oweals/openssl.git
4 years agoTest that SSL_get_servername returns what we expect
Matt Caswell [Thu, 26 Sep 2019 15:16:06 +0000 (16:16 +0100)]
Test that SSL_get_servername returns what we expect

Test this on both the client and the server after a normal handshake,
and after a resumption handshake. We also test what happens if an
inconsistent SNI is set between the original handshake and the resumption
handshake. Finally all of this is also tested in TLSv1.2 and TLSv1.3.

Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/10018)

4 years agoFix SSL_get_servername() and SNI behaviour
Matt Caswell [Wed, 25 Sep 2019 16:06:06 +0000 (17:06 +0100)]
Fix SSL_get_servername() and SNI behaviour

The SNI behaviour for TLSv1.3 and the behaviour of SSL_get_servername()
was not quite right, and not entirely consistent with the RFC.

The TLSv1.3 RFC explicitly says that SNI is negotiated on each handshake
and the server is not required to associate it with the session. This was
not quite reflected in the code so we fix that.

Additionally there were some additional checks around early_data checking
that the SNI between the original session and this session were
consistent. In fact the RFC does not require any such checks, so they are
removed.

Finally the behaviour of SSL_get_servername() was not quite right. The
behaviour was not consistent between resumption and normal handshakes,
and also not quite consistent with historical behaviour. We clarify the
behaviour in various scenarios and also attempt to make it match historical
behaviour as closely as possible.

Fixes #8822

Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/10018)

4 years agoFix type name typo in d2i/i2d documentation.
David Makepeace [Thu, 23 Jan 2020 01:07:18 +0000 (11:07 +1000)]
Fix type name typo in d2i/i2d documentation.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/10933)

4 years agoAdd -issuer_checks to verify options
Rich Salz [Fri, 24 Jan 2020 18:32:34 +0000 (13:32 -0500)]
Add -issuer_checks to verify options

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
(Merged from https://github.com/openssl/openssl/pull/10873)

4 years agoDocument most missing options
Rich Salz [Thu, 16 Jan 2020 18:40:52 +0000 (13:40 -0500)]
Document most missing options

Add cmd-nits make target.

Listing options should stop when it hits the "parameters" separator.

Add missing .pod.in files to doc/man1/build.info

Tweak find-doc-nits to try openssl-XXX before XXX for POD files and
change an error messavge to be more useful.

Fix the following pages: ca, cms, crl, dgst, enc,
engine, errstr, gendsa, genrsa, list, ocsp, passwd, pkcs7, pkcs12, rand,
rehash, req, rsautil, s_server, speed, s_time,
sess_id, smime, srp, ts, x509.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
(Merged from https://github.com/openssl/openssl/pull/10873)

4 years agoAdd more generated man1 doc files
Rich Salz [Mon, 20 Jan 2020 14:45:27 +0000 (09:45 -0500)]
Add more generated man1 doc files

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
(Merged from https://github.com/openssl/openssl/pull/10873)

4 years agoAdd RSA key validation to default provider
Shane Lontis [Wed, 29 Jan 2020 10:32:32 +0000 (20:32 +1000)]
Add RSA key validation to default provider

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10780)

4 years agoTLS: use EVP for HMAC throughout libssl.
Pauli [Fri, 24 Jan 2020 00:41:38 +0000 (10:41 +1000)]
TLS: use EVP for HMAC throughout libssl.

Backwards compatibility with the old ticket key call back is maintained.
This will be removed when the low level HMAC APIs are finally removed.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10836)

4 years agoDeprecate the low level HMAC functions
Pauli [Tue, 14 Jan 2020 02:11:50 +0000 (12:11 +1000)]
Deprecate the low level HMAC functions

Use of the low level HMAC functions has been informally discouraged for a
long time.  We now formally deprecate them.

Applications should instead use EVP_MAC_CTX_new(3), EVP_MAC_CTX_free(3),
EVP_MAC_init(3), EVP_MAC_update(3) and EVP_MAC_final(3).

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10836)

4 years agohmac: preprocessor indentation fixes
Pauli [Tue, 14 Jan 2020 00:49:28 +0000 (10:49 +1000)]
hmac: preprocessor indentation fixes

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10836)

4 years agoDeprecate the low level CMAC functions
Pauli [Tue, 14 Jan 2020 00:59:11 +0000 (10:59 +1000)]
Deprecate the low level CMAC functions

Use of the low level CMAC functions has been informally discouraged for a
long time.  We now formally deprecate them.

Applications should instead use EVP_MAC_CTX_new(3), EVP_MAC_CTX_free(3),
EVP_MAC_init(3), EVP_MAC_update(3) and EVP_MAC_final(3).

