Matt Caswell [Thu, 21 Jan 2016 12:22:58 +0000 (12:22 +0000)]
Add tests for DTLSv1_listen
Adds a set of tests for the newly rewritten DTLSv1_listen function.
The test pokes various packets at the function and then checks
the return value and the data written out to ensure it is what we
would have expected.
Matt Caswell [Wed, 4 Nov 2015 13:53:57 +0000 (13:53 +0000)]
Provide partial support for fragmented DTLS ClientHellos
The recently rewriten DTLSv1_listen code does not support fragmented
ClientHello messages because fragment reassembly requires server state
which is against the whole point of DTLSv1_listen. This change adds some
partial support for fragmented ClientHellos. It requires that the cookie
must be within the initial fragment. That way any non-initial ClientHello
fragments can be dropped and fragment reassembly is not required.
Matt Caswell [Fri, 5 Feb 2016 10:59:42 +0000 (10:59 +0000)]
Make DTLSv1_listen a first class function and change its type
The DTLSv1_listen function exposed details of the underlying BIO
abstraction and did not properly allow for IPv6. This commit changes the
"peer" argument to be a BIO_ADDR and makes it a first class function
(rather than a ctrl) to ensure proper type checking.
Reviewed-by: Richard Levitte <levitte@openssl.org>
Viktor Dukhovni [Mon, 1 Feb 2016 02:14:51 +0000 (21:14 -0500)]
Long overdue cleanup of X509 policy tree verification
Replace all magic numbers with #defined constants except in boolean
functions that return 0 for failure and 1 for success. Avoid a
couple memory leaks in error recovery code paths. Code style
improvements.
Reviewed-by: Dr. Stephen Henson <steve@openssl.org>
Matt Caswell [Wed, 16 Dec 2015 11:17:17 +0000 (11:17 +0000)]
Don't export local symbols on Solaris
Following on from earlier commits to prevent local symbols from being
exported in the shared libraries on Linux, this makes the equivalent changes
for Solaris.
Reviewed-by: Richard Levitte <levitte@openssl.org>
Don't require an application to work out the appropriate buffer size for
ASN1_bn_print(), which is unsafe. Ignore the supplied buffer and allocate
it internally instead.
New functions EC_KEY_oct2priv and EC_KEY_priv2oct. These are private key
equivalents of EC_POINT_oct2point and EC_POINT_point2oct which convert
between the private key octet format and EC_KEY.
Richard Levitte [Thu, 4 Feb 2016 22:22:52 +0000 (23:22 +0100)]
Add checks for IPv4 and IPv6 in OpenSSL::Test::Utils and use them
This uilds on the same way of checking for availability as we do in
TLSProxy. We use all IP factories we know of, starting with those who
know both IPv6 and IPv4 and ending with the one that only knows IPv4
and cache their possible success as foundation for checking the
available of each IP domain.
80-test_ssl.t has bigger chances of working on platforms that do not
run both IP domains.
Todd Short [Thu, 4 Feb 2016 16:41:25 +0000 (11:41 -0500)]
Add option to disable async
Add no-async option to Configure that forces ASYNC_NULL.
Related to RT1979
An embedded system or replacement C library (e.g. musl or uClibc)
may not support the *context APIs that are needed for async operation.
Compiles with musl. Ran unit tests, async tests skipped as expected.
Signed-off-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org>
Viktor Dukhovni [Wed, 3 Feb 2016 23:32:25 +0000 (18:32 -0500)]
Handle localhost being either 127.0.0.1 or ::1
When connecting to "localhost" the Proxy's choice of client address
family may not match the server's choice address family. Without
MultiHomed => 1, the proxy may try the wrong address family first,
and give up without trying the other.
Reviewed-by: Richard Levitte <levitte@openssl.org>
Richard Levitte [Tue, 2 Feb 2016 23:57:30 +0000 (00:57 +0100)]
Refactoring BIO: add a simple networking test of s_client and s_server
This makes use of TLSProxy, which was expanded to use IO::Socket::IP
(which is a core perl module) or IO::Socket::INET6 (which is said to
be more popular) instead IO::Socket::INET if one of them is installed.
