MIPS64 assembly pack: add Poly1305 module.

Reviewed-by: Richard Levitte <levitte@openssl.org>

Configure: replace which() with IPC::Cmd::can_run.

Reviewed-by: Richard Levitte <levitte@openssl.org>

windows-makefile.tmpl: minor adjustments.

- some Perl versions are allergic to missing ';';
- don't stop if del fails;
- omit unused environment variable;

Reviewed-by: Stephen Henson <steve@openssl.org>

util/mkdef.pl: omit ordinals from Windows DLLs.

Reviewed-by: Stephen Henson <steve@openssl.org>

Typo.

RT#4538

Reviewed-by: Matt Caswell <matt@openssl.org>

Fix the docs for ERR_remove_thread_state and ERR_remove_state

Don't primarly recommend using OPENSSL_thread_stop(), as that's a last
resort.  Instead, recommend leaving it to automatic mechanisms.

Reviewed-by: Matt Caswell <matt@openssl.org>

Restore the ERR_remove_thread_state() API and make it a no-op

The ERR_remove_thread_state() API is restored to take a pointer
argument, but does nothing more.  ERR_remove_state() is also made into
a no-op.  Both functions are deprecated and users are recommended to
use OPENSSL_thread_stop() instead.

Documentation is changed to reflect this.

Reviewed-by: Matt Caswell <matt@openssl.org>

Have [.VMS]openssl_{startup,shutdown}.com depend on respective *.in

Reviewed-by: Andy Polyakov <appro@openssl.org>

Fix VMS/openssl_{startup,shutddown}.com.in

They were using the wrong variables.

Reviewed-by: Andy Polyakov <appro@openssl.org>

Configure: adhere to $(CROSS_COMPILE)ranlib.

Reviewed-by: Richard Levitte <levitte@openssl.org>

Configure: make it work with Perl 5.10.

Reviewed-by: Richard Levitte <levitte@openssl.org>

IRIX fixes.

Reviewed-by: Richard Levitte <levitte@openssl.org>

Add NULL check in i2d_PrivateKey()

Originally submitted by Kurt Cancemi <kurt@x64architecture.com>

Closes RT#4533

Reviewed-by: Matt Caswell <matt@openssl.org>

Don't send signature algorithms when client_version is below TLS 1.2.

Per RFC 5246,

    Note: this extension is not meaningful for TLS versions prior to 1.2.
    Clients MUST NOT offer it if they are offering prior versions.
    However, even if clients do offer it, the rules specified in [TLSEXT]
    require servers to ignore extensions they do not understand.

Although second sentence would suggest that there would be no interop
problems in always offering the extension, WebRTC has reported issues
with Bouncy Castle on < TLS 1.2 ClientHellos that still include
signature_algorithms. See also
https://bugs.chromium.org/p/webrtc/issues/detail?id=4223

RT#4390

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Stephen Henson <steve@openssl.org>

Fix BIO_eof() for BIO pairs

BIO_eof() was always returning true when using a BIO pair. It should only
be true if the peer BIO is empty and has been shutdown.

RT#1215

Reviewed-by: Richard Levitte <levitte@openssl.org>

Fix: failed to open config file if not specified when using CA commands

Issue was introduced in
https://github.com/openssl/openssl/commit/a0a82324f965bbcc4faed4e1ee3fcaf81ea52166

This patch fixes an issue which causes the 'openssl ca' commands to
fail if '-config' is not specified even if it says so otherwise.
Problem is that the default config is not loaded and the conf variable
is NULL which causes an exception.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>

VMS: support VERBOSE and V in descrip.mms

With Unixly Makefiles as well as with nmake, make variables are
transferred to the shell running the commands as envinronment
variables.  This principle doesn't apply with MMS, so we must
explicitely define VERBOSE as commands when it's needed.

