Fix error message when loading engines from config

When using command line applications errors occur when trying to
load engines specified in a config file. Introduced by commit
a0a82324f96

RT#4093

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

Remove Obsolete engines

There are a number of engines in the OpenSSL source code which are now
obsolete. The following engines have been removed: 4758cca, aep, atalla,
cswift, nuron, sureware.

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

Fix self signed handling.

Don't mark a certificate as self signed if keyUsage is present and
certificate signing not asserted.

PR#3979

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

embed CRL serial number and signature fields

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

embed certificate serial number and signature fields

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

embed value field of X509_EXTENSION

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

add CHANGES entry for embed

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

Handle embed flag in ASN1_STRING_copy().

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

PACKET: fix __owur

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

Appease gcc's Wmaybe-uninitialized

False positive: gcc (4.8) can't figure out the SSL_IS_DTLS logic.

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

Remove EVP_CHECK_DES_KEY

Thanks to the OpenBSD community for bringing this to our attention.

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

ct_locl.h: fix some comments

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

make depend: prefer clang over makedepend

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

RFC5753 compliance.

RFC5753 requires that we omit parameters for AES key wrap and set them
to NULL for 3DES wrap. OpenSSL decrypt uses the received algorithm
parameters so can transparently handle either form.

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

Add Clang 3.6 and additional GCC 5 builds to travis

Follow-up to f386742.

Signed-off-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Andy Polyakov <appro@openssl.org>

GH429: Add clang to travis

Signed-off-by: Rich Salz <rsalz@akamai.com>
Reviewed-by: Andy Polyakov <appro@openssl.org>

testlib/OpenSSL/Test.pm: remove redundant 'cmd /c', MSWin32 Perl can take care of itself.

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

Harmonize do_rehash_rule with updated test/recipies/25-test_verify.t.

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

Test suite: chomp->s/\R// to harmonize with mingw 'make test'.

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

Test suite: minimal required to get mingw 'make test' work under Linux.

(part by Alessandro Ghedini)

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

Rename -set-serial command to req

Previous OpenSSL versions used -set_serial, but master was using
-set-serial - so rename it back to the old version.

RT#4059

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

Centralise loading default apps config file

Loading the config file after processing command line options can
cause problems, e.g. where an engine provides new ciphers/digests
these are not then recoginised on the command line. Move the
default config file loading to before the command line option
processing. Whilst we're doing this we might as well centralise
this instead of doing it individually for each application. Finally
if we do it before the OpenSSL_add_ssl_algorithms() call then
ciphersuites provided by an engine (e.g. GOST) can be available to
the apps.

RT#4085
RT#4086

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

Fix option name discrepancy

There used to be options -macopt and -sigopt in <=1.0.2 for the dgst
command line app. These were incorrectly spelled as -macop and -sigop in
master.

RT#4072

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

Configurations: move -Wno-pedantic-ms-format to .travis.yml.

The option is not available in older toolchains and would cause breakage.

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

embed OCSP_CERTID

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

embed support for ASN1_STRING

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

Embed various OCSP fields.

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

embed support for CHOICE type

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

Skip PSK tests for no-psk

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

Make no-psk compile without warnings.

PR#4035

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

Typo.

PR#4079

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

DTLS: remove unused cookie field

Note that this commit constifies a user callback parameter and therefore
will break compilation for applications using this callback. But unless
they are abusing write access to the buffer, the fix is trivial.

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

Fix Windows build

Add the new ct directory to mkfiles.pl and fix the ct Makefile

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

Initial commit for Certificate Transparency support

Original authors:
Rob Stradling <rob@comodo.com>
Dr. Stephen Henson <steve@openssl.org>

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

PACKET: simplify ServerKeyExchange parsing

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

Fix no-ripemd on Windows

mkdef.pl was getting confused by:

 # ifdef OPENSSL_NO_RMD160
 #  error RIPEMD is disabled.
 # endif

Changing RIPEMD to RMD160 solves it. Fix suggested by Steve Henson.

