oweals/openssl.git
23 years agoSome functions, like strdup() and strcasecmp(), are defined in
Richard Levitte [Tue, 20 Feb 2001 19:05:59 +0000 (19:05 +0000)]
Some functions, like strdup() and strcasecmp(), are defined in
strings.h according to X/Open.

23 years agoI forgot there was a reason why the inclusions and definition of u_int
Richard Levitte [Tue, 20 Feb 2001 17:37:02 +0000 (17:37 +0000)]
I forgot there was a reason why the inclusions and definition of u_int
was made in a certain sequence.  This change restores the earlier
"chain of command".

23 years agoOpenVMS catches up.
Richard Levitte [Tue, 20 Feb 2001 17:14:30 +0000 (17:14 +0000)]
OpenVMS catches up.

23 years agoThe callbacks in the NAME_FUNCS structure are not used directly as LHASH
Geoff Thorpe [Tue, 20 Feb 2001 16:31:15 +0000 (16:31 +0000)]
The callbacks in the NAME_FUNCS structure are not used directly as LHASH
callbacks, and their prototypes were consistent as they were. These casts
need reversing.

Also, I personally find line breaks during parameter lists (ie a line
ending in a comma) easier to read at a glance than line breaks at the end
of a function call and before a dereference on the return value (ie a line
ending in a closed-bracket followed by a line starting with "->").

23 years agomake it a loop as in dsa
Ulf Möller [Tue, 20 Feb 2001 16:17:25 +0000 (16:17 +0000)]
make it a loop as in dsa

23 years agoGet e_os2.h to get all the system definitions correctly.
Richard Levitte [Tue, 20 Feb 2001 14:07:03 +0000 (14:07 +0000)]
Get e_os2.h to get all the system definitions correctly.

23 years agoGet the right cast for lhash callback functions.
Richard Levitte [Tue, 20 Feb 2001 14:00:29 +0000 (14:00 +0000)]
Get the right cast for lhash callback functions.

23 years agoInclude string.h so mem*() functions get properly declared.
Richard Levitte [Tue, 20 Feb 2001 13:41:11 +0000 (13:41 +0000)]
Include string.h so mem*() functions get properly declared.

23 years agoFix typo.
Dr. Stephen Henson [Tue, 20 Feb 2001 13:30:28 +0000 (13:30 +0000)]
Fix typo.

23 years agoA new bunch of too long symbols to hack.
Richard Levitte [Tue, 20 Feb 2001 13:26:00 +0000 (13:26 +0000)]
A new bunch of too long symbols to hack.
OCSP_CRLID_new and OCSP_crlID_new clash on case-insensitive systems.

23 years agoDEC C on VMS is pedantic by definition.
Richard Levitte [Tue, 20 Feb 2001 13:24:23 +0000 (13:24 +0000)]
DEC C on VMS is pedantic by definition.

23 years agoInclude OpenSSL header files earlier so macros like OPENSSL_SYS_VMS
Richard Levitte [Tue, 20 Feb 2001 13:23:42 +0000 (13:23 +0000)]
Include OpenSSL header files earlier so macros like OPENSSL_SYS_VMS
get a chance to be defined.
Make a batter file name translator (uhm, no, that's not the finished
variant :-)).

23 years agoInclude OpenSSL header files earlier so macros like OPENSSL_SYS_VMS
Richard Levitte [Tue, 20 Feb 2001 13:22:35 +0000 (13:22 +0000)]
Include OpenSSL header files earlier so macros like OPENSSL_SYS_VMS
get a chance to be defined.

23 years agoInclude opensslconf.h or the like early to make sure system macros get
Richard Levitte [Tue, 20 Feb 2001 13:11:54 +0000 (13:11 +0000)]
Include opensslconf.h or the like early to make sure system macros get
correctly defined.

23 years agoLet VMS catch up.
Richard Levitte [Tue, 20 Feb 2001 13:10:14 +0000 (13:10 +0000)]
Let VMS catch up.