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10836)

4 years agocmac: preprocessor indentation fixes
Pauli [Tue, 14 Jan 2020 00:48:48 +0000 (10:48 +1000)]
cmac: preprocessor indentation fixes

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10836)

4 years agoAdapt X509_PUBKEY_set() for use with provided implementations
Richard Levitte [Sun, 12 Jan 2020 23:27:40 +0000 (00:27 +0100)]
Adapt X509_PUBKEY_set() for use with provided implementations

We do this by letting a serializer serialize the provider side key to
a DER blob formatted according to the SubjectPublicKeyInfo structure
(see RFC 5280), and deserialize it in libcrypto using the usual d2i
function.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10851)

4 years agotest_evp_extra_test.c: don't rely on exact parameter position
Richard Levitte [Wed, 22 Jan 2020 19:59:56 +0000 (20:59 +0100)]
test_evp_extra_test.c: don't rely on exact parameter position

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/10920)

4 years agoPROV: Adapt the DSA signature implementation to provide Algorithmidentifiers
Richard Levitte [Tue, 21 Jan 2020 14:05:56 +0000 (15:05 +0100)]
PROV: Adapt the DSA signature implementation to provide Algorithmidentifiers

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/10920)

4 years agoAdapt ASN1_item_sign_ctx() for use with provided keypairs
Richard Levitte [Tue, 21 Jan 2020 13:56:13 +0000 (14:56 +0100)]
Adapt ASN1_item_sign_ctx() for use with provided keypairs

The mechanism to do this is to ask the signature operation for the DER
encoded AlgorithmIdentifier that corresponds to the combination of
signature algorithm and digest algorithm.

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/10920)

4 years agoModify EVP_PKEY_CTX_new_from_pkey() to add a propquery parameter
Matt Caswell [Wed, 15 Jan 2020 11:10:43 +0000 (11:10 +0000)]
Modify EVP_PKEY_CTX_new_from_pkey() to add a propquery parameter

The function EVP_PKEY_CTX_new_from_pkey() infers the name of the
algorithm to fetch from the EVP_PKEY that has been supplied as an
argument. But there was no way to specify properties to be used during
that fetch.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10926)

4 years agoOpenSSL::Test: bring back the relative paths
Richard Levitte [Tue, 21 Jan 2020 06:53:40 +0000 (07:53 +0100)]
OpenSSL::Test: bring back the relative paths

Because there was a bug in File::Spec::Unix' abs2rel when it was given
relative paths as both PATH and BASE arguments, the directories we
deal with were made to be all absolute.  Unfortunately, this meant
getting paths in our verbose test output which are difficult to use
anywhere else (such as a separate test build made for comparison), due
to the constant need to edit all the paths all the time.

We're therefore getting back the relative paths, by doing an extra
abs2rel() in __srctop_file, __srctop_dir, __bldtop_file and
__bldtop_dir, with a 'Cwd::getcwd' call as BASE argument.

Fixes #10628

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
(Merged from https://github.com/openssl/openssl/pull/10913)

4 years agotest/recipes/30-test_evp.t: Fix multiple definition of @bffiles
Richard Levitte [Sat, 25 Jan 2020 06:55:36 +0000 (07:55 +0100)]
test/recipes/30-test_evp.t: Fix multiple definition of @bffiles

Curiously enough, perl only warned about the shadowing.  However, the
following 'plan' statement got disturbed somehow, as one could notice
the test counter say "11/?" instead of "11/25".

Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/10944)

4 years agoCheck that the default signature type is allowed
Kurt Roeckx [Thu, 2 Jan 2020 21:53:32 +0000 (22:53 +0100)]
Check that the default signature type is allowed

TLS < 1.2 has fixed signature algorithms: MD5+SHA1 for RSA and SHA1 for the
others. TLS 1.2 sends a list of supported ciphers, but allows not sending
it in which case SHA1 is used. TLS 1.3 makes sending the list mandatory.

When we didn't receive a list from the client, we always used the
defaults without checking that they are allowed by the configuration.

Reviewed-by: Paul Dale <paul.dale@oracle.com>
GH: #10784

4 years agoReplace apps/server.pem with certificate with a sha256 signature.
Kurt Roeckx [Sun, 12 Jan 2020 15:44:01 +0000 (16:44 +0100)]
Replace apps/server.pem with certificate with a sha256 signature.

It replaces apps/server.pem that used a sha1 signature with a copy of
test/certs/servercert.pem that is uses sha256.

This caused the dtlstest to start failing. It's testing connection
sbetween a dtls client and server. In particular it was checking that if
we drop a record that the handshake recovers and still completes
successfully. The test iterates a number of times. The first time
through it drops the first record. The second time it drops the second
one, and so on. In order to do this it has a hard-coded value for the
expected number of records it should see in a handshake. That's ok
because we completely control both sides of the handshake and know what
records we expect to see. Small changes in message size would be
tolerated because that is unlikely to have an impact on the number of
records. Larger changes in message size however could increase or
decrease the number of records and hence cause the test to fail.

