
1207 the default to have only downcase letters (and digits) in
1208 automatically generated reasons codes is not always appropriate.
1211 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
1212 using strerror. Previously, ERR_reason_error_string() returned
1213 library names as reason strings for SYSerr; but SYSerr is a special
1214 case where small numbers are errno values, not library numbers.
1217 *) Add '-dsaparam' option to 'openssl dhparam' application. This
1218 converts DSA parameters into DH parameters. (When creating parameters,
1219 DSA_generate_parameters is used.)
1222 *) Include 'length' (recommended exponent length) in C code generated
1223 by 'openssl dhparam -C'.
1226 *) The second argument to set_label in perlasm was already being used
1227 so couldn't be used as a "file scope" flag. Moved to third argument
1231 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
1232 instead of RAND_bytes for encryption IVs and salts.
1235 *) Include RAND_status() into RAND_METHOD instead of implementing
1236 it only for md_rand.c Otherwise replacing the PRNG by calling
1237 RAND_set_rand_method would be impossible.
1240 *) Don't let DSA_generate_key() enter an infinite loop if the random
1241 number generation fails.
1244 *) New 'rand' application for creating pseudo-random output.
1247 *) Added configuration support for Linux/IA64
1248 [Rolf Haberrecker <rolf@suse.de>]
1250 *) Assembler module support for Mingw32.
1253 *) Shared library support for HPUX (in shlib/).
1254 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
1256 *) Shared library support for Solaris gcc.
1257 [Lutz Behnke <behnke@trustcenter.de>]
1259 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
1261 *) PKCS7_encrypt() was adding text MIME headers twice because they
1262 were added manually and by SMIME_crlf_copy().
1265 *) In bntest.c don't call BN_rand with zero bits argument.
1266 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
1268 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
1269 case was implemented. This caused BN_div_recp() to fail occasionally.
1272 *) Add an optional second argument to the set_label() in the perl
1273 assembly language builder. If this argument exists and is set
1274 to 1 it signals that the assembler should use a symbol whose
1275 scope is the entire file, not just the current function. This
1276 is needed with MASM which uses the format label:: for this scope.
1277 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
1279 *) Change the ASN1 types so they are typedefs by default. Before
1280 almost all types were #define'd to ASN1_STRING which was causing
1281 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
1285 *) Change names of new functions to the new get1/get0 naming
1286 convention: After 'get1', the caller owns a reference count
1287 and has to call ..._free; 'get0' returns a pointer to some
1288 data structure without incrementing reference counters.
1289 (Some of the existing 'get' functions increment a reference
1290 counter, some don't.)
1291 Similarly, 'set1' and 'add1' functions increase reference
1292 counters or duplicate objects.
1295 *) Allow for the possibility of temp RSA key generation failure:
1296 the code used to assume it always worked and crashed on failure.
1299 *) Fix potential buffer overrun problem in BIO_printf().
1300 [Ulf Möller, using public domain code by Patrick Powell; problem
1301 pointed out by David Sacerdote <das33@cornell.edu>]
1303 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
1304 RAND_egd() and RAND_status(). In the command line application,
1305 the EGD socket can be specified like a seed file using RANDFILE
1309 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
1310 Some CAs (e.g. Verisign) distribute certificates in this form.
1313 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
1314 list to exclude them. This means that no special compilation option
1315 is needed to use anonymous DH: it just needs to be included in the
1319 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
1320 EVP_MD_type. The old functionality is available in a new macro called
1321 EVP_MD_md(). Change code that uses it and update docs.
1324 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
1325 where the 'void *' argument is replaced by a function pointer argument.
1326 Previously 'void *' was abused to point to functions, which works on
1327 many platforms, but is not correct. As these functions are usually
1328 called by macros defined in OpenSSL header files, most source code
1329 should work without changes.
1332 *) <openssl/opensslconf.h> (which is created by Configure) now contains
1333 sections with information on -D... compiler switches used for
1334 compiling the library so that applications can see them. To enable
1335 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
1336 must be defined. E.g.,
1337 #define OPENSSL_ALGORITHM_DEFINES
1338 #include <openssl/opensslconf.h>
1339 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
1340 [Richard Levitte, Ulf and Bodo Möller]
1342 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
1346 *) Change the 'other' type in certificate aux info to a STACK_OF
1347 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
1348 the required ASN1 format: arbitrary types determined by an OID.
1351 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
1352 argument to 'req'. This is not because the function is newer or
1353 better than others it just uses the work 'NEW' in the certificate
1354 request header lines. Some software needs this.
1357 *) Reorganise password command line arguments: now passwords can be
1358 obtained from various sources. Delete the PEM_cb function and make
1359 it the default behaviour: i.e. if the callback is NULL and the
1360 usrdata argument is not NULL interpret it as a null terminated pass
1361 phrase. If usrdata and the callback are NULL then the pass phrase
1362 is prompted for as usual.
1365 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
1366 the support is automatically enabled. The resulting binaries will
1367 autodetect the card and use it if present.
1368 [Ben Laurie and Compaq Inc.]
1370 *) Work around for Netscape hang bug. This sends certificate request
1371 and server done in one record. Since this is perfectly legal in the
1372 SSL/TLS protocol it isn't a "bug" option and is on by default. See
1373 the bugs/SSLv3 entry for more info.
1376 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
1379 *) Add -rand argument to smime and pkcs12 applications and read/write
1383 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
1386 *) Add command line password options to the remaining applications.
1389 *) Bug fix for BN_div_recp() for numerators with an even number of
1393 *) More tests in bntest.c, and changed test_bn output.
1396 *) ./config recognizes MacOS X now.
1399 *) Bug fix for BN_div() when the first words of num and divsor are
1400 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
1403 *) Add support for various broken PKCS#8 formats, and command line
1404 options to produce them.
1407 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
1408 get temporary BIGNUMs from a BN_CTX.
1411 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
1415 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
1416 include a #define from the old name to the new. The original intent
1417 was that statically linked binaries could for example just call
1418 SSLeay_add_all_ciphers() to just add ciphers to the table and not
1419 link with digests. This never worked becayse SSLeay_add_all_digests()
1420 and SSLeay_add_all_ciphers() were in the same source file so calling
1421 one would link with the other. They are now in separate source files.
1424 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
1427 *) Use a less unusual form of the Miller-Rabin primality test (it used
1428 a binary algorithm for exponentiation integrated into the Miller-Rabin
1429 loop, our standard modexp algorithms are faster).
1432 *) Support for the EBCDIC character set completed.
1433 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
1435 *) Source code cleanups: use const where appropriate, eliminate casts,
1436 use void * instead of char * in lhash.
1439 *) Bugfix: ssl3_send_server_key_exchange was not restartable
1440 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
1441 this the server could overwrite ephemeral keys that the client
1445 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
1446 using 50 iterations of the Rabin-Miller test.
1448 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
1449 iterations of the Rabin-Miller test as required by the appendix
1450 to FIPS PUB 186[-1]) instead of DSA_is_prime.
1451 As BN_is_prime_fasttest includes trial division, DSA parameter
1452 generation becomes much faster.
1454 This implies a change for the callback functions in DSA_is_prime
1455 and DSA_generate_parameters: The callback function is called once
1456 for each positive witness in the Rabin-Miller test, not just
1457 occasionally in the inner loop; and the parameters to the
1458 callback function now provide an iteration count for the outer
1459 loop rather than for the current invocation of the inner loop.
1460 DSA_generate_parameters additionally can call the callback
1461 function with an 'iteration count' of -1, meaning that a
1462 candidate has passed the trial division test (when q is generated
1463 from an application-provided seed, trial division is skipped).
1466 *) New function BN_is_prime_fasttest that optionally does trial
1467 division before starting the Rabin-Miller test and has
1468 an additional BN_CTX * argument (whereas BN_is_prime always
1469 has to allocate at least one BN_CTX).
1470 'callback(1, -1, cb_arg)' is called when a number has passed the
1471 trial division stage.
1474 *) Fix for bug in CRL encoding. The validity dates weren't being handled
1478 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
1481 *) New function BN_pseudo_rand().
1484 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
1485 bignum version of BN_from_montgomery() with the working code from
1486 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
1490 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
1491 made it impossible to use the same SSL_SESSION data structure in
1492 SSL2 clients in multiple threads.
1495 *) The return value of RAND_load_file() no longer counts bytes obtained
1496 by stat(). RAND_load_file(..., -1) is new and uses the complete file
1497 to seed the PRNG (previously an explicit byte count was required).
1498 [Ulf Möller, Bodo Möller]
1500 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
1501 used (char *) instead of (void *) and had casts all over the place.
1504 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
1507 *) Retain source code compatibility for BN_prime_checks macro:
1508 BN_is_prime(..., BN_prime_checks, ...) now uses
1509 BN_prime_checks_for_size to determine the appropriate number of
1510 Rabin-Miller iterations.
1513 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
1514 DH_CHECK_P_NOT_SAFE_PRIME.
1515 (Check if this is true? OpenPGP calls them "strong".)
1518 *) Merge the functionality of "dh" and "gendh" programs into a new program
1519 "dhparam". The old programs are retained for now but will handle DH keys
1520 (instead of parameters) in future.