Reviewed-by: Rich Salz <rsalz@openssl.org>

Recognise VERBOSE and V as well as HARNESS_VERBOSE

PR#4462

Reviewed-by: Rich Salz <rsalz@openssl.org>

Missing credit in CHANGES

Reviewed-by: Richard Levitte <levitte@openssl.org>

Fix error in the loop of ECDH

The tests was incorrectly repeated multiple times when using the
async_jobs options

Reviewed-by: Kurt Roeckx <kurt@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>

Free any existing SRTP connection profile

When setting a new SRTP connection profile using
SSL_CTX_set_tlsext_use_srtp() or SSL_set_tlsext_use_srtp() we should
free any existing profile first to avoid a memory leak.

Reviewed-by: Kurt Roeckx <kurt@openssl.org>

Configurations/windows-makefile.tmpl: expand environments early.

If environment variables are not explanded early enough, expanded
strings are passed with single backslash to C compiler, e.g.
C:\Program Files, which effectively results in OpenSSL looking for
engines and certificates in C:Program Files.

Reviewed-by: Richard Levitte <levitte@openssl.org>

fix tab-space mixed indentation

No code change

Reviewed-by: Kurt Roeckx <kurt@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>

fix check

Reviewed-by: Kurt Roeckx <kurt@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>

few missing allocation failure checks and releases on error paths

- Missing checks for allocation failure.
- releasing memory in few missing error paths

Reviewed-by: Kurt Roeckx <kurt@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>

memset() doesn't take NULL.

Reviewed-by: Emilia Käsper <emilia@openssl.org>

Add fuzzing!

Reviewed-by: Emilia Käsper <emilia@openssl.org>

Constify PKCS12_newpass()

PR#4449

Reviewed-by: Rich Salz <rsalz@openssl.org>

Add documentation of PKCS12_newpass()

PR#4478

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Stephen Henson <steve@openssl.org>

Tidy up PKCS12_newpass() fix memory leaks.

PR#4466

Reviewed-by: Rich Salz <rsalz@openssl.org>

Only set CMS parameter when encrypting

Reviewed-by: Rich Salz <rsalz@openssl.org>

RT3513: req doesn't display attributes using utf8string

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>

README.PERL: clarify "matching" Perl requirement on Windows.

Reviewed-by: Rich Salz <rsalz@openssl.org>

poly1305/asm/poly1305-x86_64.pl: contain symbols within shared lib.

We don't need it, but external users might find it handy.

Reviewed-by: Richard Levitte <levitte@openssl.org>

poly1305/asm/poly1305-x86_64.pl: make it cross-compile.

Reviewed-by: Richard Levitte <levitte@openssl.org>

testlib/OpenSSL/Test.pm: address 5.10 warnings.

Reviewed-by: Richard Levitte <levitte@openssl.org>

test/evp_test.c: exercise different combinations of data misalignment.

Reviewed-by: Emilia Käsper <emilia@openssl.org>

Use default ASN.1 for SEED.

The default ASN.1 handling can be used for SEED. This also makes
CMS work with SEED.

PR#4504

Reviewed-by: Rich Salz <rsalz@openssl.org>

typo

Reviewed-by: Rich Salz <rsalz@openssl.org>

Move 3DES from HIGH to MEDIUM

Reviewed-by: Viktor Dukhovni <viktor@openssl.org>

Always try to set ASN.1 parameters for CMS.

Try to set the ASN.1 parameters for CMS encryption even if the IV
length is zero as the underlying cipher should still set the type.

This will correctly result in errors if an attempt is made to use
an unsupported cipher type.

Reviewed-by: Rich Salz <rsalz@openssl.org>

GH356: Change assert to normal error

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>

Handle no async jobs in libssl

If the application has limited the size of the async pool using
ASYNC_init_thread() then we could run out of jobs while trying to start a
libssl io operation. However libssl was failing to handle this and treating
it like a fatal error. It should not be fatal...we just need to retry when
there are jobs available again.