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

Don't treat a bare OCTETSTRING as DigestInfo in int_rsa_verify

The function int_rsa_verify is an internal function used for verifying an
RSA signature. It takes an argument |dtype| which indicates the digest type
that was used. Dependant on that digest type the processing of the
signature data will vary. In particular if |dtype == NID_mdc2| and the
signature data is a bare OCTETSTRING then it is treated differently to the
default case where the signature data is treated as a DigestInfo (X509_SIG).

Due to a missing "else" keyword the logic actually correctly processes the
OCTETSTRING format signature first, and then attempts to continue and
process it as DigestInfo. This will invariably fail because we already know
that it is a bare OCTETSTRING.

This failure doesn't actualy make a real difference because it ends up at
the |err| label regardless and still returns a "success" result. This patch
just cleans things up to make it look a bit more sane.

RT#4076

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

When ENGINE_add finds that id or name is missing, actually return

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

Don't check pointer we just freed, always set it to NULL.

Signed-off-by: Kurt Roeckx <kurt@roeckx.be>
Reviewed-by: Rich Salz <rsalz@openssl.org>
MR #1231

Move BN_CTX_start() call so the error case can always call BN_CTX_end().

Signed-off-by: Kurt Roeckx <kurt@roeckx.be>
Reviewed-by: Rich Salz <rsalz@openssl.org>
MR #1231

Move BN_CTX_start() call so the error case can always call BN_CTX_end().

Signed-off-by: Kurt Roeckx <kurt@roeckx.be>
Reviewed-by: Rich Salz <rsalz@openssl.org>
MR #1231

Properly check return type of DH_compute_key()

It returns -1 on error, not 0.

Signed-off-by: Kurt Roeckx <kurt@roeckx.be>
Reviewed-by: Rich Salz <rsalz@openssl.org>
MR #1231

Set flags to 0 before calling BN_with_flags()

BN_with_flags() will read the dest->flags to keep the BN_FLG_MALLOCED but
overwrites everything else.

Signed-off-by: Kurt Roeckx <kurt@roeckx.be>
Reviewed-by: Rich Salz <rsalz@openssl.org>
MR #1231

Don't advance PACKET in ssl_check_for_safari

The function ssl_check_for_safari fingerprints the incoming extensions
to see whether it is one of the broken versions of safari. However it was
failing to reset the PACKET back to the same position it started in, hence
causing some extensions to be skipped incorrectly.

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

Fix travis builds on master

-Allow mingw debug builds to fail on Travis CI
-Fix Travis email notifications config
-Rename a variable to avoid a bogus warning with old GCC
 error: declaration of ``dup'' shadows a global declaration [-Werror=shadow]
-Disable pedantic ms-format warnings with mingw
-Properly define const DH parameters
-Restore --debug flag in Travis CI builds; -d would get incorrectly passed
 to ./Configure in mingw debug builds.

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

SSLv2 compat ciphers: clarify comment

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

Don't try and parse boolean type.

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

Address more Windows warnings illuminated by mingw.

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

New cipher and cipher modes standardized in Russia

This change introduces short names and NIDs for Russian GOST ciphers
according to GOST R 34.13-2015

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

Validate ClientHello extension field length

RT#4069

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

Add PACKET_copy_all

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

ssl_sess.c: grab a copy of the session ID

The user callback takes a non-const pointer, so don't pass PACKET data
to it directly; rather, grab a local copy.

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

PACKETize and clean up ssl_bytes_to_cipher_list.

Fix alerts.

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

ssl3_get_client_hello: rearrange logic

Move all packet parsing to the beginning of the method. This limits the
SSLv2 compatibility soup to the parsing, and makes the rest of the
processing uniform.

This is also needed for simpler EMS support: EMS servers need to do an
early scan for EMS to make resumption decisions. This'll be easier when
the entire ClientHello is parsed in the beginning.

As a side effect,
1) PACKETize ssl_get_prev_session and tls1_process_ticket; and
2) Delete dead code for SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG.