1523 *) Make the ciphers, s_server and s_client programs check the return values
1524 when a new cipher list is set.
1527 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
1528 ciphers. Before when the 56bit ciphers were enabled the sorting was
1531 The syntax for the cipher sorting has been extended to support sorting by
1532 cipher-strength (using the strength_bits hard coded in the tables).
1533 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
1535 Fix a bug in the cipher-command parser: when supplying a cipher command
1536 string with an "undefined" symbol (neither command nor alphanumeric
1537 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
1538 an error is flagged.
1540 Due to the strength-sorting extension, the code of the
1541 ssl_create_cipher_list() function was completely rearranged. I hope that
1542 the readability was also increased :-)
1543 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
1545 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
1546 for the first serial number and places 2 in the serial number file. This
1547 avoids problems when the root CA is created with serial number zero and
1548 the first user certificate has the same issuer name and serial number
1552 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
1553 the new code. Add documentation for this stuff.
1556 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
1557 X509_*() to X509at_*() on the grounds that they don't handle X509
1558 structures and behave in an analagous way to the X509v3 functions:
1559 they shouldn't be called directly but wrapper functions should be used
1562 So we also now have some wrapper functions that call the X509at functions
1563 when passed certificate requests. (TO DO: similar things can be done with
1564 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
1565 things. Some of these need some d2i or i2d and print functionality
1566 because they handle more complex structures.)
1569 *) Add missing #ifndefs that caused missing symbols when building libssl
1570 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
1571 NO_RSA in ssl/s2*.c.
1572 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
1574 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
1575 has a return value which indicates the quality of the random data
1576 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
1577 error queue. New function RAND_pseudo_bytes() generates output that is
1578 guaranteed to be unique but not unpredictable. RAND_add is like
1579 RAND_seed, but takes an extra argument for an entropy estimate
1580 (RAND_seed always assumes full entropy).
1583 *) Do more iterations of Rabin-Miller probable prime test (specifically,
1584 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
1585 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
1586 in crypto/bn/bn_prime.c for the complete table). This guarantees a
1587 false-positive rate of at most 2^-80 for random input.
1590 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
1593 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
1594 in the 0.9.5 release), this returns the chain
1595 from an X509_CTX structure with a dup of the stack and all
1596 the X509 reference counts upped: so the stack will exist
1597 after X509_CTX_cleanup() has been called. Modify pkcs12.c
1600 Also make SSL_SESSION_print() print out the verify return
1604 *) Add manpage for the pkcs12 command. Also change the default
1605 behaviour so MAC iteration counts are used unless the new
1606 -nomaciter option is used. This improves file security and
1607 only older versions of MSIE (4.0 for example) need it.
1610 *) Honor the no-xxx Configure options when creating .DEF files.
1613 *) Add PKCS#10 attributes to field table: challengePassword,
1614 unstructuredName and unstructuredAddress. These are taken from
1615 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
1616 international characters are used.
1618 More changes to X509_ATTRIBUTE code: allow the setting of types
1619 based on strings. Remove the 'loc' parameter when adding
1620 attributes because these will be a SET OF encoding which is sorted
1624 *) Initial changes to the 'req' utility to allow request generation
1625 automation. This will allow an application to just generate a template
1626 file containing all the field values and have req construct the
1629 Initial support for X509_ATTRIBUTE handling. Stacks of these are
1630 used all over the place including certificate requests and PKCS#7
1631 structures. They are currently handled manually where necessary with
1632 some primitive wrappers for PKCS#7. The new functions behave in a
1633 manner analogous to the X509 extension functions: they allow
1634 attributes to be looked up by NID and added.
1636 Later something similar to the X509V3 code would be desirable to
1637 automatically handle the encoding, decoding and printing of the
1638 more complex types. The string types like challengePassword can
1639 be handled by the string table functions.
1641 Also modified the multi byte string table handling. Now there is
1642 a 'global mask' which masks out certain types. The table itself
1643 can use the flag STABLE_NO_MASK to ignore the mask setting: this
1644 is useful when for example there is only one permissible type
1645 (as in countryName) and using the mask might result in no valid
1649 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
1650 SSL_get_peer_finished to allow applications to obtain the latest
1651 Finished messages sent to the peer or expected from the peer,
1652 respectively. (SSL_get_peer_finished is usually the Finished message
1653 actually received from the peer, otherwise the protocol will be aborted.)
1655 As the Finished message are message digests of the complete handshake
1656 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
1657 be used for external authentication procedures when the authentication
1658 provided by SSL/TLS is not desired or is not enough.
1661 *) Enhanced support for Alpha Linux is added. Now ./config checks if
1662 the host supports BWX extension and if Compaq C is present on the
1663 $PATH. Just exploiting of the BWX extension results in 20-30%
1664 performance kick for some algorithms, e.g. DES and RC4 to mention
1665 a couple. Compaq C in turn generates ~20% faster code for MD5 and
1669 *) Add support for MS "fast SGC". This is arguably a violation of the
1670 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
1671 weak crypto and after checking the certificate is SGC a second one
1672 with strong crypto. MS SGC stops the first handshake after receiving
1673 the server certificate message and sends a second client hello. Since
1674 a server will typically do all the time consuming operations before
1675 expecting any further messages from the client (server key exchange
1676 is the most expensive) there is little difference between the two.
1678 To get OpenSSL to support MS SGC we have to permit a second client
1679 hello message after we have sent server done. In addition we have to
1680 reset the MAC if we do get this second client hello.
1683 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
1684 if a DER encoded private key is RSA or DSA traditional format. Changed
1685 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
1686 format DER encoded private key. Newer code should use PKCS#8 format which
1687 has the key type encoded in the ASN1 structure. Added DER private key
1688 support to pkcs8 application.
1691 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
1692 ciphersuites has been selected (as required by the SSL 3/TLS 1
1693 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
1694 is set, we interpret this as a request to violate the specification
1695 (the worst that can happen is a handshake failure, and 'correct'
1696 behaviour would result in a handshake failure anyway).
1699 *) In SSL_CTX_add_session, take into account that there might be multiple
1700 SSL_SESSION structures with the same session ID (e.g. when two threads
1701 concurrently obtain them from an external cache).
1702 The internal cache can handle only one SSL_SESSION with a given ID,
1703 so if there's a conflict, we now throw out the old one to achieve
1707 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
1708 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
1709 some routines that use cipher OIDs: some ciphers do not have OIDs
1710 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
1714 *) Simplify the trust setting structure and code. Now we just have
1715 two sequences of OIDs for trusted and rejected settings. These will
1716 typically have values the same as the extended key usage extension
1717 and any application specific purposes.
1719 The trust checking code now has a default behaviour: it will just
1720 check for an object with the same NID as the passed id. Functions can
1721 be provided to override either the default behaviour or the behaviour
1722 for a given id. SSL client, server and email already have functions
1723 in place for compatibility: they check the NID and also return "trusted"
1724 if the certificate is self signed.
1727 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
1728 traditional format into an EVP_PKEY structure.
1731 *) Add a password callback function PEM_cb() which either prompts for
1732 a password if usr_data is NULL or otherwise assumes it is a null
1733 terminated password. Allow passwords to be passed on command line
1734 environment or config files in a few more utilities.
1737 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
1738 keys. Add some short names for PKCS#8 PBE algorithms and allow them
1739 to be specified on the command line for the pkcs8 and pkcs12 utilities.
1740 Update documentation.
1743 *) Support for ASN1 "NULL" type. This could be handled before by using
1744 ASN1_TYPE but there wasn't any function that would try to read a NULL
1745 and produce an error if it couldn't. For compatibility we also have
1746 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
1747 don't allocate anything because they don't need to.
1750 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
1752 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
1754 *) Rebuild of the memory allocation routines used by OpenSSL code and
1755 possibly others as well. The purpose is to make an interface that
1756 provide hooks so anyone can build a separate set of allocation and
1757 deallocation routines to be used by OpenSSL, for example memory
1758 pool implementations, or something else, which was previously hard
1759 since Malloc(), Realloc() and Free() were defined as macros having
1760 the values malloc, realloc and free, respectively (except for Win32
1761 compilations). The same is provided for memory debugging code.
1762 OpenSSL already comes with functionality to find memory leaks, but
1763 this gives people a chance to debug other memory problems.
1765 With these changes, a new set of functions and macros have appeared:
1767 CRYPTO_set_mem_debug_functions() [F]
1768 CRYPTO_get_mem_debug_functions() [F]
1769 CRYPTO_dbg_set_options() [F]
1770 CRYPTO_dbg_get_options() [F]
1771 CRYPTO_malloc_debug_init() [M]
1773 The memory debug functions are NULL by default, unless the library
1774 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
1775 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
1776 gives the standard debugging functions that come with OpenSSL) or
1777 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
1778 provided by the library user) must be used. When the standard
1779 debugging functions are used, CRYPTO_dbg_set_options can be used to
1780 request additional information:
1781 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
1782 the CRYPTO_MDEBUG_xxx macro when compiling the library.
