Rich Salz [Thu, 28 Jan 2016 03:00:55 +0000 (22:00 -0500)]
Remove outdated tests
These tests are not built, and only usable as hand-tests so not
worth moving into our test framework.
This closes https://github.com/openssl/openssl/pull/561 and RT 4252
Reviewed-by: Richard Levitte <levitte@openssl.org>
Matt Caswell [Mon, 18 Jan 2016 11:31:58 +0000 (11:31 +0000)]
Prevent small subgroup attacks on DH/DHE
Historically OpenSSL only ever generated DH parameters based on "safe"
primes. More recently (in version 1.0.2) support was provided for
generating X9.42 style parameter files such as those required for RFC
5114 support. The primes used in such files may not be "safe". Where an
application is using DH configured with parameters based on primes that
are not "safe" then an attacker could use this fact to find a peer's
private DH exponent. This attack requires that the attacker complete
multiple handshakes in which the peer uses the same DH exponent.
A simple mitigation is to ensure that y^q (mod p) == 1
Rich Salz [Wed, 27 Jan 2016 23:43:25 +0000 (18:43 -0500)]
Remove outdated legacy crypto options
Many options for supporting optimizations for legacy crypto on legacy
platforms have been removed. This simplifies the source code and
does not really penalize anyone.
DES_PTR (always on)
DES_RISC1, DES_RISC2 (always off)
DES_INT (always 'unsigned int')
DES_UNROLL (always on)
BF_PTR (always on) BF_PTR2 (removed)
MD2_CHAR, MD2_LONG (always 'unsigned char')
IDEA_SHORT, IDEA_LONG (always 'unsigned int')
RC2_SHORT, RC2_LONG (always 'unsigned int')
RC4_LONG (only int and char (for assembler) are supported)
RC4_CHUNK (always long), RC_CHUNK_LL (removed)
RC4_INDEX (always on)
And also make D_ENCRYPT macro more clear (@appro)
Rich Salz [Fri, 18 Dec 2015 17:57:23 +0000 (12:57 -0500)]
Remove /* foo.c */ comments
This was done by the following
find . -name '*.[ch]' | /tmp/pl
where /tmp/pl is the following three-line script:
print unless $. == 1 && m@/\* .*\.[ch] \*/@;
close ARGV if eof; # Close file to reset $.
Benjamin Kaduk [Thu, 21 Jan 2016 18:06:39 +0000 (12:06 -0600)]
Remove unused, undocumented clean-shared target
Also removes the make variable SHARED_LIBS_LINK_EXTS, only used by
the clean-shared target.
When shared library linking was moved to the separate Makefile.shared
in commit 30afcc072acd4f70590fec68bf0590da4e4f1883, this target was
skipped. Prior to that commit, clean-shared was invoked as a
dependency of build-shared, but afterward it was no longer referenced
anywhere in the tree.
Instead of porting the functionality over to Makefile.shared, just
remove it entirely, as it appears to be unused.
Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org>
Richard Levitte [Tue, 26 Jan 2016 13:57:21 +0000 (14:57 +0100)]
Have OpenSSL::Test handle perl scripts like any program
Since we're building some of our perl scripts and the result might not
end up in apps/ (*), we may need to treat them like the compile
programs we use for testing.
This introduces perlapp() and perltest(), which behave like app() and
test(), but will add the perl executable in the command line.
-----
(*) For example, with a mk1mf build, the result will end up in $(BIN_D)
Matt Caswell [Tue, 26 Jan 2016 11:31:41 +0000 (11:31 +0000)]
Add SSL_up_ref() and SSL_CTX_up_ref()
The SSL and SSL_CTX structures are reference counted. However since libssl
was made opaque there is no way for users of the library to manipulate the
reference counts. This adds functions to enable that.
Richard Levitte [Tue, 26 Jan 2016 01:09:33 +0000 (02:09 +0100)]
Use Configure's @disablables and %disabled through configdata.pm
Enhances the routines in OpenSSL::Test::Utils for checking disabled
stuff to get their information directly from Configure instead of
'openssl list -disabled'.
