Fix build error on some platforms
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/10757)
without error handling.
This takes up the ball from #11278
without trying to solve everything at once.
[extended tests]
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Kurt Roeckx <kurt@roeckx.be>
(Merged from https://github.com/openssl/openssl/pull/11850)
Signed-off-by: Maxim Zakharov <5158255+Maxime2@users.noreply.github.com>
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11767)
This was added before the changes to the sslap/ssl_new/ssl_old tests which
run those tests with a non-default library context. It no longer adds
anything that those tests don't already do, so it can be deleted.
This also fixes a number of run-checker build failures which were failing
in this test if TLSv1.2 was disabled.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11832)
The previous commit changed the types of various objects passed between
the core and providers. Therefore the documentation needs to be updated
to reflect that.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11758)
A provider could be linked against a different version of libcrypto than
the version of libcrypto that loaded the provider. Different versions of
libcrypto could define opaque types differently. It must never occur that
a type created in one libcrypto is used directly by the other libcrypto.
This will cause crashes.
We can "cheat" for "built-in" providers that are part of libcrypto itself,
because we know that the two libcrypto versions are the same - but not for
other providers.
To ensure this does not occur we use different types names for the handful
of opaque types that are passed between the core and providers.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11758)
The call of ossl_provider_query_operation() used |data->operation_id|,
when |cur_operation| should be used.
If any ossl_provider_query_operation() call returned NULL, the loop
was stopped, when it should just continue on to the next operation.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11837)
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/11785)
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/11785)
The main fixes were errors in itemized lists "*)" instead of "*"
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
(Merged from https://github.com/openssl/openssl/pull/11770)
Also fix a nit in recent CHANGES.md update.
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
(Merged from https://github.com/openssl/openssl/pull/11770)
This also adds the more flexible and general load_key_cert_crl()
as well as helper functions get_passwd(), cleanse(), and clear_free()
to be used also in apps/cmp.c etc.
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com>
(Merged from https://github.com/openssl/openssl/pull/11755)
The alignment calculation in ssl3_setup_write incorrectly results in an
alignment allowance of
(-SSL3_RT_HEADER_LENGTH) & (SSL3_ALIGN_PAYLOAD - 1) bytes. This equals 3
in almost all cases. The maximum alignment actually used in do_ssl3_write
is (SSL3_ALIGN_PAYLOAD - 1). This equals 7 bytes in almost all cases. So
there is a potential to overrun the buffer by up to 4 bytes.
Fortunately, the encryption overhead allowed for is 80 bytes which
consists of 16 bytes for the cipher block size and 64 bytes for the MAC
output. However the biggest MAC that we ever produce is HMAC-384 which is
48 bytes - so we have a headroom of 16 bytes (i.e. more than the 4 bytes
of potential overrun).
Thanks to Nagesh Hegde for reporting this.
Fixes#11766
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/11768)
This is a 32-bit ABI build (as opposed to linux64-mips64).
Setting SIXTY_FOUR_BIT breaks hardware optimizations, at least on
octeon processors.
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11725)
Doing this is kind of contrary to how we normally do things, as this
constitutes a kind of pre-approval. However, without this, the normal
review process will modify the reviewed commits, and render the
annotated release tag invalid, which forces the person doing the
release to re-tag manually.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11630)
flag (which is not supported by older compilers).
CLA: trivial
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11815)
The problem encountered is that some arrays were deemed unnecessary by
clang, for example:
providers/common/der/der_rsa.c:424:28: error: variable 'der_aid_sha224Identifier' is not needed and will not be emitted [-Werror,-Wunneeded-internal-declaration]
static const unsigned char der_aid_sha224Identifier[] = {
^
However, these arrays are used in sizeof() expressions in other parts
of the code that's actually used, making that warning-turned-error a
practical problem. We solve this by making the array non-static,
which guarantees that the arrays will be emitted, even though
unnecessarily. Fortunately, they are very small.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11710)
There are a few things in the OpenSSL code that are known to give
warnings that we know are harmless. We test our builds accordingly.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11710)
There were a few RSA-PSS related tests that were disabled for non-default
library contexts. We now re-enable them.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11710)
To make it easier to check the generated key manually, display it
before comparing diverse other serializations.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11710)
Instead of passing it with signature->digest_verify_init() and
signature->digest_sign_init(), we pass it with signature->newctx().
This allows the digests that are indicated by RSA PSS parameters
to have a useful propquery.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11710)
We want to catch errors in passed parameters early, which requires
kowledge of the ongoing operation. Fortunately, that's possible by
re-using the EVP_PKEY_OP macros in specific init functions.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11710)
We separate out the NIST arc OIDs to a separate file, so it can be
re-used, and also the DIGEST OIDs.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11710)
In some cases, a SEQUENCE that contains only optional fields may end
up empty. In some cases, this may be represented by dropping the
SEQUENCE entirely from the encoded DER.
To do this, we detect the case where WPACKET_FLAGS_ABANDON_ON_ZERO_LENGTH
is used, and adapt accordingly.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11710)
The import and export functions are affected. We also refactor them
to assign the RSA key type more carefully.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11710)
RSA_PSS_PARAMS carries with it a lot of baggage in form of X509_ALGOR
and ASN1_INTEGER, which we would rather avoid in our providers.
Therefore, we create a parallell structure - RSA_PSS_PARAMS_30 - that
contains the same information, but uses numeric identities (*) and C
integers (**). This makes it simpler to handle.
Note that neither this structure nor its contents are passed between
libcrypto and the providers. Instead, the numeric identities are
translated to and from names, which are then passed over that
boundary.
For future considerations, we might consider dropping RSA_PSS_PARAMS
entirely. For now, it's still reserved for EVP_PKEY_ASN1_METHOD code,
which RSA_PSS_PARAMS_30 is (almost entirely) reserved for use in our
providers.
(*) We use NIDs in this case, because we already have them and because
only algorithms that libcrypto knows about are permitted in PSS
restrictions. We could use any number series we want, as long as we
know for sure what they represent.
(**) That's for saltlen and for trailerfield, which are never expect
to surpass the set of numbers that fit in a regular 'int'.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11710)
The scheme currently added is OAEP-PSSDigestAlgorithms codified.
The translator functions translate an EVP_MD into a NID, and a NID
into a name, to support the creation and parsing of OSSL_PARAM items.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11710)
The resulting function, rsa_todata(), is designed to be usable by
providers as well.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11710)
The support of restricted RSA key types (OAEP and PSS) was lacking, or
dependent on the presence of restriction parameters. For example,
this means that even though an RSA-PSS key may have been generated, it
may appear as a plain unrestricted RSA key if parameters weren't
present (which is the case when default restriction parameters are
used)
To make it clearer what an RSA key is intended for, and avoid
depending in an EVP_PKEY, we introduce RSA key types. This is done by
reserving a section of the RSA flags (4 bits, which allows a total of
16 different types).
This isn't terribly important for EVP_PKEY_ASN1_METHOD code, as that
has access to the wrapping EVP_PKEY. This is very important for
provider code, which has no access to the wrapping EVP_PKEY.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11710)