1784 Also, things like CRYPTO_set_mem_functions will always give the
1785 expected result (the new set of functions is used for allocation
1786 and deallocation) at all times, regardless of platform and compiler
1789 To finish it up, some functions that were never use in any other
1790 way than through macros have a new API and new semantic:
1793 CRYPTO_dbg_realloc()
1796 All macros of value have retained their old syntax.
1797 [Richard Levitte and Bodo Moeller]
1799 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
1800 ordering of SMIMECapabilities wasn't in "strength order" and there
1801 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
1805 *) Some ASN1 types with illegal zero length encoding (INTEGER,
1806 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
1807 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
1809 *) Merge in my S/MIME library for OpenSSL. This provides a simple
1810 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
1811 functionality to handle multipart/signed properly) and a utility
1812 called 'smime' to call all this stuff. This is based on code I
1813 originally wrote for Celo who have kindly allowed it to be
1814 included in OpenSSL.
1817 *) Add variants des_set_key_checked and des_set_key_unchecked of
1818 des_set_key (aka des_key_sched). Global variable des_check_key
1819 decides which of these is called by des_set_key; this way
1820 des_check_key behaves as it always did, but applications and
1821 the library itself, which was buggy for des_check_key == 1,
1822 have a cleaner way to pick the version they need.
1825 *) New function PKCS12_newpass() which changes the password of a
1829 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
1830 dynamic mix. In both cases the ids can be used as an index into the
1831 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
1832 functions so they accept a list of the field values and the
1833 application doesn't need to directly manipulate the X509_TRUST
1837 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
1841 *) Modify the way the V3 extension code looks up extensions. This now
1842 works in a similar way to the object code: we have some "standard"
1843 extensions in a static table which is searched with OBJ_bsearch()
1844 and the application can add dynamic ones if needed. The file
1845 crypto/x509v3/ext_dat.h now has the info: this file needs to be
1846 updated whenever a new extension is added to the core code and kept
1847 in ext_nid order. There is a simple program 'tabtest.c' which checks
1848 this. New extensions are not added too often so this file can readily
1849 be maintained manually.
1851 There are two big advantages in doing things this way. The extensions
1852 can be looked up immediately and no longer need to be "added" using
1853 X509V3_add_standard_extensions(): this function now does nothing.
1854 [Side note: I get *lots* of email saying the extension code doesn't
1855 work because people forget to call this function]
1856 Also no dynamic allocation is done unless new extensions are added:
1857 so if we don't add custom extensions there is no need to call
1858 X509V3_EXT_cleanup().
1861 *) Modify enc utility's salting as follows: make salting the default. Add a
1862 magic header, so unsalted files fail gracefully instead of just decrypting
1863 to garbage. This is because not salting is a big security hole, so people
1864 should be discouraged from doing it.
1867 *) Fixes and enhancements to the 'x509' utility. It allowed a message
1868 digest to be passed on the command line but it only used this
1869 parameter when signing a certificate. Modified so all relevant
1870 operations are affected by the digest parameter including the
1871 -fingerprint and -x509toreq options. Also -x509toreq choked if a
1872 DSA key was used because it didn't fix the digest.
1875 *) Initial certificate chain verify code. Currently tests the untrusted
1876 certificates for consistency with the verify purpose (which is set
1877 when the X509_STORE_CTX structure is set up) and checks the pathlength.
1879 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
1880 this is because it will reject chains with invalid extensions whereas
1881 every previous version of OpenSSL and SSLeay made no checks at all.
1883 Trust code: checks the root CA for the relevant trust settings. Trust
1884 settings have an initial value consistent with the verify purpose: e.g.
1885 if the verify purpose is for SSL client use it expects the CA to be
1886 trusted for SSL client use. However the default value can be changed to
1887 permit custom trust settings: one example of this would be to only trust
1888 certificates from a specific "secure" set of CAs.
1890 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
1891 which should be used for version portability: especially since the
1892 verify structure is likely to change more often now.
1894 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
1895 to set them. If not set then assume SSL clients will verify SSL servers
1898 Two new options to the verify program: -untrusted allows a set of
1899 untrusted certificates to be passed in and -purpose which sets the
1900 intended purpose of the certificate. If a purpose is set then the
1901 new chain verify code is used to check extension consistency.
1904 *) Support for the authority information access extension.
1907 *) Modify RSA and DSA PEM read routines to transparently handle
1908 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
1909 public keys in a format compatible with certificate
1910 SubjectPublicKeyInfo structures. Unfortunately there were already
1911 functions called *_PublicKey_* which used various odd formats so
1912 these are retained for compatibility: however the DSA variants were
1913 never in a public release so they have been deleted. Changed dsa/rsa
1914 utilities to handle the new format: note no releases ever handled public
1915 keys so we should be OK.
1917 The primary motivation for this change is to avoid the same fiasco
1918 that dogs private keys: there are several incompatible private key
1919 formats some of which are standard and some OpenSSL specific and
1920 require various evil hacks to allow partial transparent handling and
1921 even then it doesn't work with DER formats. Given the option anything
1922 other than PKCS#8 should be dumped: but the other formats have to
1923 stay in the name of compatibility.
1925 With public keys and the benefit of hindsight one standard format
1926 is used which works with EVP_PKEY, RSA or DSA structures: though
1927 it clearly returns an error if you try to read the wrong kind of key.
1929 Added a -pubkey option to the 'x509' utility to output the public key.
1930 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
1931 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
1932 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
1933 that do the same as the EVP_PKEY_assign_*() except they up the
1934 reference count of the added key (they don't "swallow" the
1938 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
1939 CRLs would fail if the file contained no certificates or no CRLs:
1940 added a new function to read in both types and return the number
1941 read: this means that if none are read it will be an error. The
1942 DER versions of the certificate and CRL reader would always fail
1943 because it isn't possible to mix certificates and CRLs in DER format
1944 without choking one or the other routine. Changed this to just read
1945 a certificate: this is the best we can do. Also modified the code
1946 in apps/verify.c to take notice of return codes: it was previously
1947 attempting to read in certificates from NULL pointers and ignoring
1948 any errors: this is one reason why the cert and CRL reader seemed
1949 to work. It doesn't check return codes from the default certificate
1950 routines: these may well fail if the certificates aren't installed.
1953 *) Code to support otherName option in GeneralName.
1956 *) First update to verify code. Change the verify utility
1957 so it warns if it is passed a self signed certificate:
1958 for consistency with the normal behaviour. X509_verify
1959 has been modified to it will now verify a self signed
1960 certificate if *exactly* the same certificate appears
1961 in the store: it was previously impossible to trust a
1962 single self signed certificate. This means that:
1963 openssl verify ss.pem
1964 now gives a warning about a self signed certificate but
1965 openssl verify -CAfile ss.pem ss.pem
1969 *) For servers, store verify_result in SSL_SESSION data structure
1970 (and add it to external session representation).
1971 This is needed when client certificate verifications fails,
1972 but an application-provided verification callback (set by
1973 SSL_CTX_set_cert_verify_callback) allows accepting the session
1974 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
1975 but returns 1): When the session is reused, we have to set
1976 ssl->verify_result to the appropriate error code to avoid
1978 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
1980 *) Fix a bug in the new PKCS#7 code: it didn't consider the
1981 case in PKCS7_dataInit() where the signed PKCS7 structure
1982 didn't contain any existing data because it was being created.
1983 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
1985 *) Add a salt to the key derivation routines in enc.c. This
1986 forms the first 8 bytes of the encrypted file. Also add a
1987 -S option to allow a salt to be input on the command line.
1990 *) New function X509_cmp(). Oddly enough there wasn't a function
1991 to compare two certificates. We do this by working out the SHA1
1992 hash and comparing that. X509_cmp() will be needed by the trust
1996 *) SSL_get1_session() is like SSL_get_session(), but increments
1997 the reference count in the SSL_SESSION returned.
1998 [Geoff Thorpe <geoff@eu.c2.net>]
2000 *) Fix for 'req': it was adding a null to request attributes.
2001 Also change the X509_LOOKUP and X509_INFO code to handle
2002 certificate auxiliary information.
2005 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
2009 *) Add the possibility to add extra information to the memory leak
2010 detecting output, to form tracebacks, showing from where each
2011 allocation was originated: CRYPTO_push_info("constant string") adds
2012 the string plus current file name and line number to a per-thread
2013 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
2014 is like calling CYRPTO_pop_info() until the stack is empty.
2015 Also updated memory leak detection code to be multi-thread-safe.
2018 *) Add options -text and -noout to pkcs7 utility and delete the
2019 encryption options which never did anything. Update docs.
2022 *) Add options to some of the utilities to allow the pass phrase
2023 to be included on either the command line (not recommended on
2024 OSes like Unix) or read from the environment. Update the
2025 manpages and fix a few bugs.
2028 *) Add a few manpages for some of the openssl commands.
2031 *) Fix the -revoke option in ca. It was freeing up memory twice,
2032 leaking and not finding already revoked certificates.
2035 *) Extensive changes to support certificate auxiliary information.
2036 This involves the use of X509_CERT_AUX structure and X509_AUX
2037 functions. An X509_AUX function such as PEM_read_X509_AUX()
2038 can still read in a certificate file in the usual way but it
2039 will also read in any additional "auxiliary information". By
2040 doing things this way a fair degree of compatibility can be
2041 retained: existing certificates can have this information added
2042 using the new 'x509' options.