This particular test uses a mem bio which doesn't have all the CTRLs
that the dgram BIO has. When we are using a dgram BIO we query that BIO
to determine the MTU size. The smaller the MTU the more fragmented
handshakes become. Since the mem BIO doesn't report an MTU we use a
rather small default value and get quite a lot of records in our
handshake. This has the tendency to increase the likelihood of the
number of records changing in the test if the message size changes.

It so happens that the new server certificate is smaller than the old
one. AFAICT this is probably because the DNs for the Subject and Issuer
are significantly shorter than previously. The result is that the number
of records used to transmit the Certificate message is one less than it
was before. This actually has a knock on impact for subsequent messages
and how we fragment them resulting in one less ServerKeyExchange record
too (the actual size of the ServerKeyExchange message hasn't changed,
but where in that message it gets fragmented has). In total the number
of records used in the handshake has decreased by 2 with the new
server.pem file.

Reviewed-by: Paul Dale <paul.dale@oracle.com>
GH: #10784

4 years agoEVP: Adapt EVP_PKEY Seal and Open for provider keys
Richard Levitte [Fri, 10 Jan 2020 23:04:56 +0000 (00:04 +0100)]
EVP: Adapt EVP_PKEY Seal and Open for provider keys

This affects the following function, which can now deal with provider
side keys:

- EVP_SealInit()
- EVP_OpenInit()

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/10808)

4 years agoopenssl-config: add example libssl system-defaults
Benjamin Kaduk [Fri, 24 Jan 2020 01:08:34 +0000 (17:08 -0800)]
openssl-config: add example libssl system-defaults

Provide a "simple" example for affecting the systemwide default behavior
of libssl.  The large number of mandatory nested sections makes this
less simple than the main description might suggest.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10937)

4 years agoDeprecate the low level DES functions.
Pauli [Thu, 16 Jan 2020 03:50:03 +0000 (13:50 +1000)]
Deprecate the low level DES functions.

Use of the low level DES functions has been informally discouraged for a
long time. We now formally deprecate them.

Applications should instead use the EVP APIs, e.g. EVP_EncryptInit_ex,
EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the equivalently named decrypt
functions.

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10858)

4 years agodes: fix header file preprocessor indentation
Pauli [Thu, 16 Jan 2020 03:17:25 +0000 (13:17 +1000)]
des: fix header file preprocessor indentation

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10858)

4 years agoMake sure we use RAND_bytes_ex and RAND_priv_bytes_ex in libssl
Matt Caswell [Wed, 15 Jan 2020 18:12:59 +0000 (18:12 +0000)]
Make sure we use RAND_bytes_ex and RAND_priv_bytes_ex in libssl

Now that libssl knows about libctx we should use it wherever we generate
a random number.

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10927)

4 years agoNew functions for PKCS8 attributes management - documentation
Dmitry Belyavskiy [Mon, 20 Jan 2020 13:02:26 +0000 (16:02 +0300)]
New functions for PKCS8 attributes management - documentation

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/10900)

4 years agoNew functions for PKCS8 attributes management
Dmitry Belyavskiy [Mon, 20 Jan 2020 13:01:02 +0000 (16:01 +0300)]
New functions for PKCS8 attributes management

This commit introduces functions PKCS8_pkey_add1_attr_by_OBJ and PKCS8_pkey_add1_attr

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/10900)

4 years agoModify DSA and DH keys to use a shared FFC_PARAMS struct
Shane Lontis [Fri, 24 Jan 2020 04:09:33 +0000 (14:09 +1000)]
Modify DSA and DH keys to use a shared FFC_PARAMS struct

This is required in order to share code for FIPS related parameter generation and validation routinues.
Note the 'counter' field is now stored as a integer (as that is the form required for generation/validation functions).

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10860)

4 years agoUpdate man3/verify documentation, error text
Rich Salz [Sat, 12 Oct 2019 21:45:56 +0000 (17:45 -0400)]
Update man3/verify documentation, error text

Move the x509_V_ERR_xxx definitions from openssl-verify to
X509_STORE_CTX_get_error.pod.  Add some missing ones.  Consistently
start with a lowercase letter, unless it's an acronym.

Fix some markup mistakes in X509_verify_cert.

Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
(Merged from https://github.com/openssl/openssl/pull/10132)

4 years agoConfigure: Better detection of '-static' in @{$config{LDFLAGS}}
Richard Levitte [Fri, 17 Jan 2020 11:47:44 +0000 (12:47 +0100)]
Configure: Better detection of '-static' in @{$config{LDFLAGS}}

@{$config{LDFLAGS}} isn't necessarily split up in pieces, so we need
to check for '-static' with a regexp rather than with an exact string
match.