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

make depend

Reviewed-by: Ben Laurie <ben@openssl.org>

Fix make depend for things being built in subdirectories

Some makedepend mechanisms remove all directory information in the
target, so a dependency can looks like this:

ssl3_record.o: record/ssl3_record.c

However, that doesn't quite suit us, our Makefile has us build
record/ssl3_record.o rather than ssl3_record.o.

To clear this up, a change to util/clean-depend.pl takes care of this
case by looking up the original file in the dependencies and restoring
the directory information from it.

Reviewed-by: Ben Laurie <ben@openssl.org>

Harmonize pointer printing and size_t-fy casts.

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

bio/bss_log.c: harmonize format string to silence -Wformat.

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

Address Windows warnings in apps/.

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

Explicitly cast INVALID_SOCKET to (int) to address warnings on Windows.

Even though SOCKET is effectively declared as (void *) on Windows, it's
not actually a pointer, but an index within per-process table of
kernel objects. The table size is actually limited and its upper limit
is far below upper limit for signed 32-bit integer. This is what makes
cast in question possible.

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

Free up ASN.1 structures at top level only.

When a decoding error in ASN.1 occurs only free up the partial structure
at the top level. This simplifies embedded handling and fixes freeing
up of structures when presented with malformed input.

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

Fix more d2i cases to properly update the input pointer

Thanks to David Benjamin <davidben@google.com> for pointing them out.

Reviewed-by: Steve Henson <steve@openssl.org>
MR #1198

Remove BIO_s_file_internal macro.

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

Revert "Custom cipher constants"

This reverts commit 3aabc1dd1c969e60bdea8f827d6645ff04c56b67.

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

Custom cipher constants

Add some custom ctrls for Engine specific use.

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

Remove the hard coded -DOPENSSL_NO_DEPRECATED from DEPFLAG

Any time you configure with enable-deprecated, make depend would
scream bloody murder.  This change has it quiet down a bit.

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

Change the DEFAULT ciphersuites to exclude DES, RC4 and RC2

This patch updates the "DEFAULT" cipherstring to be
"ALL:!COMPLEMENTOFDEFAULT:!eNULL". COMPLEMENTOFDEFAULT is now defined
internally by a flag on each ciphersuite indicating whether it should be
excluded from DEFAULT or not. This gives us control at an individual
ciphersuite level as to exactly what is in DEFAULT and what is not.

Finally all DES, RC4 and RC2 ciphersuites are added to COMPLEMENTOFDEFAULT
and hence removed from DEFAULT.

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

bn/asm/armv4-mont.pl: boost NEON performance.

Close difference gap on Cortex-A9, which resulted in further improvement
even on other processors.

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

Make update / libeay.num fix

Looks like someone forgot to do a "make update" since crypto/ts/Makefile
keeps changing.  So include that.

Second is that the declare_dh_bn macro fools the libeay.num script.
The declarations are only needed in one file (dh_rfc5114) so remove
them from the header and put the "raw" declarations directly into that
file.

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

Fix libeay.num

Removed duplicated ordinals from libeay.num

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

Add a test for duplicated ordinals

Occaisionally we have had problems where there are duplicated ordinals in
libeay.num or ssleay.num. This adds a test for this issue.

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

Display brief help if no options for list.

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

Add GOST12 cms/smime capabilities

Add additional NID references in the CMS/SMIME capabilities code to cater
for GOST12.

Patch supplied by Dmitry Belyavsky <beldmit@gmail.com>

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

Fix no-stdio build

Much related/similar work also done by
Ivan Nestlerode <ivan.nestlerode@sonos.com>