2044 Current auxiliary information includes an "alias" and some trust
2045 settings. The trust settings will ultimately be used in enhanced
2046 certificate chain verification routines: currently a certificate
2047 can only be trusted if it is self signed and then it is trusted
2051 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
2052 The problem was that one of the replacement routines had not been working
2053 since SSLeay releases. For now the offending routine has been replaced
2054 with non-optimised assembler. Even so, this now gives around 95%
2055 performance improvement for 1024 bit RSA signs.
2058 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
2059 handling. Most clients have the effective key size in bits equal to
2060 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
2061 A few however don't do this and instead use the size of the decrypted key
2062 to determine the RC2 key length and the AlgorithmIdentifier to determine
2063 the effective key length. In this case the effective key length can still
2064 be 40 bits but the key length can be 168 bits for example. This is fixed
2065 by manually forcing an RC2 key into the EVP_PKEY structure because the
2066 EVP code can't currently handle unusual RC2 key sizes: it always assumes
2067 the key length and effective key length are equal.
2070 *) Add a bunch of functions that should simplify the creation of
2071 X509_NAME structures. Now you should be able to do:
2072 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
2073 and have it automatically work out the correct field type and fill in
2074 the structures. The more adventurous can try:
2075 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
2076 and it will (hopefully) work out the correct multibyte encoding.
2079 *) Change the 'req' utility to use the new field handling and multibyte
2080 copy routines. Before the DN field creation was handled in an ad hoc
2081 way in req, ca, and x509 which was rather broken and didn't support
2082 BMPStrings or UTF8Strings. Since some software doesn't implement
2083 BMPStrings or UTF8Strings yet, they can be enabled using the config file
2084 using the dirstring_type option. See the new comment in the default
2085 openssl.cnf for more info.
2088 *) Make crypto/rand/md_rand.c more robust:
2089 - Assure unique random numbers after fork().
2090 - Make sure that concurrent threads access the global counter and
2091 md serializably so that we never lose entropy in them
2092 or use exactly the same state in multiple threads.
2093 Access to the large state is not always serializable because
2094 the additional locking could be a performance killer, and
2095 md should be large enough anyway.
2098 *) New file apps/app_rand.c with commonly needed functionality
2099 for handling the random seed file.
2101 Use the random seed file in some applications that previously did not:
2103 dsaparam -genkey (which also ignored its '-rand' option),
2106 x509 (when signing).
2107 Except on systems with /dev/urandom, it is crucial to have a random
2108 seed file at least for key creation, DSA signing, and for DH exchanges;
2109 for RSA signatures we could do without one.
2111 gendh and gendsa (unlike genrsa) used to read only the first byte
2112 of each file listed in the '-rand' option. The function as previously
2113 found in genrsa is now in app_rand.c and is used by all programs
2114 that support '-rand'.
2117 *) In RAND_write_file, use mode 0600 for creating files;
2118 don't just chmod when it may be too late.
2121 *) Report an error from X509_STORE_load_locations
2122 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
2125 *) New function ASN1_mbstring_copy() this copies a string in either
2126 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
2127 into an ASN1_STRING type. A mask of permissible types is passed
2128 and it chooses the "minimal" type to use or an error if not type
2132 *) Add function equivalents to the various macros in asn1.h. The old
2133 macros are retained with an M_ prefix. Code inside the library can
2134 use the M_ macros. External code (including the openssl utility)
2135 should *NOT* in order to be "shared library friendly".
2138 *) Add various functions that can check a certificate's extensions
2139 to see if it usable for various purposes such as SSL client,
2140 server or S/MIME and CAs of these types. This is currently
2141 VERY EXPERIMENTAL but will ultimately be used for certificate chain
2142 verification. Also added a -purpose flag to x509 utility to
2143 print out all the purposes.
2146 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
2150 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
2151 for, obtain and decode and extension and obtain its critical flag.
2152 This allows all the necessary extension code to be handled in a
2153 single function call.
2156 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
2157 platforms. See crypto/rc4/rc4_enc.c for further details.
2160 *) New -noout option to asn1parse. This causes no output to be produced
2161 its main use is when combined with -strparse and -out to extract data
2162 from a file (which may not be in ASN.1 format).
2165 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
2166 when producing the local key id.
2167 [Richard Levitte <levitte@stacken.kth.se>]
2169 *) New option -dhparam in s_server. This allows a DH parameter file to be
2170 stated explicitly. If it is not stated then it tries the first server
2171 certificate file. The previous behaviour hard coded the filename
2175 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
2176 a public key to be input or output. For example:
2177 openssl rsa -in key.pem -pubout -out pubkey.pem
2178 Also added necessary DSA public key functions to handle this.
2181 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
2182 in the message. This was handled by allowing
2183 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
2184 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
2186 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
2187 to the end of the strings whereas this didn't. This would cause problems
2188 if strings read with d2i_ASN1_bytes() were later modified.
2189 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
2191 *) Fix for base64 decode bug. When a base64 bio reads only one line of
2192 data and it contains EOF it will end up returning an error. This is
2193 caused by input 46 bytes long. The cause is due to the way base64
2194 BIOs find the start of base64 encoded data. They do this by trying a
2195 trial decode on each line until they find one that works. When they
2196 do a flag is set and it starts again knowing it can pass all the
2197 data directly through the decoder. Unfortunately it doesn't reset
2198 the context it uses. This means that if EOF is reached an attempt
2199 is made to pass two EOFs through the context and this causes the
2200 resulting error. This can also cause other problems as well. As is
2201 usual with these problems it takes *ages* to find and the fix is
2202 trivial: move one line.
2203 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
2205 *) Ugly workaround to get s_client and s_server working under Windows. The
2206 old code wouldn't work because it needed to select() on sockets and the
2207 tty (for keypresses and to see if data could be written). Win32 only
2208 supports select() on sockets so we select() with a 1s timeout on the
2209 sockets and then see if any characters are waiting to be read, if none
2210 are present then we retry, we also assume we can always write data to
2211 the tty. This isn't nice because the code then blocks until we've
2212 received a complete line of data and it is effectively polling the
2213 keyboard at 1s intervals: however it's quite a bit better than not
2214 working at all :-) A dedicated Windows application might handle this
2215 with an event loop for example.
2218 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
2219 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
2220 will be called when RSA_sign() and RSA_verify() are used. This is useful
2221 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
2222 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
2223 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
2224 This necessitated the support of an extra signature type NID_md5_sha1
2225 for SSL signatures and modifications to the SSL library to use it instead
2226 of calling RSA_public_decrypt() and RSA_private_encrypt().
2229 *) Add new -verify -CAfile and -CApath options to the crl program, these
2230 will lookup a CRL issuers certificate and verify the signature in a
2231 similar way to the verify program. Tidy up the crl program so it
2232 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
2233 less strict. It will now permit CRL extensions even if it is not
2234 a V2 CRL: this will allow it to tolerate some broken CRLs.
2237 *) Initialize all non-automatic variables each time one of the openssl
2238 sub-programs is started (this is necessary as they may be started
2239 multiple times from the "OpenSSL>" prompt).
2240 [Lennart Bang, Bodo Moeller]
2242 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
2243 removing all other RSA functionality (this is what NO_RSA does). This
2244 is so (for example) those in the US can disable those operations covered
2245 by the RSA patent while allowing storage and parsing of RSA keys and RSA
2249 *) Non-copying interface to BIO pairs.
2250 (still largely untested)
2253 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
2254 ASCII string. This was handled independently in various places before.
2257 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
2258 UTF8 strings a character at a time.
2261 *) Use client_version from client hello to select the protocol
2262 (s23_srvr.c) and for RSA client key exchange verification
2263 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
2266 *) Add various utility functions to handle SPKACs, these were previously
2267 handled by poking round in the structure internals. Added new function
2268 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
2269 print, verify and generate SPKACs. Based on an original idea from
2270 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
2273 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
2276 *) Allow the config file extension section to be overwritten on the
2277 command line. Based on an original idea from Massimiliano Pala
2278 <madwolf@comune.modena.it>. The new option is called -extensions
2279 and can be applied to ca, req and x509. Also -reqexts to override
2280 the request extensions in req and -crlexts to override the crl extensions
2284 *) Add new feature to the SPKAC handling in ca. Now you can include
2285 the same field multiple times by preceding it by "XXXX." for example:
2288 this is the same syntax as used in the req config file.
2291 *) Allow certificate extensions to be added to certificate requests. These
2292 are specified in a 'req_extensions' option of the req section of the
2293 config file. They can be printed out with the -text option to req but
2294 are otherwise ignored at present.
2297 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
2298 data read consists of only the final block it would not decrypted because
2299 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
2300 A misplaced 'break' also meant the decrypted final block might not be
2301 copied until the next read.
2304 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
2305 a few extra parameters to the DH structure: these will be useful if
2306 for example we want the value of 'q' or implement X9.42 DH.