Reviewed-by: Richard Levitte <levitte@openssl.org>

Document inversion ladder in curve25519

This demystifies two for-loops that do nothing. They were used to write
the ladder in a unified way. Now that the ladder is otherwise commented,
remove the dead loops.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>

Script changed; update the generated file.

Reviewed-by: Dr. Stephen Henson <steve@openssl.org>

Improve heartbeats coding style

Based on an orignal commit by GitHub user BertramScharpf. Rebased and
updated to take account of all the updates since this was first raised.

GH PR#62

Reviewed-by: Rich Salz <rsalz@openssl.org>

Tweak generated warning lines.

Reviewed-by: Andy Polyakov <appro@openssl.org>

remove unused macros in list -disabled

list -disabled was checking OPENSSL_NO_SSL/OPENSSL_NO_TLS, which are
not used to disable SSL/TLS respectively.
Building with these macros wrongly show as SSL/TLS disabled, hence
removing this code.

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>

Fix spelling

Signed-off-by: Kurt Roeckx <kurt@roeckx.be>
Reviewed-by: Matt Caswell <matt@openssl.org>
GH: #1021

support embed in ASN.1 print

Reviewed-by: Richard Levitte <levitte@openssl.org>

Complete the list of names in doc/ssl/SSL_CTX_load_verify_locations.pod

Reviewed-by: Rich Salz <rsalz@openssl.org>

Fix name length limit check.

The name length limit check in x509_name_ex_d2i() includes
the containing structure as well as the actual X509_NAME. This will
cause large CRLs to be rejected.

Fix by limiting the length passed to ASN1_item_ex_d2i() which will
then return an error if the passed X509_NAME exceeds the length.

RT#4531

Reviewed-by: Rich Salz <rsalz@openssl.org>

Check return of PEM_write_* functions and report possible errors

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/1025)

DEFINE_STACK_OF(ASN1_UTF8STRING) moved from ts_lcl.h to asn1.h

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/1020)

Handle malloc failures in BIO_accept

The old BIO_accept() function can encounter errors during malloc. We need
to ensure we properly clean up if that occurs.

GH Issue #817

Reviewed-by: Richard Levitte <levitte@openssl.org>

reject zero block length in PKCS12 keygen

Reviewed-by: Richard Levitte <levitte@openssl.org>

Fix Blake block length

PR#4514

Reviewed-by: Richard Levitte <levitte@openssl.org>

add documentation

Reviewed-by: Matt Caswell <matt@openssl.org>

Fix double free in d2i_PrivateKey().

RT#4527

Reviewed-by: Matt Caswell <matt@openssl.org>

GH932: Add more help messages to some apps options.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>

Fix the no-tls option

The TLSProxy based tests don't work when TLS is disabled so we shouldn't
run them.

Reviewed-by: Richard Levitte <levitte@openssl.org>

Remove stale errors from early connection attempts in a client

The init_client() function in the apps sets up the client connection. It
may try multiple addresses until it finds one that works. We should clear
the error queue if we eventually get a successful connection because
otherwise we get stale errors hanging around. This can cause problems in
subsequent calls to SSL_get_error(), i.e. non-fatal NBIO events appear as
fatal.

Reviewed-by: Richard Levitte <levitte@openssl.org>

Allow spaces in filenames when using perl's glob

Reviewed-by: Matt Caswell <matt@openssl.org>

Alpha assembly pack: make it work on Linux.

Reviewed-by: Richard Levitte <levitte@openssl.org>

MIPS assembly pack: fix MIPS64 assembler warnings.

Reviewed-by: Richard Levitte <levitte@openssl.org>

Configurations/unix-Makefile.tmpl: don't count on -E -P.

Some non-Gnu compilers interpret -E -P combination differently.
some prioritize -E over -P, others -P over -E (in which case .i
file is generated and sometimes truncated because of redirection).

Reviewed-by: Richard Levitte <levitte@openssl.org>

Tru64 fixes.

Reviewed-by: Richard Levitte <levitte@openssl.org>

Drop duplicate  ctx->verify_cb assignment

The right variant is ~18 lines below.