Richard Levitte [Mon, 25 Jan 2016 20:51:22 +0000 (21:51 +0100)]
Small cleanups in Configure
- Small rearrangement of the TABLE and HASH printouts, and adding
printout of the "build_scheme" item
- Renamed "engines_obj" to "padlock_obj"
- Moved the runs of dofile down... it didn't quite make sense to have
that in the middle of a printout
Richard Levitte [Mon, 25 Jan 2016 20:19:59 +0000 (21:19 +0100)]
Generate warning text
Now that we're using templates, we should warn people not to edit the
resulting file. We do it through util/dofile.pl, which is enhanced
with an option to tell what file it was called from. We also change
the calls so the template files are on the command line instead of
being redirected through standard input. That way, we can display
something like this (example taken from include/openssl/opensslconf.h):
/* WARNING: do not edit! */
/* Generated by Configure from include/openssl/opensslconf.h.in */
Richard Levitte [Wed, 20 May 2015 18:03:20 +0000 (20:03 +0200)]
Refactor file writing - Remake Makefile.org into a template
It is time for Makefile.org to fold into the new regime and have a run
through util/dofile.pl. This forces some information out of there and
into Configure, which isn't a bad thing, it makes Configure
increasingly the center of build information, which is as it should
be.
A few extra defaults were needed in the BASE template to get rid of
warnings about missing values.
Richard Levitte [Sun, 27 Dec 2015 01:27:30 +0000 (02:27 +0100)]
Refactor file writing - make configdata.pm the info center for "reconf"
Now that configdata.pm is the centre of information, use that instead
of Makefile to figure out reconfiguration parameters. This will help
future development with different Makefile file names.
The code to read necessary configuration data from Makefile is retained
for an easy transition to configdata.pm based information gathering. It
will be removed later on.
This change includes moving the variable $cross_compile_prefix to %config.
Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Viktor Dukhovni <viktor@openssl.org>
Richard Levitte [Mon, 18 May 2015 23:40:54 +0000 (01:40 +0200)]
Refactor file writing - rewrite crypto/opensslconf.h.in as template
The turn has come to have crypto/opensslconf.h.in get run through
util/dofile.pl. The consequence is that a large number of variables
get moved to the %config table.
Also, the string variables $openssl_*, which were populated with cpp
lines, all being of the form "#define SOMETHING", were converted into
ARRAY refs in %config values, containing just the list of macros to be
defined.
Richard Levitte [Mon, 11 Jan 2016 23:17:12 +0000 (00:17 +0100)]
Refactor file writing - Adapt util/mkdef.pl to use configdata.pm
For this adaptation, the variables $options and $version needed to
move to %config in Configure, and why not move all other variables
holding diverse version numbers at the same time?
Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org>
Richard Levitte [Fri, 22 Jan 2016 11:37:31 +0000 (12:37 +0100)]
Refactor file writing - arrange for use of bundled Perl modules as fallback
For our own convenience, we need a mechanism to be able to fall back
on bundled Perl modules. It's a minimal package that's called like
this:
use with_fallback qw(Module1 Module2 ...);
For each module, it will try to require them from the system
installation, and failing that, it will temporarly add external/perl
and try to require transfer::{ModuleName}. It requires that each
bundled Perl modules is accompanied by a small transfer module
(external/perl/transfer/ModuleName.pm in our example) that knows
exactly what to load.
apps/CA.pl and tools/c_rehash are built from template files. So far,
this was done by Configure, which created its own problems as it
forced everyone to reconfigure just because one of the template files
had changed.
Instead, have those files created as part of the normal build in apps/
and in tools/.
Furthermore, this prepares for a future where Configure may produce
entirely other build files than Makefile, and the latter can't be
guaranteed to be the holder of all information for other scripts.
Instead, configdata.pm (described below) becomes the center of
configuration information.
This introduces a few new things:
%config a hash table to hold all kinds of configuration data
that can be used by any other script.
configdata.pm a perl module that Configure writes. It currently
holds the hash tables %config and %target.
util/dofile.pl a script that takes a template on STDIN and outputs
the result after applying configuration data on it.
It's supposed to be called like this:
perl -I$(TOP) -Mconfigdata < template > result
or
perl -I$(TOP) -Mconfigdata templ1 templ2 ... > result
Note: util/dofile.pl requires Text::Template.