2309 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
2310 provides hooks that allow the default DSA functions or functions on a
2311 "per key" basis to be replaced. This allows hardware acceleration and
2312 hardware key storage to be handled without major modification to the
2313 library. Also added low level modexp hooks and CRYPTO_EX structure and
2314 associated functions.
2317 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
2318 as "read only": it can't be written to and the buffer it points to will
2319 not be freed. Reading from a read only BIO is much more efficient than
2320 a normal memory BIO. This was added because there are several times when
2321 an area of memory needs to be read from a BIO. The previous method was
2322 to create a memory BIO and write the data to it, this results in two
2323 copies of the data and an O(n^2) reading algorithm. There is a new
2324 function BIO_new_mem_buf() which creates a read only memory BIO from
2325 an area of memory. Also modified the PKCS#7 routines to use read only
2329 *) Bugfix: ssl23_get_client_hello did not work properly when called in
2330 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
2331 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
2332 but a retry condition occured while trying to read the rest.
2335 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
2336 NID_pkcs7_encrypted by default: this was wrong since this should almost
2337 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
2338 the encrypted data type: this is a more sensible place to put it and it
2339 allows the PKCS#12 code to be tidied up that duplicated this
2343 *) Changed obj_dat.pl script so it takes its input and output files on
2344 the command line. This should avoid shell escape redirection problems
2348 *) Initial support for certificate extension requests, these are included
2349 in things like Xenroll certificate requests. Included functions to allow
2350 extensions to be obtained and added.
2353 *) -crlf option to s_client and s_server for sending newlines as
2354 CRLF (as required by many protocols).
2357 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
2359 *) Install libRSAglue.a when OpenSSL is built with RSAref.
2360 [Ralf S. Engelschall]
2362 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
2363 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
2365 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
2369 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
2370 DH parameters/keys (q is lost during that conversion, but the resulting
2371 DH parameters contain its length).
2373 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
2374 much faster than DH_generate_parameters (which creates parameters
2375 where p = 2*q + 1), and also the smaller q makes DH computations
2376 much more efficient (160-bit exponentiation instead of 1024-bit
2377 exponentiation); so this provides a convenient way to support DHE
2378 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
2379 utter importance to use
2380 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
2382 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
2383 when such DH parameters are used, because otherwise small subgroup
2384 attacks may become possible!
2387 *) Avoid memory leak in i2d_DHparams.
2390 *) Allow the -k option to be used more than once in the enc program:
2391 this allows the same encrypted message to be read by multiple recipients.
2394 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
2395 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
2396 it will always use the numerical form of the OID, even if it has a short
2400 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
2401 method only got called if p,q,dmp1,dmq1,iqmp components were present,
2402 otherwise bn_mod_exp was called. In the case of hardware keys for example
2403 no private key components need be present and it might store extra data
2404 in the RSA structure, which cannot be accessed from bn_mod_exp.
2405 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
2406 private key operations.
2409 *) Added support for SPARC Linux.
2412 *) pem_password_cb function type incompatibly changed from
2413 typedef int pem_password_cb(char *buf, int size, int rwflag);
2415 ....(char *buf, int size, int rwflag, void *userdata);
2416 so that applications can pass data to their callbacks:
2417 The PEM[_ASN1]_{read,write}... functions and macros now take an
2418 additional void * argument, which is just handed through whenever
2419 the password callback is called.
2420 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
2422 New function SSL_CTX_set_default_passwd_cb_userdata.
2424 Compatibility note: As many C implementations push function arguments
2425 onto the stack in reverse order, the new library version is likely to
2426 interoperate with programs that have been compiled with the old
2427 pem_password_cb definition (PEM_whatever takes some data that
2428 happens to be on the stack as its last argument, and the callback
2429 just ignores this garbage); but there is no guarantee whatsoever that
2432 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
2433 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
2434 problems not only on Windows, but also on some Unix platforms.
2435 To avoid problematic command lines, these definitions are now in an
2436 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
2437 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
2440 *) MIPS III/IV assembler module is reimplemented.
2443 *) More DES library cleanups: remove references to srand/rand and
2444 delete an unused file.
2447 *) Add support for the the free Netwide assembler (NASM) under Win32,
2448 since not many people have MASM (ml) and it can be hard to obtain.
2449 This is currently experimental but it seems to work OK and pass all
2450 the tests. Check out INSTALL.W32 for info.
2453 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
2454 without temporary keys kept an extra copy of the server key,
2455 and connections with temporary keys did not free everything in case
2459 *) New function RSA_check_key and new openssl rsa option -check
2460 for verifying the consistency of RSA keys.
2461 [Ulf Moeller, Bodo Moeller]
2463 *) Various changes to make Win32 compile work:
2464 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
2465 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
2466 comparison" warnings.
2467 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
2470 *) Add a debugging option to PKCS#5 v2 key generation function: when
2471 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
2472 derived keys are printed to stderr.
2475 *) Copy the flags in ASN1_STRING_dup().
2476 [Roman E. Pavlov <pre@mo.msk.ru>]
2478 *) The x509 application mishandled signing requests containing DSA
2479 keys when the signing key was also DSA and the parameters didn't match.
2481 It was supposed to omit the parameters when they matched the signing key:
2482 the verifying software was then supposed to automatically use the CA's
2483 parameters if they were absent from the end user certificate.
2485 Omitting parameters is no longer recommended. The test was also
2486 the wrong way round! This was probably due to unusual behaviour in
2487 EVP_cmp_parameters() which returns 1 if the parameters match.
2488 This meant that parameters were omitted when they *didn't* match and
2489 the certificate was useless. Certificates signed with 'ca' didn't have
2491 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
2493 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
2494 The interface is as follows:
2495 Applications can use
2496 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
2497 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
2498 "off" is now the default.
2499 The library internally uses
2500 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
2501 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
2502 to disable memory-checking temporarily.
2504 Some inconsistent states that previously were possible (and were
2505 even the default) are now avoided.
2507 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
2508 with each memory chunk allocated; this is occasionally more helpful
2509 than just having a counter.
2511 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
2513 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
2517 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
2518 which largely parallels "options", but is for changing API behaviour,
2519 whereas "options" are about protocol behaviour.
2520 Initial "mode" flags are:
2522 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
2523 a single record has been written.
2524 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
2525 retries use the same buffer location.
2526 (But all of the contents must be
2530 *) Bugfix: SSL_set_mode ignored its parameter, only SSL_CTX_set_mode
2533 *) Fix problems with no-hmac etc.
2534 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
2536 *) New functions RSA_get_default_method(), RSA_set_method() and
2537 RSA_get_method(). These allows replacement of RSA_METHODs without having
2538 to mess around with the internals of an RSA structure.
2541 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
2542 Also really enable memory leak checks in openssl.c and in some
2544 [Chad C. Mulligan, Bodo Moeller]
2546 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
2547 up the length of negative integers. This has now been simplified to just
2548 store the length when it is first determined and use it later, rather
2549 than trying to keep track of where data is copied and updating it to
2551 [Steve Henson, reported by Brien Wheeler
2552 <bwheeler@authentica-security.com>]
2554 *) Add a new function PKCS7_signatureVerify. This allows the verification
2555 of a PKCS#7 signature but with the signing certificate passed to the
2556 function itself. This contrasts with PKCS7_dataVerify which assumes the
2557 certificate is present in the PKCS#7 structure. This isn't always the
2558 case: certificates can be omitted from a PKCS#7 structure and be
2559 distributed by "out of band" means (such as a certificate database).
2562 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
2563 function prototypes in pem.h, also change util/mkdef.pl to add the
2564 necessary function names.
2567 *) mk1mf.pl (used by Windows builds) did not properly read the
2568 options set by Configure in the top level Makefile, and Configure
2569 was not even able to write more than one option correctly.
2570 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
2573 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
2574 file to be loaded from a BIO or FILE pointer. The BIO version will
2575 for example allow memory BIOs to contain config info.
2578 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
2579 Whoever hopes to achieve shared-library compatibility across versions
2580 must use this, not the compile-time macro.
2581 (Exercise 0.9.4: Which is the minimum library version required by
2583 Note: All this applies only to multi-threaded programs, others don't
2587 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
2588 through a BIO pair triggered the default case, i.e.
2589 SSLerr(...,SSL_R_UNKNOWN_STATE).
2592 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
2593 can use the SSL library even if none of the specific BIOs is
2597 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
2598 for the encoded length.
2599 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
2601 *) Add initial documentation of the X509V3 functions.
2604 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
2605 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
2606 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
2607 secure PKCS#8 private key format with a high iteration count.
2610 *) Fix determination of Perl interpreter: A perl or perl5
2611 _directory_ in $PATH was also accepted as the interpreter.
2612 [Ralf S. Engelschall]
2614 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
2615 wrong with it but it was very old and did things like calling
2616 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
2620 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
2621 to use the new extension code.
2624 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
2625 with macros. This should make it easier to change their form, add extra
2626 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
2630 *) Add to configuration table a new entry that can specify an alternative
2631 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
2632 according to Mark Crispin <MRC@Panda.COM>.
2636 *) DES CBC did not update the IV. Weird.
2639 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
2640 Changing the behaviour of the former might break existing programs --
2641 where IV updating is needed, des_ncbc_encrypt can be used.