Fixes #10867

Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/10878)

4 years agoAdd answers for EVP_PKEY_get_default_digest_name() in RSA and DSA keymgmt
Richard Levitte [Mon, 13 Jan 2020 07:54:47 +0000 (08:54 +0100)]
Add answers for EVP_PKEY_get_default_digest_name() in RSA and DSA keymgmt

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10824)

4 years agoEVP: Add EVP_PKEY_get_default_digest_name() and use it
Richard Levitte [Mon, 13 Jan 2020 07:49:44 +0000 (08:49 +0100)]
EVP: Add EVP_PKEY_get_default_digest_name() and use it

It is the provider version of EVP_PKEY_get_default_digest_nid().  We make
sure to use it in the non-legacy section of do_sigver_init() (internal
implementation for EVP_DigestSignInit() and EVP_DigestVerifyInit())

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10824)

4 years agoAdd internal maxsize macros
Richard Levitte [Wed, 22 Jan 2020 13:00:21 +0000 (14:00 +0100)]
Add internal maxsize macros

We've started to see "magic" numbers being used for certain sizes,
such as algorithm names and property query strings.

This change takes care of the few items where buffers for algorithm
names and property query strings are used.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10928)

4 years agoPROV: Fix bignum printout in text serializers
Richard Levitte [Sun, 19 Jan 2020 08:04:08 +0000 (09:04 +0100)]
PROV: Fix bignum printout in text serializers

The common routine ossl_prov_print_labeled_bignum() didn't print the
BIGNUM quite the way it should.  It treated the limbs in a big endian
fashion, when they are really organised in a little endian fashion.

Furthermore, we make it inherit the behaviour from the print of legacy
keys, where a number starting with the high bit set gets an extra zero
printed first.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10891)

4 years agoAdd DH key exchange to fips provider
Shane Lontis [Thu, 23 Jan 2020 10:33:28 +0000 (20:33 +1000)]
Add DH key exchange to fips provider

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10826)

4 years agoCheck ECC-CDH is compliant with SP800-56A-r3
Shane Lontis [Thu, 23 Jan 2020 10:17:05 +0000 (20:17 +1000)]
Check ECC-CDH is compliant with SP800-56A-r3

Added comments and cleared an intermediate result.
KAT tests already exist in evppkey.txt (Search for "KAS_ECC_CDH_PrimitiveTest")

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10838)

4 years agoRe-remove TS_VERIFY_CTX_set_certs entry from missingcrypto.txt
Shane Lontis [Thu, 23 Jan 2020 00:18:34 +0000 (10:18 +1000)]
Re-remove TS_VERIFY_CTX_set_certs entry from missingcrypto.txt

make doc-nits was failing due to a merge error on master in missingcrypto.txt.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10932)

4 years agoAvoid leak in error path of PKCS5_PBE_keyivgen
kinichiro [Sun, 12 Jan 2020 08:35:39 +0000 (17:35 +0900)]
Avoid leak in error path of PKCS5_PBE_keyivgen

CLA: trivial

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10816)

4 years agoLegacy digests can have custom control values
Dmitry Belyavskiy [Tue, 21 Jan 2020 17:08:38 +0000 (20:08 +0300)]
Legacy digests can have custom control values

Fixes #10915.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10916)

4 years agoDo not silently truncate files on perlasm errors
David Benjamin [Fri, 17 Jan 2020 21:53:56 +0000 (16:53 -0500)]
Do not silently truncate files on perlasm errors

If one of the perlasm xlate drivers crashes, OpenSSL's build will
currently swallow the error and silently truncate the output to however
far the driver got. This will hopefully fail to build, but better to
check such things.

Handle this by checking for errors when closing STDOUT (which is a pipe
to the xlate driver).

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10883)

4 years agoRussian Elbrus processors support
Dmitry Belyavskiy [Mon, 20 Jan 2020 13:25:43 +0000 (16:25 +0300)]
Russian Elbrus processors support

Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10901)

4 years agoAdd missing files to generated
Rich Salz [Mon, 13 Jan 2020 18:48:08 +0000 (13:48 -0500)]
Add missing files to generated

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10256)

4 years agoFix some missing doc links.
Rich Salz [Thu, 24 Oct 2019 20:40:11 +0000 (16:40 -0400)]
Fix some missing doc links.

Replace "=for openssl foreign manuals" with simpler syntax, it looks
like the "=for openssl ifdef" construct.
Fix some broken L<> links; add some missing foreign references and fixed
some typo's.
The WARNINGS in dhparam referred to non-existant commands so reword it.

Fixes #10109

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10256)

4 years agoDocument the SSL_CTX_with_libctx() function
Matt Caswell [Thu, 16 Jan 2020 12:29:01 +0000 (12:29 +0000)]
Document the SSL_CTX_with_libctx() function

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/10866)

4 years agoIntroduce SSL_CTX_new_with_libex()
Matt Caswell [Thu, 16 Jan 2020 12:13:09 +0000 (12:13 +0000)]
Introduce SSL_CTX_new_with_libex()

We add the ability to specify an OPENSSL_CTX (which may be NULL for the
default context) and a property query string for use during algorithm
fetch operations.

For example, in this way one SSL_CTX could be used the default provider,
and another one could be used with the FIPS provider.