As part of this changed, remove a number of variables that are really
just copies of entries in %target, and use %target directly. The
exceptions are $target{cflags} and $target{lflags}, they do get copied
to $cflags and $lflags. The reason for this is that those variable
potentially go through a lot of changes and would rather deserve a
place in %config. That, however, is for another commit.
Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org>
Richard Levitte [Mon, 18 May 2015 10:53:38 +0000 (12:53 +0200)]
Refactor config - consolidate and refresh print_table_entry
It's time for print_table_entry to get a bit of refreshment. The way it
was put together, we needed to maintain the list of known configuration
keys of interest twice, in different shapes. This is error prone, so
move the list of strings to a common list for all printing cases, and
use simple formatting of lines to do the actual printout based on that
list.
Richard Levitte [Fri, 22 Jan 2016 12:06:26 +0000 (13:06 +0100)]
Cleanup .gitignore
Some things to ignore need to be properly rooted, and use a bit more
precision on ignoring 'lib', as that maybe be a perfectly valid
directory name to add into git elsewhere in the source tree.
Richard Levitte [Mon, 18 May 2015 12:31:49 +0000 (14:31 +0200)]
Refactor config - @MK1MF_Builds out, general build scheme in
Time to get rid of @MK1MF_Builds and introduce a more flexible
'build_scheme' configuration key. Its value may be a string or an
array of strings, meaning we need to teach resolve_config how to
handle ARRAY referenses.
The build scheme is a word that selects a function to create the
appropriate result files for a certain configuration. Currently valid
build schemes aer "mk1mf" and "unixmake", the plan is however to add
at least one other for a more universal build scheme.
Incidently, this also adds the functions 'add' and 'add_before', which
can be used in a configuration, so instead of having to repeatedly
write a sub like this:
key1 => sub { join(" ", @_, "myvalues"); },
key2 => sub { join(" ", "myvalues", @_); },
The good point with 'add' and 'add_before' is that they handle
inheritances where the values are a misture of scalars and ARRAYs. If
there are any ARRAY to be found, the resulting value will be an ARRAY,
otherwise it will be a scalar with all the incoming valued joined
together with the separator given as first argument to add/add_before.
Richard Levitte [Mon, 18 May 2015 00:54:28 +0000 (02:54 +0200)]
Refactor config - move templates and template docs to Configurations
Move the documentation of the target configuration form to
Configurations/README.
Move initial assembler object templates to
Configurations/00-BASE-templates.conf.
Furthermore, remove all variables containing the names of the
non-assembler object files and make a BASE template of them instead.
The values from this templates are used as defaults as is. The
remaining manipulation of data when assembler modules are used is done
only when $no_asm is false.
While doing this, clean out some other related variables that aren't
used anywhere.
Also, we had to move the resolution of the chosen target a bit, or the
function 'asm' would never catch a true $no_asm... this hasn't
mattered before we've moved it all to the BASE template, but now it
does.
At the same time, add the default for the 'unistd' key to the BASE
template.
Richard Levitte [Wed, 20 Jan 2016 17:11:51 +0000 (18:11 +0100)]
Refresh the thinking of --prefix and --openssldir
--prefix is now exclusively used for software and manual installation.
--openssldir is not exclusively used as a default location for certs,
keys and the default openssl.cnf.
This change is made to bring clarity, to have the two less
intertwined, and to be more compatible with the usual ways of software
installation.
Please change your habits and scripts to use --prefix rather than
--openssldir for installation location now.
Richard Levitte [Mon, 18 May 2015 09:03:47 +0000 (11:03 +0200)]
Refactor config - consolidate handling of disabled stuff
It's time to refactor the handling of %disabled so that all
information of value is in the same place. We have so far had a few
cascading disable rules in form of code, far away from %disabled.
Instead, bring that information to the array @disable_cascade, which
is a list of pairs of the form 'test => descendents'. The test part
can be a string, and it's simply checked if that string is a key in
%disabled, or it can be a CODEref to do a more complex test. If the
test comes true, then all descendents are disabled. This check is
performed until there are no more things that need to be disabled.
Also, $default_depflags is constructed from the information in
%disabled instead of being a separate string. While a string of its
own is visually appealing, it's much too easy to forget to update it
when something is changed in %disabled.