23 years agoDEC C on VMS is pedantic by definition.
Richard Levitte [Tue, 20 Feb 2001 13:06:55 +0000 (13:06 +0000)]
DEC C on VMS is pedantic by definition.

23 years agoUse sk_*_new_null() instead of sk_*_new(NULL). That avoids getting
Richard Levitte [Tue, 20 Feb 2001 13:06:10 +0000 (13:06 +0000)]
Use sk_*_new_null() instead of sk_*_new(NULL).  That avoids getting
lots of silly warnings from the compiler.

23 years agohonour '-no_tmp_rsa'
Bodo Möller [Tue, 20 Feb 2001 12:59:48 +0000 (12:59 +0000)]
honour '-no_tmp_rsa'

23 years agoInclude string.h so mem* functions get properly declared.
Richard Levitte [Tue, 20 Feb 2001 12:51:56 +0000 (12:51 +0000)]
Include string.h so mem* functions get properly declared.

23 years agoInclude OpenSSL header files earlier so macros like OPENSSL_SYS_VMS
Richard Levitte [Tue, 20 Feb 2001 12:44:46 +0000 (12:44 +0000)]
Include OpenSSL header files earlier so macros like OPENSSL_SYS_VMS
get a chance to be defined.

23 years agoInclude string.h so mem* functions get properly declared.
Richard Levitte [Tue, 20 Feb 2001 12:43:11 +0000 (12:43 +0000)]
Include string.h so mem* functions get properly declared.

23 years agoUse 0 instead of NULL, at least for function casts, since there are
Richard Levitte [Tue, 20 Feb 2001 12:40:42 +0000 (12:40 +0000)]
Use 0 instead of NULL, at least for function casts, since there are
variants of stdio.h that define NULL in such a way that it's "unsafe"
to use for function pointer casting.

23 years agoupdate
Bodo Möller [Tue, 20 Feb 2001 11:36:02 +0000 (11:36 +0000)]
update

23 years agoUse new-style system-id macros everywhere possible. I hope I haven't
Richard Levitte [Tue, 20 Feb 2001 08:13:47 +0000 (08:13 +0000)]
Use new-style system-id macros everywhere possible.  I hope I haven't
missed any.

This compiles and runs on Linux, and external applications have no
problems with it.  The definite test will be to build this on VMS.

23 years agoFix BN_[pseudo_]rand: 'mask' must be used even if top=-1.
Bodo Möller [Tue, 20 Feb 2001 08:10:38 +0000 (08:10 +0000)]
Fix BN_[pseudo_]rand: 'mask' must be used even if top=-1.

Mention BN_[pseudo_]rand with top=-1 in CHANGES.

23 years agoUse new-style system-id macros.
Richard Levitte [Tue, 20 Feb 2001 07:43:22 +0000 (07:43 +0000)]
Use new-style system-id macros.

23 years agoI'm sick of the warnings about long long...
Richard Levitte [Tue, 20 Feb 2001 07:22:11 +0000 (07:22 +0000)]
I'm sick of the warnings about long long...

23 years agoUse BN_rand_range().
Ulf Möller [Tue, 20 Feb 2001 00:43:59 +0000 (00:43 +0000)]
Use BN_rand_range().

23 years agoFix warning.
Ulf Möller [Tue, 20 Feb 2001 00:43:03 +0000 (00:43 +0000)]
Fix warning.

23 years agoBN_rand_range() needs a BN_rand() variant that doesn't set the MSB.
Ulf Möller [Tue, 20 Feb 2001 00:23:07 +0000 (00:23 +0000)]
BN_rand_range() needs a BN_rand() variant that doesn't set the MSB.

23 years agonote OPENSSL_issetugid().
Ulf Möller [Mon, 19 Feb 2001 23:58:56 +0000 (23:58 +0000)]
note OPENSSL_issetugid().