At this stage we don't use these values. That will come later.

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/10866)

4 years agoPROV: Add support for error queue marks and implement in FIPS module
Richard Levitte [Wed, 15 Jan 2020 13:09:54 +0000 (14:09 +0100)]
PROV: Add support for error queue marks and implement in FIPS module

This propagates ERR_set_mark(), and ERR_clear_last_mark() and
ERR_pop_to_mark() for provider use.

Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/10803)

4 years agoEVP: clear error when falling back from failed EVP_KEYMGMT_fetch()
Richard Levitte [Fri, 10 Jan 2020 16:50:03 +0000 (17:50 +0100)]
EVP: clear error when falling back from failed EVP_KEYMGMT_fetch()

Since we're falling back to legacy, this isn't an error any more.
Among others the failed EVP_KEYMGMT_fetch() error shadows other errors
produced by the legacy code, which disrupts our test/evp_test runs.

We use the error stack mark to restore the error stack just right,
i.e. ERR_set_mark(), ERR_clear_last_mark() and ERR_pop_to_mark()

Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/10803)

4 years agoFix unwind info in crypto/rc4/asm/rc4-x86_64.pl
H.J. Lu [Thu, 16 Jan 2020 21:37:14 +0000 (13:37 -0800)]
Fix unwind info in crypto/rc4/asm/rc4-x86_64.pl

Move .cfi_startproc to the right place for RC4.  Add missing
.cfi_startproc and .cfi_endproc to RC4_options.

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/10872)

4 years agoImplement the NULL cipher in the default provider
Matt Caswell [Wed, 8 Jan 2020 16:16:22 +0000 (16:16 +0000)]
Implement the NULL cipher in the default provider

Libssl uses the null cipher in certain situations. It should be
converted to a provided cipher.

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10865)

4 years agoConvert rand_bytes_ex and rand_priv_bytes_ex to public functions
Matt Caswell [Wed, 15 Jan 2020 16:34:55 +0000 (16:34 +0000)]
Convert rand_bytes_ex and rand_priv_bytes_ex to public functions

These were initially added as internal functions only. However they will
also need to be used by libssl as well. Therefore it make sense to move
them into the public API.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10864)

4 years agoDon't register drbg_delete_thread_state twice
Matt Caswell [Wed, 15 Jan 2020 18:11:04 +0000 (18:11 +0000)]
Don't register drbg_delete_thread_state twice

drbg_delete_thread_state cleans up after both the public and the private
DRBG. It can be registered automtically by getting either of those DRBGs,
but it should not be registered twice.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10862)

4 years agoFix init_thread_stop
Matt Caswell [Wed, 15 Jan 2020 18:10:03 +0000 (18:10 +0000)]
Fix init_thread_stop

init_thread_stop maintains a linked lists of handlers that it should
call when a thread finishes. The linked list handling wasn't quite right
resulting in corrupted data.

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/10863)

4 years agolibssl: Eliminate as much use of EVP_PKEY_size() as possible
Matt Caswell [Fri, 10 Jan 2020 14:16:30 +0000 (14:16 +0000)]
libssl: Eliminate as much use of EVP_PKEY_size() as possible

Some uses were going against documented recommendations.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10798)

4 years agolibcrypto: Eliminate as much use of EVP_PKEY_size() as possible
Richard Levitte [Thu, 9 Jan 2020 20:38:47 +0000 (21:38 +0100)]
libcrypto: Eliminate as much use of EVP_PKEY_size() as possible

Some uses were going against documented recommendations.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10798)

4 years agoAPPS & TEST: Eliminate as much use of EVP_PKEY_size() as possible
Richard Levitte [Thu, 9 Jan 2020 20:37:32 +0000 (21:37 +0100)]
APPS & TEST: Eliminate as much use of EVP_PKEY_size() as possible

Some uses were going against documented recommendations.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10798)

4 years agomdc2: use evp_test instead of a separate test application.
Pauli [Mon, 13 Jan 2020 22:01:34 +0000 (08:01 +1000)]
mdc2: use evp_test instead of a separate test application.

One of the MDC2 test applications can be done using evp_test.
This makes it so.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10831)

4 years agoapps: Fix deprecation conditional in speed.c
Pauli [Thu, 16 Jan 2020 04:15:19 +0000 (14:15 +1000)]
apps: Fix deprecation conditional in speed.c

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10819)

4 years agoDeprecate the low level IDEA functions.
Pauli [Mon, 13 Jan 2020 03:02:45 +0000 (13:02 +1000)]
Deprecate the low level IDEA functions.

Use of the low level IDEA functions has been informally discouraged for a
long time. We now formally deprecate them.