Richard Levitte [Mon, 18 May 2015 01:33:55 +0000 (03:33 +0200)]
Refactor config - rewrite handling of "reconf"
The way the "reconf"/"reconfigure" argument is handled is overly
complicated. Just grep for it first, and if it is there in the
current arguments, get the old command line arguments from Makefile.
While we're at it, make the Makefile variable CONFIGURE_ARGS hold the
value as a perl list of strings. This makes things much safer in case
one of the arguments would contain a space. Since CONFIGURE_ARGS is
used for nothing else, there's no harm in this.
Viktor Dukhovni [Sat, 16 Jan 2016 06:15:02 +0000 (01:15 -0500)]
Multiple -trusted/-untrusted/-CRLfile options in verify
It is sometimes useful (especially in automated tests) to supply
multiple trusted or untrusted certificates via separate files rather
than have to prepare a single file containing them all.
To that end, change verify(1) to accept these options zero or more
times. Also automatically set -no-CAfile and -no-CApath when
-trusted is specified.
Improve verify(1) documentation, which could still use some work.
Reviewed-by: Richard Levitte <levitte@openssl.org>
Viktor Dukhovni [Fri, 15 Jan 2016 08:49:11 +0000 (03:49 -0500)]
Reject when explicit trust EKU are set and none match.
Returning untrusted is enough for for full chains that end in
self-signed roots, because when explicit trust is specified it
suppresses the default blanket trust of self-signed objects.
But for partial chains, this is not enough, because absent a similar
trust-self-signed policy, non matching EKUs are indistinguishable
from lack of EKU constraints.
Therefore, failure to match any trusted purpose must trigger an
explicit reject.
Reviewed-by: Richard Levitte <levitte@openssl.org>
Viktor Dukhovni [Sun, 17 Jan 2016 21:50:52 +0000 (16:50 -0500)]
Check Suite-B constraints with EE DANE records
When DANE-EE(3) matches or either of DANE-EE/PKIX-EE fails, we don't
build a chain at all, but rather succeed or fail with just the leaf
certificate. In either case also check for Suite-B violations.
As unlikely as it may seem that anyone would enable both DANE and
Suite-B, we should do what the application asks.
Took the opportunity to eliminate the "cb" variables in x509_vfy.c,
just call ctx->verify_cb(ok, ctx)
Reviewed-by: Dr. Stephen Henson <steve@openssl.org>
Richard Levitte [Mon, 18 May 2015 00:57:24 +0000 (02:57 +0200)]
Refactor config - split read_config into read_config and resolve_config
Split the read_config function into read_config that ONLY reads the
configuration files but doesn't try to resolve any of the
inheritances, and resolve_config which resolves the inheritance chain
of a given target. Move them to the bottom of Configure, with the
rest of the helpers.
Have a new small hash table, %target, which will hold the values for
the target the user requested. This also means that all access to the
current target data can be reduced from '$table{$target}->{key}' to a
mere '$target{key}'.
While we're at it, the old string formatted configurations are getting
obsolete, so they may as well get deprecated entirely.
Richard Levitte [Mon, 18 May 2015 00:46:21 +0000 (02:46 +0200)]
Refactor config - a small cosmetic touchup of Configure
Start simple, removed some unused variables and change all '<<EOF' to
'<<"EOF"'. The latter is because some code colorizers (notably, in
emacs) cannot recognise the here document end marker unless it's
quoted and therefore assume the rest of the file is part of the here
document.
Rich Salz [Sun, 17 Jan 2016 21:53:53 +0000 (16:53 -0500)]
Consolidate "make update"
Rename 'update' to 'generate'. Rather than recurse, just explicitly
call the three generate targets directly. Reviewed-by: Richard Levitte <levitte@openssl.org>
Matt Caswell [Mon, 7 Dec 2015 16:50:38 +0000 (16:50 +0000)]
Handle SSL_shutdown while in init more appropriately
Calling SSL_shutdown while in init previously gave a "1" response, meaning
everything was successfully closed down (even though it wasn't). Better is
to send our close_notify, but fail when trying to receive one.
The problem with doing a shutdown while in the middle of a handshake is
that once our close_notify is sent we shouldn't really do anything else
(including process handshake/CCS messages) until we've received a
close_notify back from the peer. However the peer might send a CCS before
acting on our close_notify - so we won't be able to read it because we're
not acting on CCS messages!