23 years agoTemporary fix for build break.
Ulf Möller [Mon, 19 Feb 2001 23:42:09 +0000 (23:42 +0000)]
Temporary fix for build break.
It's still inconsistent - probably better to undo the whole OPENSSL_NO_* thing.

23 years agoMake sure time() is properly declared.
Richard Levitte [Mon, 19 Feb 2001 22:47:40 +0000 (22:47 +0000)]
Make sure time() is properly declared.

23 years agoUse the new-style system-identity macros.
Richard Levitte [Mon, 19 Feb 2001 22:04:02 +0000 (22:04 +0000)]
Use the new-style system-identity macros.

23 years agoI forgot to document the system identification macros
Richard Levitte [Mon, 19 Feb 2001 16:15:13 +0000 (16:15 +0000)]
I forgot to document the system identification macros

23 years agoMake all configuration macros available for application by making
Richard Levitte [Mon, 19 Feb 2001 16:06:34 +0000 (16:06 +0000)]
Make all configuration macros available for application by making
sure they are available in opensslconf.h, by giving them names starting
with "OPENSSL_" to avoid conflicts with other packages and by making
sure e_os2.h will cover all platform-specific cases together with
opensslconf.h.

I've checked fairly well that nothing breaks with this (apart from
external software that will adapt if they have used something like
NO_KRB5), but I can't guarantee it completely, so a review of this
change would be a good thing.

23 years agoRemove temporary files when done.
Richard Levitte [Mon, 19 Feb 2001 15:23:37 +0000 (15:23 +0000)]
Remove temporary files when done.

23 years agomake update
Richard Levitte [Mon, 19 Feb 2001 14:00:38 +0000 (14:00 +0000)]
make update

23 years agoAn obvious but hard-to-see cut'n'paste error corrected.
Richard Levitte [Mon, 19 Feb 2001 13:57:40 +0000 (13:57 +0000)]
An obvious but hard-to-see cut'n'paste error corrected.

23 years agoNew -set_serial options to 'req' and 'x509'.
Dr. Stephen Henson [Mon, 19 Feb 2001 13:38:32 +0000 (13:38 +0000)]
New -set_serial options to 'req' and 'x509'.

Remove the old broken bio read of serial numbers in the 'ca' index
file. This would choke if a revoked certificate was specified with
a negative serial number.

Fix typo in uid.c

23 years agoDo not insert things in syms{} and kind{} when parsing the header
Richard Levitte [Mon, 19 Feb 2001 13:33:04 +0000 (13:33 +0000)]
Do not insert things in syms{} and kind{} when parsing the header
files.  Instead, insert proper information in the $def string, which
will be properly munged later on.

23 years agoMake the choice of "makedepend" program choosable through a switch.
Richard Levitte [Mon, 19 Feb 2001 12:51:14 +0000 (12:51 +0000)]
Make the choice of "makedepend" program choosable through a switch.

23 years agoVMS follows suit.
Richard Levitte [Mon, 19 Feb 2001 11:30:22 +0000 (11:30 +0000)]
VMS follows suit.

23 years agoMemory leak detection bugfixes for multi-threading.
Bodo Möller [Mon, 19 Feb 2001 10:32:53 +0000 (10:32 +0000)]
Memory leak detection bugfixes for multi-threading.

23 years agoAdd uid.{c,o}
Bodo Möller [Mon, 19 Feb 2001 10:31:04 +0000 (10:31 +0000)]
Add uid.{c,o}

23 years agoPerl code patch contributed by "Kurt J. Pires" <kjpires@iat.com>
Richard Levitte [Mon, 19 Feb 2001 09:29:53 +0000 (09:29 +0000)]
Perl code patch contributed by "Kurt J. Pires" <kjpires@iat.com>
His own words are:

The patch adds no new functionality (other than a simple test package)
to the libraries, but it allows them to be compiled with Perl5.6.0.
It has only been tested under "Red Hat Linux release 7.0 (Guinness)"
with the unpatched verion of OpenSSL 0.9.6 released last September.