Applications should instead use the EVP APIs, e.g. EVP_EncryptInit_ex,
EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the equivalently named decrypt
functions.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10819)

4 years agoidea: fix preprocessor indention
Pauli [Mon, 13 Jan 2020 02:50:08 +0000 (12:50 +1000)]
idea: fix preprocessor indention

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10819)

4 years agoparam_bld: add a padded BN call.
Pauli [Tue, 14 Jan 2020 09:36:39 +0000 (19:36 +1000)]
param_bld: add a padded BN call.

To aviod leaking size information when passing private value using the
OSSL_PARAM builder, a padded BN call is required.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10840)

4 years agoTODO: undo md5.h and sha.h changes temporarily
Pauli [Fri, 10 Jan 2020 02:43:47 +0000 (12:43 +1000)]
TODO: undo md5.h and sha.h changes temporarily

Undo the changes to md5.h and sha.h so that the low level symbols are
exported from libcrypto again.  This allows libssl to build and link.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10791)

4 years agoDeprecate the low level MD5 functions.
Pauli [Wed, 15 Jan 2020 01:13:03 +0000 (11:13 +1000)]
Deprecate the low level MD5 functions.

Use of the low level MD5 functions has been informally discouraged for a long
time.  We now formally deprecate them.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10791)

4 years agoDeprecate the low level SHA functions.
Pauli [Thu, 9 Jan 2020 03:14:13 +0000 (13:14 +1000)]
Deprecate the low level SHA functions.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10791)

4 years agosha: fix preprocessor indentation
Pauli [Thu, 9 Jan 2020 02:04:54 +0000 (12:04 +1000)]
sha: fix preprocessor indentation

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10791)

4 years agoAdd GNU properties note for Intel CET in x86_64-xlate.pl
Richard Levitte [Fri, 17 Jan 2020 07:29:28 +0000 (08:29 +0100)]
Add GNU properties note for Intel CET in x86_64-xlate.pl

This appears to be emitted with gcc and clang with -fcf-protection
selected, so we should do the same.

We're trying to be smart, and only emit this when the 'endbranch'
pseudo-mnemonic has been used at least once.

This is inspired by and owes to work done by @hjl-tools (github)

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/10875)

4 years agoEVP: Add evp_pkey_make_provided() and refactor around it
Richard Levitte [Tue, 14 Jan 2020 13:11:47 +0000 (14:11 +0100)]
EVP: Add evp_pkey_make_provided() and refactor around it

The code to ensure that an EVP_PKEY is exported to providers is
repeated all over the place, enough that copying it again has the
usual future hazards with code copying.

Instead, we refactor that code into one function,
evp_pkey_make_provided(), and make sure to use that everywhere.
It relies on the creation of EVP_PKEY_CTX to figure out facts about
the input key, should it need to.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10850)

4 years agoCORE: renumber OSSL_FUNC_KEYMGMT macros
Richard Levitte [Fri, 10 Jan 2020 17:02:05 +0000 (18:02 +0100)]
CORE: renumber OSSL_FUNC_KEYMGMT macros

An amount of upcoming work does this to make space for new functions
in different groups.

Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/10804)

4 years agoDSA: Move DSA_security_bits() and DSA_bits()
Richard Levitte [Mon, 13 Jan 2020 11:28:05 +0000 (12:28 +0100)]
DSA: Move DSA_security_bits() and DSA_bits()

... to make them accessible from the FIPS provider module.

Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/10778)

4 years agoAdd CHANGES entry regarding the documentation of EVP_PKEY_size() et al
Richard Levitte [Sun, 12 Jan 2020 00:23:43 +0000 (01:23 +0100)]
Add CHANGES entry regarding the documentation of EVP_PKEY_size() et al

Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/10778)

4 years agoDOC: Make EVP_SignInit.pod conform with man-pages(7)
Richard Levitte [Wed, 8 Jan 2020 10:08:06 +0000 (11:08 +0100)]
DOC: Make EVP_SignInit.pod conform with man-pages(7)

Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/10778)

4 years agoDOC: New file for EVP_PKEY_size(), EVP_PKEY_bits() and EVP_PKEY_security_bits()
Richard Levitte [Wed, 8 Jan 2020 10:04:15 +0000 (11:04 +0100)]
DOC: New file for EVP_PKEY_size(), EVP_PKEY_bits() and EVP_PKEY_security_bits()

We change the description to be about the key rather than the
signature.  How the key size is related to the signature is explained
in the description of EVP_SignFinal() anyway.

Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/10778)

4 years agoTEST: Adapt test/evp_pkey_provided_test.c to check the key size
Richard Levitte [Wed, 8 Jan 2020 02:50:33 +0000 (03:50 +0100)]
TEST: Adapt test/evp_pkey_provided_test.c to check the key size

This is for the case where we build keys from user data

Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/10778)

4 years agoPROV: Adapt the RSA, DSA and DH KEYMGMT implementations
Richard Levitte [Wed, 8 Jan 2020 02:49:08 +0000 (03:49 +0100)]
PROV: Adapt the RSA, DSA and DH KEYMGMT implementations

They now all respond to requests for key size, bits and security bits.

Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/10778)

4 years agoEVP: make EVP_PKEY_{bits,security_bits,size} work with provider only keys
Richard Levitte [Wed, 8 Jan 2020 02:44:28 +0000 (03:44 +0100)]
EVP: make EVP_PKEY_{bits,security_bits,size} work with provider only keys

These functions relied entirely on the presence of 'pkey->pmeth',
which is NULL on provider only keys.  This adds an interface to get
domparam and key data from a provider, given corresponding provider
data (the actual domparam or key).

The retrieved data is cached in the EVP_PKEY structure (lending the
idea from provided EVP_CIPHER).

Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/10778)

4 years agoModify EVP_CIPHER_is_a() and EVP_MD_is_a() to handle legacy methods too
Richard Levitte [Wed, 15 Jan 2020 00:04:37 +0000 (01:04 +0100)]
Modify EVP_CIPHER_is_a() and EVP_MD_is_a() to handle legacy methods too

These functions would only handle provided methods, but there are
cases where the caller just passes along a received method without
knowing the underlying method tech, so might pass along a legacy
method.  We therefore need to have them handle this case as well so
they don't cause any unnecessary surprises.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10845)

4 years agoFor all assembler scripts where it matters, recognise clang > 9.x
Richard Levitte [Wed, 15 Jan 2020 20:14:05 +0000 (21:14 +0100)]
For all assembler scripts where it matters, recognise clang > 9.x

Fixes #10853

Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/10855)

4 years agoBuild file templates: Use explicit files instead of $< or $? for pods
Richard Levitte [Wed, 15 Jan 2020 07:28:46 +0000 (08:28 +0100)]
Build file templates: Use explicit files instead of $< or $? for pods

When generating html or manpages from POD files, we used $< or $? to
get the file name to process.  It turns out, though, that some make
implementations only define $< with implicit rules, so its expansion
remains empty in explicit rules.  $? is a fine replacement, but only
as long as we have one dependency, so it may cause problems in the
future.

The final solution seems to be to use explicit POD file names
instead.  That leaves no doubts.

Fixes #10817

Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/10849)

4 years agonews: combined NEWS entry for deprecated low level cipher functions
Pauli [Thu, 16 Jan 2020 04:05:05 +0000 (14:05 +1000)]
news: combined NEWS entry for deprecated low level cipher functions

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10820)

4 years agochanges: combined CHANGES entry for deprecated low level cipher functions.
Pauli [Mon, 13 Jan 2020 03:19:00 +0000 (13:19 +1000)]
changes: combined CHANGES entry for deprecated low level cipher functions.

[skip ci]

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10820)

4 years agoAdd duplication APIs to ASN1_TIME and related types
Paul Yang [Mon, 13 Jan 2020 06:26:11 +0000 (14:26 +0800)]
Add duplication APIs to ASN1_TIME and related types

Fixes #10600.

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10823)

4 years agoDigest function deprecation CHANGES.
Pauli [Fri, 10 Jan 2020 12:14:27 +0000 (22:14 +1000)]
Digest function deprecation CHANGES.

Add a changes entry to cover the deprecation of the low level digest functions:
    MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256, SHA384, SHA512 and
    Whirlpool

[skip ci]

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10802)

4 years agoUpdate SSL_CTX_sess_set_new_cb(3) docs for refcounts
Benjamin Kaduk [Wed, 15 Jan 2020 00:22:52 +0000 (16:22 -0800)]
Update SSL_CTX_sess_set_new_cb(3) docs for refcounts

The existing documentation for the new-session callback was unclear
about the requirements on the callback with respect to reference-handling
of the session object being created.  Be more explicit about the
(non-)requirements on the callback code for "success" (1) and "ignore"
(0) return values.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10848)

4 years agoMove the stored namemap pre-population to namemap construction
Richard Levitte [Wed, 15 Jan 2020 00:10:42 +0000 (01:10 +0100)]
Move the stored namemap pre-population to namemap construction

Prepopulation of the stored namemap from the legacy method object
database happened on first EVP fetch.  However, there are moments when
that prepopulation needs to happen even though no fetching has been
performed yet.  We therefore move pre-population to happen when the
namemap is constructed.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10846)

4 years agoAdd doc for TS_VERIFY_CTX_set_certs()
Paul Yang [Tue, 10 Sep 2019 05:08:29 +0000 (13:08 +0800)]
Add doc for TS_VERIFY_CTX_set_certs()

This addition is based on PR #9472.

Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
(Merged from https://github.com/openssl/openssl/pull/9838)

4 years agoPROV: Adapt the DSA keymgmt implementation to no ex_fields
Richard Levitte [Tue, 14 Jan 2020 01:35:29 +0000 (02:35 +0100)]
PROV: Adapt the DSA keymgmt implementation to no ex_fields

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/10837)

4 years agoCRYPTO: Remove support for ex_data fields when building the FIPS module
Richard Levitte [Tue, 14 Jan 2020 01:32:42 +0000 (02:32 +0100)]
CRYPTO: Remove support for ex_data fields when building the FIPS module

These fields are purely application data, and applications don't reach
into the bowels of the FIPS module, so these fields are never used
there.