2644 *) When bntest is run from "make test" it drives bc to check its
2645 calculations, as well as internally checking them. If an internal check
2646 fails, it needs to cause bc to give a non-zero result or make test carries
2647 on without noticing the failure. Fixed.
2650 *) DES library cleanups.
2653 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
2654 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
2655 ciphers. NOTE: although the key derivation function has been verified
2656 against some published test vectors it has not been extensively tested
2657 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
2661 *) Instead of "mkdir -p", which is not fully portable, use new
2662 Perl script "util/mkdir-p.pl".
2665 *) Rewrite the way password based encryption (PBE) is handled. It used to
2666 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
2667 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
2668 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
2669 the 'parameter' field of the AlgorithmIdentifier is passed to the
2670 underlying key generation function so it must do its own ASN1 parsing.
2671 This has also changed the EVP_PBE_CipherInit() function which now has a
2672 'parameter' argument instead of literal salt and iteration count values
2673 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
2676 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
2677 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
2678 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
2679 KEY" because this clashed with PKCS#8 unencrypted string. Since this
2680 value was just used as a "magic string" and not used directly its
2681 value doesn't matter.
2684 *) Introduce some semblance of const correctness to BN. Shame C doesn't
2688 *) "linux-sparc64" configuration (ultrapenguin).
2689 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
2690 "linux-sparc" configuration.
2691 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
2693 *) config now generates no-xxx options for missing ciphers.
2696 *) Support the EBCDIC character set (work in progress).
2697 File ebcdic.c not yet included because it has a different license.
2698 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
2700 *) Support BS2000/OSD-POSIX.
2701 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
2703 *) Make callbacks for key generation use void * instead of char *.
2706 *) Make S/MIME samples compile (not yet tested).
2709 *) Additional typesafe stacks.
2712 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
2716 Changes between 0.9.3 and 0.9.3a [29 May 1999]
2718 *) New configuration variant "sco5-gcc".
2720 *) Updated some demos.
2721 [Sean O Riordain, Wade Scholine]
2723 *) Add missing BIO_free at exit of pkcs12 application.
2726 *) Fix memory leak in conf.c.
2729 *) Updates for Win32 to assembler version of MD5.
2732 *) Set #! path to perl in apps/der_chop to where we found it
2733 instead of using a fixed path.
2736 *) SHA library changes for irix64-mips4-cc.
2739 *) Improvements for VMS support.
2743 Changes between 0.9.2b and 0.9.3 [24 May 1999]
2745 *) Bignum library bug fix. IRIX 6 passes "make test" now!
2746 This also avoids the problems with SC4.2 and unpatched SC5.
2747 [Andy Polyakov <appro@fy.chalmers.se>]
2749 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
2750 These are required because of the typesafe stack would otherwise break
2751 existing code. If old code used a structure member which used to be STACK
2752 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
2753 sk_num or sk_value it would produce an error because the num, data members
2754 are not present in STACK_OF. Now it just produces a warning. sk_set
2755 replaces the old method of assigning a value to sk_value
2756 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
2757 that does this will no longer work (and should use sk_set instead) but
2758 this could be regarded as a "questionable" behaviour anyway.
2761 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
2762 correctly handle encrypted S/MIME data.
2765 *) Change type of various DES function arguments from des_cblock
2766 (which means, in function argument declarations, pointer to char)
2767 to des_cblock * (meaning pointer to array with 8 char elements),
2768 which allows the compiler to do more typechecking; it was like
2769 that back in SSLeay, but with lots of ugly casts.
2771 Introduce new type const_des_cblock.
2774 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
2775 problems: find RecipientInfo structure that matches recipient certificate
2776 and initialise the ASN1 structures properly based on passed cipher.
2779 *) Belatedly make the BN tests actually check the results.
2782 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
2783 to and from BNs: it was completely broken. New compilation option
2784 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
2785 key elements as negative integers.
2788 *) Reorganize and speed up MD5.
2789 [Andy Polyakov <appro@fy.chalmers.se>]
2792 [Richard Levitte <richard@levitte.org>]
2794 *) New option -out to asn1parse to allow the parsed structure to be
2795 output to a file. This is most useful when combined with the -strparse
2796 option to examine the output of things like OCTET STRINGS.
2799 *) Make SSL library a little more fool-proof by not requiring any longer
2800 that SSL_set_{accept,connect}_state be called before
2801 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
2802 in many applications because usually everything *appeared* to work as
2803 intended anyway -- now it really works as intended).
2806 *) Move openssl.cnf out of lib/.
2809 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
2810 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
2811 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
2812 [Ralf S. Engelschall]
2814 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
2815 handle PKCS#7 enveloped data properly.
2816 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
2818 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
2819 copying pointers. The cert_st handling is changed by this in
2820 various ways (and thus what used to be known as ctx->default_cert
2821 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
2822 any longer when s->cert does not give us what we need).
2823 ssl_cert_instantiate becomes obsolete by this change.
2824 As soon as we've got the new code right (possibly it already is?),
2825 we have solved a couple of bugs of the earlier code where s->cert
2826 was used as if it could not have been shared with other SSL structures.
2828 Note that using the SSL API in certain dirty ways now will result
2829 in different behaviour than observed with earlier library versions:
2830 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
2831 does not influence s as it used to.
2833 In order to clean up things more thoroughly, inside SSL_SESSION
2834 we don't use CERT any longer, but a new structure SESS_CERT
2835 that holds per-session data (if available); currently, this is
2836 the peer's certificate chain and, for clients, the server's certificate
2837 and temporary key. CERT holds only those values that can have
2838 meaningful defaults in an SSL_CTX.
2841 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
2842 from the internal representation. Various PKCS#7 fixes: remove some
2843 evil casts and set the enc_dig_alg field properly based on the signing
2847 *) Allow PKCS#12 password to be set from the command line or the
2848 environment. Let 'ca' get its config file name from the environment
2849 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
2853 *) Allow certificate policies extension to use an IA5STRING for the
2854 organization field. This is contrary to the PKIX definition but
2855 VeriSign uses it and IE5 only recognises this form. Document 'x509'
2859 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
2860 without disallowing inline assembler and the like for non-pedantic builds.
2863 *) Support Borland C++ builder.
2864 [Janez Jere <jj@void.si>, modified by Ulf Möller]
2869 *) SHA-1 cleanups and performance enhancements.
2870 [Andy Polyakov <appro@fy.chalmers.se>]
2872 *) Sparc v8plus assembler for the bignum library.
2873 [Andy Polyakov <appro@fy.chalmers.se>]
2875 *) Accept any -xxx and +xxx compiler options in Configure.
2878 *) Update HPUX configuration.
2881 *) Add missing sk_<type>_unshift() function to safestack.h
2882 [Ralf S. Engelschall]
2884 *) New function SSL_CTX_use_certificate_chain_file that sets the
2885 "extra_cert"s in addition to the certificate. (This makes sense
2886 only for "PEM" format files, as chains as a whole are not
2890 *) Support verify_depth from the SSL API.
2891 x509_vfy.c had what can be considered an off-by-one-error:
2892 Its depth (which was not part of the external interface)
2893 was actually counting the number of certificates in a chain;
2894 now it really counts the depth.
2897 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
2898 instead of X509err, which often resulted in confusing error
2899 messages since the error codes are not globally unique
2900 (e.g. an alleged error in ssl3_accept when a certificate
2901 didn't match the private key).
2903 *) New function SSL_CTX_set_session_id_context that allows to set a default
2904 value (so that you don't need SSL_set_session_id_context for each
2905 connection using the SSL_CTX).
2908 *) OAEP decoding bug fix.
2911 *) Support INSTALL_PREFIX for package builders, as proposed by
2915 *) New Configure options "threads" and "no-threads". For systems
2916 where the proper compiler options are known (currently Solaris
2917 and Linux), "threads" is the default.
2920 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
2923 *) Install various scripts to $(OPENSSLDIR)/misc, not to
2924 $(INSTALLTOP)/bin -- they shouldn't clutter directories
2925 such as /usr/local/bin.
2928 *) "make linux-shared" to build shared libraries.
2929 [Niels Poppe <niels@netbox.org>]
2931 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
2934 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
2935 extension adding in x509 utility.
2938 *) Remove NOPROTO sections and error code comments.
2941 *) Partial rewrite of the DEF file generator to now parse the ANSI
2945 *) New Configure options --prefix=DIR and --openssldir=DIR.
2948 *) Complete rewrite of the error code script(s). It is all now handled
2949 by one script at the top level which handles error code gathering,
2950 header rewriting and C source file generation. It should be much better
2951 than the old method: it now uses a modified version of Ulf's parser to
2952 read the ANSI prototypes in all header files (thus the old K&R definitions
2953 aren't needed for error creation any more) and do a better job of
2954 translating function codes into names. The old 'ASN1 error code imbedded
2955 in a comment' is no longer necessary and it doesn't use .err files which
2956 have now been deleted. Also the error code call doesn't have to appear all
2957 on one line (which resulted in some large lines...).
2960 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
2963 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
2964 0 (which usually indicates a closed connection), but continue reading.
2967 *) Fix some race conditions.