23 years agoMake it possible to use gcc to generate the dependency tables.
Richard Levitte [Fri, 16 Feb 2001 13:55:05 +0000 (13:55 +0000)]
Make it possible to use gcc to generate the dependency tables.

23 years agoispell
Ulf Möller [Fri, 16 Feb 2001 02:09:53 +0000 (02:09 +0000)]
ispell

23 years agopod format error
Ulf Möller [Fri, 16 Feb 2001 01:44:24 +0000 (01:44 +0000)]
pod format error

23 years agoNew options to 'ca' utility to support CRL entry extensions.
Dr. Stephen Henson [Fri, 16 Feb 2001 01:35:44 +0000 (01:35 +0000)]
New options to 'ca' utility to support CRL entry extensions.

Add revelant new X509V3 extensions.

Add OIDs.

Fix ASN1 memory leak code to pop info if external allocation used.

23 years agoMove entry to match chronologic orderering.
Lutz Jänicke [Thu, 15 Feb 2001 14:18:53 +0000 (14:18 +0000)]
Move entry to match chronologic orderering.

23 years agoDon't forget to mention minor change.
Lutz Jänicke [Thu, 15 Feb 2001 10:35:56 +0000 (10:35 +0000)]
Don't forget to mention minor change.

23 years agoAdd "-rand" option to s_client and s_server.
Lutz Jänicke [Thu, 15 Feb 2001 10:22:07 +0000 (10:22 +0000)]
Add "-rand" option to s_client and s_server.

23 years agoThat was misleading. The problem won't happen with 0.9.6a anyway.
Ulf Möller [Wed, 14 Feb 2001 16:55:22 +0000 (16:55 +0000)]
That was misleading. The problem won't happen with 0.9.6a anyway.

23 years agoInclude the newly reported problem with bc on FreeBSD 4.2.
Richard Levitte [Wed, 14 Feb 2001 07:45:59 +0000 (07:45 +0000)]
Include the newly reported problem with bc on FreeBSD 4.2.

23 years agoOption to disable standard block padding with EVP API.
Dr. Stephen Henson [Wed, 14 Feb 2001 02:11:52 +0000 (02:11 +0000)]
Option to disable standard block padding with EVP API.

Add -nopad option to enc command.

Update docs.

23 years agoNew function OPENSSL_issetugid(). Needs more work.
Ulf Möller [Wed, 14 Feb 2001 01:35:44 +0000 (01:35 +0000)]
New function OPENSSL_issetugid(). Needs more work.

23 years agoInitial OCSP SSL support.
Dr. Stephen Henson [Wed, 14 Feb 2001 01:12:41 +0000 (01:12 +0000)]
Initial OCSP SSL support.

23 years agoIRIX bugfix
Ulf Möller [Wed, 14 Feb 2001 00:14:09 +0000 (00:14 +0000)]
IRIX bugfix

23 years agoFinish first round of session cache documentation.
Lutz Jänicke [Tue, 13 Feb 2001 14:00:09 +0000 (14:00 +0000)]
Finish first round of session cache documentation.

23 years agoUpdate of linux-ppc. Contributed by MATSUURA Takanori
Richard Levitte [Tue, 13 Feb 2001 13:08:12 +0000 (13:08 +0000)]
Update of linux-ppc.  Contributed by MATSUURA Takanori
<t-matsuu@protein.osaka-u.ac.jp>

23 years agoNew manual page: SSL_CTX_set_mode.
Lutz Jänicke [Tue, 13 Feb 2001 11:43:11 +0000 (11:43 +0000)]
New manual page: SSL_CTX_set_mode.

23 years agoNew function OCSP_parse_url() and -url option for ocsp utility.
Dr. Stephen Henson [Tue, 13 Feb 2001 00:37:44 +0000 (00:37 +0000)]
New function OCSP_parse_url() and -url option for ocsp utility.

Doesn't handle SSL URLs yet.