Fixes #10835

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/10837)

4 years agoDeprecate the low level RC5 functions
Pauli [Tue, 14 Jan 2020 00:33:03 +0000 (10:33 +1000)]
Deprecate the low level RC5 functions

Use of the low level RC5 functions has been informally discouraged for a long
time.  We now formally deprecate them.

Applications should instead use the EVP APIs, e.g. EVP_EncryptInit_ex,
EVP_EncryptUpdate, EVP_EncryptFinal_ex and the equivalently named decrypt
functions.

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10834)

4 years agorc5: fix preprocessor indentation
Pauli [Mon, 13 Jan 2020 23:03:55 +0000 (09:03 +1000)]
rc5: fix preprocessor indentation

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10834)

4 years agoDeprecate the low level RC4 functions
Pauli [Mon, 13 Jan 2020 23:38:09 +0000 (09:38 +1000)]
Deprecate the low level RC4 functions

Use of the low level RC4 functions has been informally discouraged for a long
time.  We now formally deprecate them.

Applications should instead use the EVP APIs, e.g. EVP_EncryptInit_ex,
EVP_EncryptUpdate, EVP_EncryptFinal_ex and the equivalently named decrypt
functions.

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10834)

4 years agorc4: fix preprocessor indentation
Pauli [Mon, 13 Jan 2020 23:03:55 +0000 (09:03 +1000)]
rc4: fix preprocessor indentation

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10834)

4 years agoDeprecate the low level RC2 functions
Pauli [Mon, 13 Jan 2020 23:15:18 +0000 (09:15 +1000)]
Deprecate the low level RC2 functions

Use of the low level RC2 functions has been informally discouraged for a
long time. We now formally deprecate them.

Applications should instead use the EVP APIs, e.g. EVP_EncryptInit_ex,
EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the equivalently named decrypt
functions.

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10834)

4 years agorc2: fix preprocessor indentation
Pauli [Mon, 13 Jan 2020 23:03:54 +0000 (09:03 +1000)]
rc2: fix preprocessor indentation

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10834)

4 years agoDeprecate the low level SEED functions
Pauli [Mon, 13 Jan 2020 22:35:12 +0000 (08:35 +1000)]
Deprecate the low level SEED functions

Use of the low level SEED functions has been informally discouraged for a
long time. We now formally deprecate them.

Applications should instead use the EVP APIs, e.g. EVP_EncryptInit_ex,
EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the equivalently named decrypt
functions.

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10833)

4 years agoseed: fix preprocessor indentation
Pauli [Mon, 13 Jan 2020 22:18:58 +0000 (08:18 +1000)]
seed: fix preprocessor indentation

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10833)

4 years agoRestoring correct check for legacy PKEY
Dmitry Belyavskiy [Tue, 14 Jan 2020 08:09:30 +0000 (11:09 +0300)]
Restoring correct check for legacy PKEY

The fix inroduced in #10758 was rolled back by accident.
Restoring it.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10839)

4 years agoAdd FIPS Self test kats for digests
Shane Lontis [Wed, 15 Jan 2020 00:48:01 +0000 (10:48 +1000)]
Add FIPS Self test kats for digests

Added an API to optionally set a self test callback.
The callback has the following 2 purposes
(1) Output information about the KAT tests.
(2) Allow the ability to corrupt one of the KAT's
The fipsinstall program uses the API.

Some KATS are not included in this PR since the required functionality did not yet exist in the provider.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10374)

4 years agoChange returned -2 to 0 in EVP_Digest{Sign,Verify}Init()
Richard Levitte [Sun, 12 Jan 2020 00:05:01 +0000 (01:05 +0100)]
Change returned -2 to 0 in EVP_Digest{Sign,Verify}Init()

The returned -2 was to mark when these operations are unsupported.
However, that breaks away from the previous API and expectations, and
there's not enough justification for that not being zero.

Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com>
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10815)

4 years agoAvoid leak in error path of asn1_parse2
kinichiro [Thu, 9 Jan 2020 14:22:25 +0000 (23:22 +0900)]
Avoid leak in error path of asn1_parse2

CLA: trivial

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10794)

4 years agoPROV: Fix mixup between general and specialized GCM implementations
Richard Levitte [Wed, 8 Jan 2020 13:58:34 +0000 (14:58 +0100)]
PROV: Fix mixup between general and specialized GCM implementations

providers/implementations/ciphers/ciphercommon_gcm_hw.c had an AES
specific GCM update function, while
providers/implementations/ciphers/cipher_aria_gcm_hw.c had the more
general implementation.

This moves them around to have the more general implementation in the
common source, and place the AES specialiation where it belongs.

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/10783)