2970 *) Add support for CRL distribution points extension. Add Certificate
2971 Policies and CRL distribution points documentation.
2974 *) Move the autogenerated header file parts to crypto/opensslconf.h.
2977 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
2978 8 of keying material. Merlin has also confirmed interop with this fix
2979 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
2980 [Merlin Hughes <merlin@baltimore.ie>]
2982 *) Fix lots of warnings.
2983 [Richard Levitte <levitte@stacken.kth.se>]
2985 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
2986 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
2987 [Richard Levitte <levitte@stacken.kth.se>]
2989 *) Fix problems with sizeof(long) == 8.
2990 [Andy Polyakov <appro@fy.chalmers.se>]
2992 *) Change functions to ANSI C.
2995 *) Fix typos in error codes.
2996 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
2998 *) Remove defunct assembler files from Configure.
3001 *) SPARC v8 assembler BIGNUM implementation.
3002 [Andy Polyakov <appro@fy.chalmers.se>]
3004 *) Support for Certificate Policies extension: both print and set.
3005 Various additions to support the r2i method this uses.
3008 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
3009 return a const string when you are expecting an allocated buffer.
3012 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
3013 types DirectoryString and DisplayText.
3016 *) Add code to allow r2i extensions to access the configuration database,
3017 add an LHASH database driver and add several ctx helper functions.
3020 *) Fix an evil bug in bn_expand2() which caused various BN functions to
3021 fail when they extended the size of a BIGNUM.
3024 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
3025 support typesafe stack.
3028 *) Fix typo in SSL_[gs]et_options().
3029 [Nils Frostberg <nils@medcom.se>]
3031 *) Delete various functions and files that belonged to the (now obsolete)
3032 old X509V3 handling code.
3035 *) New Configure option "rsaref".
3038 *) Don't auto-generate pem.h.
3041 *) Introduce type-safe ASN.1 SETs.
3044 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
3045 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
3047 *) Introduce type-safe STACKs. This will almost certainly break lots of code
3048 that links with OpenSSL (well at least cause lots of warnings), but fear
3049 not: the conversion is trivial, and it eliminates loads of evil casts. A
3050 few STACKed things have been converted already. Feel free to convert more.
3051 In the fullness of time, I'll do away with the STACK type altogether.
3054 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
3055 specified in <certfile> by updating the entry in the index.txt file.
3056 This way one no longer has to edit the index.txt file manually for
3057 revoking a certificate. The -revoke option does the gory details now.
3058 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
3060 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
3061 `-text' option at all and this way the `-noout -text' combination was
3062 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
3063 [Ralf S. Engelschall]
3065 *) Make sure a corresponding plain text error message exists for the
3066 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
3067 verify callback function determined that a certificate was revoked.
3068 [Ralf S. Engelschall]
3070 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
3071 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
3072 all available cipers including rc5, which was forgotten until now.
3073 In order to let the testing shell script know which algorithms
3074 are available, a new (up to now undocumented) command
3075 "openssl list-cipher-commands" is used.
3078 *) Bugfix: s_client occasionally would sleep in select() when
3079 it should have checked SSL_pending() first.
3082 *) New functions DSA_do_sign and DSA_do_verify to provide access to
3083 the raw DSA values prior to ASN.1 encoding.
3086 *) Tweaks to Configure
3087 [Niels Poppe <niels@netbox.org>]
3089 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
3093 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
3096 *) New config option to avoid instructions that are illegal on the 80386.
3097 The default code is faster, but requires at least a 486.
3100 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
3101 SSL2_SERVER_VERSION (not used at all) macros, which are now the
3102 same as SSL2_VERSION anyway.
3105 *) New "-showcerts" option for s_client.
3108 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
3109 application. Various cleanups and fixes.
3112 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
3113 modify error routines to work internally. Add error codes and PBE init
3114 to library startup routines.
3117 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
3118 packing functions to asn1 and evp. Changed function names and error
3119 codes along the way.
3122 *) PKCS12 integration: and so it begins... First of several patches to
3123 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
3124 objects to objects.h
3127 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
3128 and display support for Thawte strong extranet extension.
3131 *) Add LinuxPPC support.
3132 [Jeff Dubrule <igor@pobox.org>]
3134 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
3135 bn_div_words in alpha.s.
3136 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
3138 *) Make sure the RSA OAEP test is skipped under -DRSAref because
3139 OAEP isn't supported when OpenSSL is built with RSAref.
3140 [Ulf Moeller <ulf@fitug.de>]
3142 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
3143 so they no longer are missing under -DNOPROTO.
3144 [Soren S. Jorvang <soren@t.dk>]
3147 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
3149 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
3150 doesn't work when the session is reused. Coming soon!
3153 *) Fix a security hole, that allows sessions to be reused in the wrong
3154 context thus bypassing client cert protection! All software that uses
3155 client certs and session caches in multiple contexts NEEDS PATCHING to
3156 allow session reuse! A fuller solution is in the works.
3157 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
3159 *) Some more source tree cleanups (removed obsolete files
3160 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
3161 permission on "config" script to be executable) and a fix for the INSTALL
3163 [Ulf Moeller <ulf@fitug.de>]
3165 *) Remove some legacy and erroneous uses of malloc, free instead of
3167 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
3169 *) Make rsa_oaep_test return non-zero on error.
3170 [Ulf Moeller <ulf@fitug.de>]
3172 *) Add support for native Solaris shared libraries. Configure
3173 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
3174 if someone would make that last step automatic.
3175 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
3177 *) ctx_size was not built with the right compiler during "make links". Fixed.
3180 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
3181 except NULL ciphers". This means the default cipher list will no longer
3182 enable NULL ciphers. They need to be specifically enabled e.g. with
3183 the string "DEFAULT:eNULL".
3186 *) Fix to RSA private encryption routines: if p < q then it would
3187 occasionally produce an invalid result. This will only happen with
3188 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
3191 *) Be less restrictive and allow also `perl util/perlpath.pl
3192 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
3193 because this way one can also use an interpreter named `perl5' (which is
3194 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
3195 installed as `perl').
3196 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
3198 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
3199 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
3201 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
3202 advapi32.lib to Win32 build and change the pem test comparision
3203 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
3204 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
3205 and crypto/des/ede_cbcm_enc.c.
3208 *) DES quad checksum was broken on big-endian architectures. Fixed.
3211 *) Comment out two functions in bio.h that aren't implemented. Fix up the
3212 Win32 test batch file so it (might) work again. The Win32 test batch file
3213 is horrible: I feel ill....
3216 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
3217 in e_os.h. Audit of header files to check ANSI and non ANSI
3218 sections: 10 functions were absent from non ANSI section and not exported
3219 from Windows DLLs. Fixed up libeay.num for new functions.
3222 *) Make `openssl version' output lines consistent.
3223 [Ralf S. Engelschall]
3225 *) Fix Win32 symbol export lists for BIO functions: Added
3226 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
3227 to ms/libeay{16,32}.def.
3228 [Ralf S. Engelschall]
3230 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
3231 fine under Unix and passes some trivial tests I've now added. But the
3232 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
3233 added to make sure no one expects that this stuff really works in the
3234 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
3235 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
3237 [Ralf S. Engelschall]
3239 *) Fix the generation of two part addresses in perl.
3240 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
3242 *) Add config entry for Linux on MIPS.
3243 [John Tobey <jtobey@channel1.com>]
3245 *) Make links whenever Configure is run, unless we are on Windoze.
3248 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
3249 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
3253 *) Add a useful kludge to allow package maintainers to specify compiler and
3254 other platforms details on the command line without having to patch the
3255 Configure script everytime: One now can use ``perl Configure
3256 <id>:<details>'', i.e. platform ids are allowed to have details appended
3257 to them (seperated by colons). This is treated as there would be a static
3258 pre-configured entry in Configure's %table under key <id> with value
3259 <details> and ``perl Configure <id>'' is called. So, when you want to
3260 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
3261 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
3262 now, which overrides the FreeBSD-elf entry on-the-fly.
3263 [Ralf S. Engelschall]
3265 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
3268 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
3269 on the `perl Configure ...' command line. This way one can compile
3270 OpenSSL libraries with Position Independent Code (PIC) which is needed
3271 for linking it into DSOs.
3272 [Ralf S. Engelschall]
3274 *) Remarkably, export ciphers were totally broken and no-one had noticed!
3278 *) Cleaned up the LICENSE document: The official contact for any license
3279 questions now is the OpenSSL core team under openssl-core@openssl.org.
3280 And add a paragraph about the dual-license situation to make sure people
3281 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
3282 to the OpenSSL toolkit.
3283 [Ralf S. Engelschall]
3285 *) General source tree makefile cleanups: Made `making xxx in yyy...'
3286 display consistent in the source tree and replaced `/bin/rm' by `rm'.
3287 Additonally cleaned up the `make links' target: Remove unnecessary
3288 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
3289 to speed processing and no longer clutter the display with confusing
3290 stuff. Instead only the actually done links are displayed.