23 years agoModify OCSP nonce behaviour.
Dr. Stephen Henson [Mon, 12 Feb 2001 23:28:45 +0000 (23:28 +0000)]
Modify OCSP nonce behaviour.

23 years agoWork around for libsafe "error".
Dr. Stephen Henson [Mon, 12 Feb 2001 03:22:49 +0000 (03:22 +0000)]
Work around for libsafe "error".

23 years agoRe-order a couple of static functions and "#if 0" out unused ones - this
Geoff Thorpe [Mon, 12 Feb 2001 02:30:19 +0000 (02:30 +0000)]
Re-order a couple of static functions and "#if 0" out unused ones - this
gets rid of gcc warnings.

23 years agoThis change was a quick experiment that I'd wanted to try that works quite
Geoff Thorpe [Mon, 12 Feb 2001 02:28:29 +0000 (02:28 +0000)]
This change was a quick experiment that I'd wanted to try that works quite
well (and is a good demonstration of how encapsulating the SSL in a
memory-based state machine can make it easier to apply to different
situations).

The change implements a new command-line switch "-flipped <0|1>" which, if
set to 1, reverses the usual interpretation of a client and server for SSL
tunneling. Normally, an ssl client (ie. "-server 0") accepts "cleartext"
connections and conducts SSL/TLS over a proxied connection acting as an SSL
client. Likewise, an ssl server (ie. "-server 1") accepts connections and
conducts SSL/TLS (as an SSL server) over them and passes "cleartext" over
the proxied connection. With "-flipped 1", an SSL client (specified with
"-server 0") in fact accepts SSL connections and proxies clear, whereas an
SSL server ("-server 1") accepts clear and proxies SSL. NB: most of this
diff is command-line handling, the actual meat of the change is simply the
line or two that plugs "clean" and "dirty" file descriptors into the item
that holds the state-machine - reverse them and you get the desired
behaviour.

This allows a network server to be an SSL client, and a network client to
be an SSL server. Apart from curiosity value, there's a couple of possibly
interesting applications - SSL/TLS is inherently vulnerable to trivial DoS
attacks, because the SSL server usually has to perform a private key
operation first, even if the client is authenticated. With this scenario,
the network client is the SSL server and performs the first private key
operation, whereas the network server serves as the SSL client. Another
possible application is when client-only authentication is required (ie.
the underlying protocol handles (or doesn't care about) authenticating the
server). Eg. an SSL/TLS version of 'ssh' could be concocted where the
client's signed certificate is used to validate login to a server system -
whether or not the client needs to validate who the server is can be
configured at the client end rather than at the server end (ie. a complete
inversion of what happens in normal SSL/TLS).

NB: This is just an experiment/play-thing, using "-flipped 1" probably
creates something that is interoperable with exactly nothing. :-)

23 years agoMore about session caching.
Lutz Jänicke [Sun, 11 Feb 2001 17:01:36 +0000 (17:01 +0000)]
More about session caching.

23 years agoInclude information that automatic query is a new feature.
Lutz Jänicke [Sat, 10 Feb 2001 19:10:36 +0000 (19:10 +0000)]
Include information that automatic query is a new feature.

23 years agoUpdate for 0.9.7 with SSL_OP_CIPHER_SERVER_PREFERENCE.
Lutz Jänicke [Sat, 10 Feb 2001 16:21:38 +0000 (16:21 +0000)]
Update for 0.9.7 with SSL_OP_CIPHER_SERVER_PREFERENCE.

23 years agoManual page for SSL_CTX_set_options(). Unfortunately for some of the
Lutz Jänicke [Sat, 10 Feb 2001 16:18:35 +0000 (16:18 +0000)]
Manual page for SSL_CTX_set_options(). Unfortunately for some of the
options someone much longer working with OpenSSL/SSLeay is needed.

23 years agoOops: It's RegTP, not RegPT ...
Bodo Möller [Sat, 10 Feb 2001 13:35:34 +0000 (13:35 +0000)]
Oops: It's RegTP, not RegPT ...