Reviewed-by: Matt Caswell <matt@openssl.org>

Update Diffie-Hellman parameters to IANA standards

This replaces the old SKIP primes with the most current Diffie-Hellman
MODP groups defined by RFC 7296 and RFC 3526.

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from GitHub https://github.com/openssl/openssl/pull/775)

Update CHANGES and NEWS for the new release

Reviewed-by: Richard Levitte <levitte@openssl.org>

Check that we have enough padding characters.

Reviewed-by: Emilia Käsper <emilia@openssl.org>
CVE-2016-2107

MR: #2572

Add ASN.1 INTEGER tests.

Add tests for ASN.1 INTEGER: invalid tag, valid 0, 1, -1 and 0, -1 with
illegal padding.

Also add ASN1_ANY tests for 0, 1 and -1.

Reviewed-by: Emilia Käsper <emilia@openssl.org>

add ASN1_INTEGER type to d2i_test

Reviewed-by: Emilia Käsper <emilia@openssl.org>

Add test for CVE-2016-2018

Reviewed-by: Emilia Käsper <emilia@openssl.org>

Fix ASN1_INTEGER handling.

Only treat an ASN1_ANY type as an integer if it has the V_ASN1_INTEGER
tag: V_ASN1_NEG_INTEGER is an internal only value which is never used
for on the wire encoding.

Thanks to David Benjamin <davidben@google.com> for reporting this bug.

This was found using libFuzzer.

RT#4364 (part)CVE-2016-2108.

Reviewed-by: Emilia Käsper <emilia@openssl.org>

Add documentation for EVP_EncodeInit() and similar functions

Reviewed-by: Richard Levitte <levitte@openssl.org>

Ensure EVP_EncodeUpdate handles an output length that is too long

With the EVP_EncodeUpdate function it is the caller's responsibility to
determine how big the output buffer should be. The function writes the
amount actually used to |*outl|. However this could go negative with a
sufficiently large value for |inl|. We add a check for this error
condition.

Reviewed-by: Richard Levitte <levitte@openssl.org>

Avoid overflow in EVP_EncodeUpdate

An overflow can occur in the EVP_EncodeUpdate function which is used for
Base64 encoding of binary data. If an attacker is able to supply very large
amounts of input data then a length check can overflow resulting in a heap
corruption. Due to the very large amounts of data involved this will most
likely result in a crash.

Internally to OpenSSL the EVP_EncodeUpdate function is primarly used by the
PEM_write_bio* family of functions. These are mainly used within the
OpenSSL command line applications, so any application which processes
data from an untrusted source and outputs it as a PEM file should be
considered vulnerable to this issue.

User applications that call these APIs directly with large amounts of
untrusted data may also be vulnerable.

Issue reported by Guido Vranken.

CVE-2016-2105

Reviewed-by: Richard Levitte <levitte@openssl.org>

Properly own the duplicated string

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>

Improve comment

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>

Try 'make install' with one of the Travis configurations

Reviewed-by: Matt Caswell <matt@openssl.org>

Prevent EBCDIC overread for very long strings

ASN1 Strings that are over 1024 bytes can cause an overread in
applications using the X509_NAME_oneline() function on EBCDIC systems.
This could result in arbitrary stack data being returned in the buffer.

Issue reported by Guido Vranken.

CVE-2016-2176

Reviewed-by: Andy Polyakov <appro@openssl.org>

Fix encrypt overflow

An overflow can occur in the EVP_EncryptUpdate function. If an attacker is
able to supply very large amounts of input data after a previous call to
EVP_EncryptUpdate with a partial block then a length check can overflow
resulting in a heap corruption.

Following an analysis of all OpenSSL internal usage of the
EVP_EncryptUpdate function all usage is one of two forms.

The first form is like this:
EVP_EncryptInit()
EVP_EncryptUpdate()

i.e. where the EVP_EncryptUpdate() call is known to be the first called
function after an EVP_EncryptInit(), and therefore that specific call
must be safe.