   +Replace FILE BIO's with dummy ops that fail.
   +Include <stdio.h> for sscanf() even with no-stdio (since the declaration
    is there). We rely on sscanf() to parse the OPENSSL_ia32cap environment
    variable, since it can be larger than a 'long'. And we don't rely on the
    availability of strtoull().
   +Remove OPENSSL_stderr(); not used.
   +Make OPENSSL_showfatal() do nothing (currently without stdio there's
    nothing we can do).
   +Remove file-based functionality from ssl/. The function
    prototypes were already gone, but not the functions themselves.
   +Remove unviable conf functionality via SYS_UEFI
   +Add fallback definition of BUFSIZ.
   +Remove functions taking FILE * from header files.
   +Add missing DECLARE_PEM_write_fp_const
   +Disable X509_LOOKUP_hash_dir(). X509_LOOKUP_file() was already compiled out,
    so remove its prototype.
   +Use OPENSSL_showfatal() in CRYPTO_destroy_dynlockid().
   +Eliminate SRP_VBASE_init() and supporting functions. Users will need to
    build the verifier manually instead.
   +Eliminate compiler warning for unused do_pk8pkey_fp().
   +Disable TEST_ENG_OPENSSL_PKEY.
   +Disable GOST engine as is uses [f]printf all over the place.
   +Eliminate compiler warning for unused send_fp_chars().

Signed-off-by: Rich Salz <rsalz@akamai.com>
Reviewed-by: Tim Hudson <tjh@openssl.org>

engine/e_capi.c: fix various warnings.

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

Fix -Wshadow warnings in mingw builds.

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

Fix pedantic warnings in mingw builds.

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

Fix prototypes in e_ossttest.c.

Problem was exposed in mingw64 build, or in other words on P64 platform.

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

asn1t.h: silence -Wmissing-prototype in Windows builds.

On Windows OPENSSL_EXPORT_VAR_AS_FUNCTION is defined and in a sense
this modification simply harmonizes it with "VAR_AS_VAR".

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

Rationalize .gitignore and harmonize pair of Makefiles.

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

RT3948: Some structs have confusing names.

Signed-off-by: Rich Salz <rsalz@akamai.com>
Reviewed-by: Tim Hudson <tjh@openssl.org>

Print debug info for extended master secret extension

Signed-off-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Tim Hudson <tjh@openssl.org>

RT4053: Typo in error message

Signed-off-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Tim Hudson <tjh@openssl.org>

Remove obsolete b64 demo's

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

PACKET: simplify ServerHello parsing

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

Empty NewSessionTicket: test session resumption

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

Empty session ticket: add a test

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

RT2772: accept empty SessionTicket

RFC 5077 section 3.3 says:
If the server determines that it does not want to include a
ticket after it has included the SessionTicket extension in the
ServerHello, then it sends a zero-length ticket in the
NewSessionTicket handshake message.

Previously the client would fail upon attempting to allocate a
zero-length buffer. Now, we have the client ignore the empty ticket and
keep the existing session.

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

Update year in Windows builds.

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

Harmonize util/mkrc.pl with header move.

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

SRP memory leak fix

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

Silence Wconditional-uninitialized

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

Fixing typo in PROBLEMS

Signed-off-by: Rich Salz <rsalz@akamai.com>
Reviewed-by: Andy Polyakov <appro@openssl.org>

Skylake performance results.

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

GH408 follow-on: update buflen

Some builds break, as documented in:
  https://github.com/openssl/openssl/pull/408#issuecomment-142971427
This fixes it.

Signed-off-by: Rich Salz <rsalz@akamai.com>
Reviewed-by: Tim Hudson <tjh@openssl.org>

Fix typo in previous commit.

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

Change --debug to -d for compat with old releases.

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

Remove obsolete OCSP demo

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

Change ossltest engine to manually allocate cipher_data

The ossltest engine wraps the built-in implementation of aes128-cbc.
Normally in an engine the cipher_data structure is automatically allocated
by the EVP layer. However this relies on the engine specifying up front
the size of that cipher_data structure. In the case of ossltest this value
isn't available at compile time. This change makes the ossltest engine
allocate its own cipher_data structure instead of leaving it to the EVP
layer.

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

Properly format linux-arm64ilp32 target config

Otherwise the ./config script fails with errors like:

> Operating system: x86_64-whatever-linux2
> This system (linux-x86_64) is not supported. See file INSTALL for details.

The failure was introduced by a93d3e0.

RT#4062

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