23 years agoAdd German SiG root certificates (extracted from the official cert registry
Bodo Möller [Sat, 10 Feb 2001 13:16:16 +0000 (13:16 +0000)]
Add German SiG root certificates (extracted from the official cert registry
file http://www.nrca-ds.de/ftp/pkd.ttp, which contains a total of 288
certificates issued by the RegPT so far)

23 years agodisable stdin buffering in load_cert
Bodo Möller [Sat, 10 Feb 2001 13:12:35 +0000 (13:12 +0000)]
disable stdin buffering in load_cert

23 years agouse case-insensitive comparison in set_table_opts
Bodo Möller [Sat, 10 Feb 2001 11:21:29 +0000 (11:21 +0000)]
use case-insensitive comparison in set_table_opts
(similar to how arguments such as -inform/-outform specifications
are treated)

23 years agoFix CRL printing to correctly show when there are no revoked certificates.
Dr. Stephen Henson [Sat, 10 Feb 2001 00:56:45 +0000 (00:56 +0000)]
Fix CRL printing to correctly show when there are no revoked certificates.

Make ca.c correctly initialize the revocation date.

Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string() set the
string type: so they can initialize ASN1_TIME structures properly.

23 years agoSimplify BN_rand_range
Bodo Möller [Sat, 10 Feb 2001 00:34:02 +0000 (00:34 +0000)]
Simplify BN_rand_range

23 years agoNew Option SSL_OP_CIPHER_SERVER_PREFERENCE allows TLS/SSLv3 server to override
Lutz Jänicke [Fri, 9 Feb 2001 19:56:31 +0000 (19:56 +0000)]
New Option SSL_OP_CIPHER_SERVER_PREFERENCE allows TLS/SSLv3 server to override
the clients choice; in SSLv2 the client uses the server's preferences.

23 years agoTypo
Lutz Jänicke [Fri, 9 Feb 2001 19:05:49 +0000 (19:05 +0000)]
Typo

23 years agoFix "wierd" typo as submitted by Jeroen Ruigrok/Asmodai <asmodai@wxs.nl>.
Lutz Jänicke [Fri, 9 Feb 2001 19:03:53 +0000 (19:03 +0000)]
Fix "wierd" typo as submitted by Jeroen Ruigrok/Asmodai <asmodai@wxs.nl>.

23 years agoVarious Win32 related fixed. Make no-krb5 work in mkdef.pl .
Dr. Stephen Henson [Fri, 9 Feb 2001 18:16:12 +0000 (18:16 +0000)]
Various Win32 related fixed. Make no-krb5 work in mkdef.pl .

Fix warning in apps/engine.c

Remove definitions of deleted functions.

Add missing definition of X509_VAL.

23 years agoVarious updates to mkdef.pl to cope with new aes
Dr. Stephen Henson [Fri, 9 Feb 2001 13:16:21 +0000 (13:16 +0000)]
Various updates to mkdef.pl to cope with new aes
and ASN1 code.

23 years agofix editing error
Bodo Möller [Fri, 9 Feb 2001 09:40:18 +0000 (09:40 +0000)]
fix editing error

23 years agoadd linux-s390 configuration (based on information submitted by
Bodo Möller [Fri, 9 Feb 2001 08:34:29 +0000 (08:34 +0000)]
add linux-s390 configuration (based on information submitted by
Denis Beauchemin <Denis.Beauchemin@Courrier.USherb.ca>)

23 years agoAdd missing \n's to ocsp usage message.
Dr. Stephen Henson [Fri, 9 Feb 2001 03:09:05 +0000 (03:09 +0000)]
Add missing \n's to ocsp usage message.

23 years agoAllow various options to be included for signing and verify of
Dr. Stephen Henson [Thu, 8 Feb 2001 19:36:10 +0000 (19:36 +0000)]
Allow various options to be included for signing and verify of
OCSP responses.