Richard Levitte [Tue, 2 Feb 2016 23:47:42 +0000 (00:47 +0100)]
Refactoring BIO: Adapt s_client and s_server
s_socket.c gets brutally cleaned out and now consists of only two
functions, one for client and the other for server. They both handle
AF_INET, AF_INET6 and additionally AF_UNIX where supported. The rest
is just easy adaptation.
Both s_client and s_server get the new flags -4 and -6 to force the
use of IPv4 or IPv6 only.
Also, the default host "localhost" in s_client is removed. It's not
certain that this host is set up for both IPv4 and IPv6. For example,
Debian has "ip6-localhost" as the default hostname for [::1]. The
better way is to default |host| to NULL and rely on BIO_lookup() to
return a BIO_ADDRINFO with the appropriate loopback address for IPv4
or IPv6 as indicated by the |family| parameter.
Richard Levitte [Tue, 2 Feb 2016 22:50:52 +0000 (23:50 +0100)]
Refactoring BIO: add a test, using test/ssltest
This adds a couple of simple tests to see that SSL traffic using the
reimplemented BIO_s_accept() and BIO_s_connect() works as expected,
both on IPv4 and on IPv6.
Richard Levitte [Tue, 2 Feb 2016 15:23:26 +0000 (16:23 +0100)]
Refactoring BIO: Small adjustments
Because of the way bio_lcl.h is organised, we must not include
internal/cryptlib.h before it. As a matter of fact, bio_lcl.h
includes internal/cryptlib.h on its own.
Richard Levitte [Tue, 2 Feb 2016 14:54:57 +0000 (15:54 +0100)]
Refactoring BIO: add wrappers around sockaddr et al
Because different platforms have different levels of support for IPv6,
different kinds of sockaddr variants, and some have getaddrinfo et al
while others don't, we could end up with a mess if ifdefs, duplicate
code and other maintainance nightmares.
Instead, we're introducing wrappers around the common form for socket
communication:
BIO_ADDR, closely related to struct sockaddr and some of its variants.
BIO_ADDRINFO, closely related to struct addrinfo.
With that comes support routines, both convenient creators and
accessors, plus a few utility functions:
BIO_parse_hostserv, takes a string of the form host:service and
splits it into host and service. It checks for * in both parts, and
converts any [ipv6-address] syntax to ust the IPv6 address.
BIO_lookup, looks up information on a host.
All routines handle IPv4 (AF_INET) and IPv6 (AF_INET6) addresses, and
there is support for local sockets (AF_UNIX) as well.
Emilia Kasper [Tue, 2 Feb 2016 15:26:38 +0000 (16:26 +0100)]
RT3234: disable compression
CRIME protection: disable compression by default, even if OpenSSL is
compiled with zlib enabled. Applications can still enable compression by
calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by using
the SSL_CONF library to configure compression. SSL_CONF continues to
work as before:
Richard Levitte [Fri, 29 Jan 2016 22:33:10 +0000 (23:33 +0100)]
unified build scheme: add build.info files
Now that we have the foundation for the "unified" build scheme in
place, we add build.info files. They have been generated from the
Makefiles in the same directories. Things that are platform specific
will appear in later commits.
Richard Levitte [Fri, 29 Jan 2016 18:45:51 +0000 (19:45 +0100)]
unified build scheme: a first introduction
The "unified" build scheme revolves around small information files,
build.info, which each describe their own bit of everything that needs
to be built, using a mini-language described in Configurations/README.
The information in build.info file contain references to source files
and final result. Object files are not mentioned at all, they are
simply from source files. Because of this, all the *_obj items in
Configurations/*.conf are renamed to *_asm_src and the files listed
in the values are change from object files to their corresponding
source files. For the sake of the other build schemes, Configure
generates corresponding *_obj entries in %target.
Furthermore, the "unified" build scheme supports having a build
directory tree separate from the source directry tree.
All paths in a build.info file is assumed to be relative to its
location, either within the source tree or within the build tree.