The second form is where the length passed to EVP_EncryptUpdate() can be
seen from the code to be some small value and therefore there is no
possibility of an overflow.

Since all instances are one of these two forms, I believe that there can
be no overflows in internal code due to this problem.

It should be noted that EVP_DecryptUpdate() can call EVP_EncryptUpdate()
in certain code paths. Also EVP_CipherUpdate() is a synonym for
EVP_EncryptUpdate(). Therefore I have checked all instances of these
calls too, and came to the same conclusion, i.e. there are no instances
in internal usage where an overflow could occur.

This could still represent a security issue for end user code that calls
this function directly.

CVE-2016-2106

Issue reported by Guido Vranken.

Reviewed-by: Tim Hudson <tjh@openssl.org>

Fix i2d_X509_AUX: pp can be NULL.

Reported by David Benjamin

Reviewed-by: Emilia Käsper <emilia@openssl.org>

GH875: Document -no_check_time
Date:   Tue Mar 15 15:19:44 2016 +0100

This commit updates the documentation of cms, ocsp, s_client,
s_server, and verify to reflect the new "-no_check_time"
option introduced in commit d35ff2c0ade0a12e84aaa2e9841b4983a2f3cf45
on 2015-07-31.

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>

Remove confusing comment.

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>

Issue #719:

If no serverinfo extension is found in some cases, do not abort the handshake,
but simply omit/skip that extension.

Check for already-registered serverinfo callbacks during serverinfo
registration.

Update SSL_CTX_use_serverinfo() documentation to mention the need to reload the
same serverinfo per certificate, for servers with multiple server certificates.

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>

Document the addition of OPENSSL_LOCAL_CONFIG_DIR

Reviewed-by: Stephen Henson <steve@openssl.org>

Add the possibility for local build file templates

Use the environment variable OPENSSL_LOCAL_CONFIG_DIR to find build
file templates as well.

Reviewed-by: Stephen Henson <steve@openssl.org>

Add the possibility to have an extra local OpenSSL configs directory

The environment variable OPENSSL_LOCAL_CONFIG_DIR is used to indicate
that there's a local directory with extra configuration files.

Reviewed-by: Stephen Henson <steve@openssl.org>

Remove personal configs from version control

As per a team decision back in 2014.

Reviewed-by: Stephen Henson <steve@openssl.org>

Secure memory fixes

Fix some of the variables to be (s)size_t, so that more than 1GB of
secure memory can be allocated. The arena has to be a power of 2, and
2GB fails because it ends up being a negative 32-bit signed number.

The |too_late| flag is not strictly necessary; it is easy to figure
out if something is secure memory by looking at the arena. As before,
secure memory allocations will not fail, but now they can be freed
correctly. Once initialized, secure memory can still be used, even if
allocations occured before initialization.

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>

ssl/record/rec_layer_s3.c: fix typo from previous commit.

Reviewed-by: Richard Levitte <levitte@openssl.org>

Handle multi-line "written by/for" comments.

Reviewed-by: Richard Levitte <levitte@openssl.org>

Android build fixes.

Reviewed-by: Rich Salz <rsalz@openssl.org>

Remove obsolete defined(__INTEL__) condition.

This macro was defined by no-longer-supported __MWERKS__ compiler.

Reviewed-by: Richard Levitte <levitte@openssl.org>

chacha/asm/chacha-x86.pl: make it compile on legacy systems.

Usage of $ymm variable is a bit misleading here, it doesn't refer
to %ymm register bank, but rather to VEX instruction encoding,
which AMD XOP code path depends on.

Reviewed-by: Richard Levitte <levitte@openssl.org>

Fix spelling in pod files

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>

make update

Reviewed-by: Rich Salz <rsalz@openssl.org>

Add the adjusted perl scripts to the set of "update" scripts

Reviewed-by: Rich Salz <rsalz@openssl.org>