3291 [Ralf S. Engelschall]
3293 *) Permit null encryption ciphersuites, used for authentication only. It used
3294 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
3295 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
3299 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
3300 signed attributes when verifying signatures (this would break them),
3301 the detached data encoding was wrong and public keys obtained using
3302 X509_get_pubkey() weren't freed.
3305 *) Add text documentation for the BUFFER functions. Also added a work around
3306 to a Win95 console bug. This was triggered by the password read stuff: the
3307 last character typed gets carried over to the next fread(). If you were
3308 generating a new cert request using 'req' for example then the last
3309 character of the passphrase would be CR which would then enter the first
3313 *) Added the new `Includes OpenSSL Cryptography Software' button as
3314 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
3315 button and can be used by applications based on OpenSSL to show the
3316 relationship to the OpenSSL project.
3317 [Ralf S. Engelschall]
3319 *) Remove confusing variables in function signatures in files
3320 ssl/ssl_lib.c and ssl/ssl.h.
3321 [Lennart Bong <lob@kulthea.stacken.kth.se>]
3323 *) Don't install bss_file.c under PREFIX/include/
3324 [Lennart Bong <lob@kulthea.stacken.kth.se>]
3326 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
3327 functions that return function pointers and has support for NT specific
3328 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
3329 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
3330 unsigned to signed types: this was killing the Win32 compile.
3333 *) Add new certificate file to stack functions,
3334 SSL_add_dir_cert_subjects_to_stack() and
3335 SSL_add_file_cert_subjects_to_stack(). These largely supplant
3336 SSL_load_client_CA_file(), and can be used to add multiple certs easily
3337 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
3338 This means that Apache-SSL and similar packages don't have to mess around
3339 to add as many CAs as they want to the preferred list.
3342 *) Experiment with doxygen documentation. Currently only partially applied to
3344 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
3345 openssl.doxy as the configuration file.
3348 *) Get rid of remaining C++-style comments which strict C compilers hate.
3349 [Ralf S. Engelschall, pointed out by Carlos Amengual]
3351 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
3352 compiled in by default: it has problems with large keys.
3355 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
3356 DH private keys and/or callback functions which directly correspond to
3357 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
3358 is needed for applications which have to configure certificates on a
3359 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
3361 For the RSA certificate situation is makes no difference, but
3362 for the DSA certificate situation this fixes the "no shared cipher"
3363 problem where the OpenSSL cipher selection procedure failed because the
3364 temporary keys were not overtaken from the context and the API provided
3365 no way to reconfigure them.
3366 The new functions now let applications reconfigure the stuff and they
3367 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
3368 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
3369 non-public-API function ssl_cert_instantiate() is used as a helper
3370 function and also to reduce code redundancy inside ssl_rsa.c.
3371 [Ralf S. Engelschall]
3373 *) Move s_server -dcert and -dkey options out of the undocumented feature
3374 area because they are useful for the DSA situation and should be
3375 recognized by the users.
3376 [Ralf S. Engelschall]
3378 *) Fix the cipher decision scheme for export ciphers: the export bits are
3379 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
3380 SSL_EXP_MASK. So, the original variable has to be used instead of the
3381 already masked variable.
3382 [Richard Levitte <levitte@stacken.kth.se>]
3384 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
3385 [Richard Levitte <levitte@stacken.kth.se>]
3387 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
3388 from `int' to `unsigned int' because it's a length and initialized by
3389 EVP_DigestFinal() which expects an `unsigned int *'.
3390 [Richard Levitte <levitte@stacken.kth.se>]
3392 *) Don't hard-code path to Perl interpreter on shebang line of Configure
3393 script. Instead use the usual Shell->Perl transition trick.
3394 [Ralf S. Engelschall]
3396 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
3397 (in addition to RSA certificates) to match the behaviour of `openssl dsa
3398 -noout -modulus' as it's already the case for `openssl rsa -noout
3399 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
3400 currently the public key is printed (a decision which was already done by
3401 `openssl dsa -modulus' in the past) which serves a similar purpose.
3402 Additionally the NO_RSA no longer completely removes the whole -modulus
3403 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
3405 [Ralf S. Engelschall]
3407 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
3408 BIO. See the source (crypto/evp/bio_ok.c) for more info.
3409 [Arne Ansper <arne@ats.cyber.ee>]
3411 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
3412 to be added. Now both 'req' and 'ca' can use new objects defined in the
3416 *) Add cool BIO that does syslog (or event log on NT).
3417 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
3419 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
3420 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
3421 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
3422 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
3425 *) Add preliminary config info for new extension code.
3428 *) Make RSA_NO_PADDING really use no padding.
3429 [Ulf Moeller <ulf@fitug.de>]
3431 *) Generate errors when private/public key check is done.
3434 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
3435 for some CRL extensions and new objects added.
3438 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
3439 key usage extension and fuller support for authority key id.
3442 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
3443 padding method for RSA, which is recommended for new applications in PKCS
3444 #1 v2.0 (RFC 2437, October 1998).
3445 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
3446 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
3447 against Bleichbacher's attack on RSA.
3448 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
3451 *) Updates to the new SSL compression code
3452 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3454 *) Fix so that the version number in the master secret, when passed
3455 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
3456 (because the server will not accept higher), that the version number
3457 is 0x03,0x01, not 0x03,0x00
3458 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3460 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
3461 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
3462 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
3465 *) Support for RAW extensions where an arbitrary extension can be
3466 created by including its DER encoding. See apps/openssl.cnf for
3470 *) Make sure latest Perl versions don't interpret some generated C array
3471 code as Perl array code in the crypto/err/err_genc.pl script.
3472 [Lars Weber <3weber@informatik.uni-hamburg.de>]
3474 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
3475 not many people have the assembler. Various Win32 compilation fixes and
3476 update to the INSTALL.W32 file with (hopefully) more accurate Win32
3480 *) Modify configure script 'Configure' to automatically create crypto/date.h
3481 file under Win32 and also build pem.h from pem.org. New script
3482 util/mkfiles.pl to create the MINFO file on environments that can't do a
3483 'make files': perl util/mkfiles.pl >MINFO should work.
3486 *) Major rework of DES function declarations, in the pursuit of correctness
3487 and purity. As a result, many evil casts evaporated, and some weirdness,
3488 too. You may find this causes warnings in your code. Zapping your evil
3489 casts will probably fix them. Mostly.
3492 *) Fix for a typo in asn1.h. Bug fix to object creation script
3493 obj_dat.pl. It considered a zero in an object definition to mean
3494 "end of object": none of the objects in objects.h have any zeros
3495 so it wasn't spotted.
3496 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
3498 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
3499 Masking (CBCM). In the absence of test vectors, the best I have been able
3500 to do is check that the decrypt undoes the encrypt, so far. Send me test
3501 vectors if you have them.
3504 *) Correct calculation of key length for export ciphers (too much space was
3505 allocated for null ciphers). This has not been tested!
3508 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
3509 message is now correct (it understands "crypto" and "ssl" on its
3510 command line). There is also now an "update" option. This will update
3511 the util/ssleay.num and util/libeay.num files with any new functions.
3513 perl util/mkdef.pl crypto ssl update
3514 it will update them.
3517 *) Overhauled the Perl interface (perl/*):
3518 - ported BN stuff to OpenSSL's different BN library
3519 - made the perl/ source tree CVS-aware
3520 - renamed the package from SSLeay to OpenSSL (the files still contain
3521 their history because I've copied them in the repository)
3522 - removed obsolete files (the test scripts will be replaced
3523 by better Test::Harness variants in the future)
3524 [Ralf S. Engelschall]
3526 *) First cut for a very conservative source tree cleanup:
3527 1. merge various obsolete readme texts into doc/ssleay.txt
3528 where we collect the old documents and readme texts.
3529 2. remove the first part of files where I'm already sure that we no
3530 longer need them because of three reasons: either they are just temporary
3531 files which were left by Eric or they are preserved original files where
3532 I've verified that the diff is also available in the CVS via "cvs diff
3533 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
3534 the crypto/md/ stuff).
3535 [Ralf S. Engelschall]
3537 *) More extension code. Incomplete support for subject and issuer alt
3538 name, issuer and authority key id. Change the i2v function parameters
3539 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
3540 what that's for :-) Fix to ASN1 macro which messed up
3541 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
3544 *) Preliminary support for ENUMERATED type. This is largely copied from the
3548 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
3549 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3551 *) Make sure `make rehash' target really finds the `openssl' program.
3552 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
3554 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
3555 like to hear about it if this slows down other processors.
3558 *) Add CygWin32 platform information to Configure script.
3559 [Alan Batie <batie@aahz.jf.intel.com>]
3561 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
3562 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
3564 *) New program nseq to manipulate netscape certificate sequences
3567 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
3571 *) Fixes to BN code. Previously the default was to define BN_RECURSION
3572 but the BN code had some problems that would cause failures when
3573 doing certificate verification and some other functions.
3574 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3576 *) Add ASN1 and PEM code to support netscape certificate sequences.
3579 *) Add ASN1 and PEM code to support netscape certificate sequences.
3582 *) Add several PKIX and private extended key usage OIDs.
3585 *) Modify the 'ca' program to handle the new extension code. Modify
3586 openssl.cnf for new extension format, add comments.
projects / openssl.git / blob