Documentation to follow...

Urgh.. this conflicted with the -VAfile patch I hope I haven't
broken it.

23 years agoThe check for request including a nonce and response not having it was
Richard Levitte [Thu, 8 Feb 2001 19:28:10 +0000 (19:28 +0000)]
The check for request including a nonce and response not having it was
inversed.  Corrected.  Hopefully, this will make it work without
dumping core.

23 years agoAdd the -VAfile option to 'openssl ocsp'. This option will give the
Richard Levitte [Thu, 8 Feb 2001 17:59:29 +0000 (17:59 +0000)]
Add the -VAfile option to 'openssl ocsp'.  This option will give the
client code certificates to use to only check response signatures.
I'm not entirely sure if the way I just implemented the verification
is the right way to do it, and would be happy if someone would like to
review this.

23 years agouse <= instead of ==
Ulf Möller [Thu, 8 Feb 2001 17:45:32 +0000 (17:45 +0000)]
use <= instead of ==

23 years agopoint out that RAND_load_file() etc are only for seed files, not for
Ulf Möller [Thu, 8 Feb 2001 17:22:56 +0000 (17:22 +0000)]
point out that RAND_load_file() etc are only for seed files, not for
entropy devices or sockets.

23 years agoNote that EGD is used automatically.
Ulf Möller [Thu, 8 Feb 2001 17:16:44 +0000 (17:16 +0000)]
Note that EGD is used automatically.

23 years agocleanup
Ulf Möller [Thu, 8 Feb 2001 17:14:07 +0000 (17:14 +0000)]
cleanup

23 years agoAnother comment change. (Previous comment does not apply
Bodo Möller [Thu, 8 Feb 2001 12:34:08 +0000 (12:34 +0000)]
Another comment change.  (Previous comment does not apply
for range = 11000000... or range = 100000...)

23 years agoChange comments. (The expected number of iterations in BN_rand_range
Bodo Möller [Thu, 8 Feb 2001 12:27:22 +0000 (12:27 +0000)]
Change comments.  (The expected number of iterations in BN_rand_range
never exceeds 1.333...).

23 years agooops -- remove observation code
Bodo Möller [Thu, 8 Feb 2001 12:24:41 +0000 (12:24 +0000)]
oops -- remove observation code

23 years agoIntegrate my implementation of a countermeasure against
Bodo Möller [Thu, 8 Feb 2001 12:14:51 +0000 (12:14 +0000)]
Integrate my implementation of a countermeasure against
Bleichenbacher's DSA attack.  With this implementation, the expected
number of iterations never exceeds 2.

New semantics for BN_rand_range():
BN_rand_range(r, min, range) now generates r such that
     min <= r < min+range.
(Previously, BN_rand_range(r, min, max) generated r such that
     min <= r < max.
It is more convenient to have the range; also the previous
prototype was misleading because max was larger than
the actual maximum.)

23 years agoplatform specific CFLAGS don't belong into this Makefile
Bodo Möller [Thu, 8 Feb 2001 11:15:50 +0000 (11:15 +0000)]
platform specific CFLAGS don't belong into this Makefile

23 years agoUpdate documentation to match the state at 0.9.6 _and_ the recent changes.
Lutz Jänicke [Thu, 8 Feb 2001 10:42:01 +0000 (10:42 +0000)]
Update documentation to match the state at 0.9.6 _and_ the recent changes.

23 years agoBleichenbacher's DSA attack
Ulf Möller [Wed, 7 Feb 2001 22:24:35 +0000 (22:24 +0000)]
Bleichenbacher's DSA attack

23 years agoModify access to EGD socket to deal with EINTR etc that can appear
Lutz Jänicke [Wed, 7 Feb 2001 22:13:38 +0000 (22:13 +0000)]
Modify access to EGD socket to deal with EINTR etc that can appear
during connect() and other calls. First seen on Unixware-7.

Unify access to EGD-socket for all RAND_egd_*() methods.