Fix some unguarded references to EC code inside the FIPS provider.
Reviewed-by: Paul Yang <kaishen.yy@antfin.com>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9543)
When creating a BN_CTX, make sure we store it in the right variable!
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/9546)
Commit ed57f7f935 implemented the macro ERR_raise and updated err.h to use
it. A typo in err.h means that errors in the BN library are mistakenly
attributed to the RSA library.
This was found due to the following error appearing in a travis log:
00:07:CB:13:05:7F:00:00:error:0400006C:rsa routines::data greater than mod
len:crypto/bn/bn_gcd.c:613:
00:07:CB:13:05:7F:00:00:error:04000003:rsa routines::BN
lib:crypto/rsa/rsa_gen.c:393:
/home/travis/build/openssl/openssl/util/shlib_wrap.sh
/home/travis/build/openssl/openssl/apps/openssl genrsa -out rsamptest.pem
-primes 5 8192 => 1
not ok 12 - genrsa 8192p5
The line in question (crypto/bn/bn_gcd.c:613) actually looks like this:
BNerr(BN_F_BN_MOD_INVERSE_NO_BRANCH, BN_R_NO_INVERSE);
The test was checking for that error being raised, but was instead seeing
a different error and thus failing.
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Paul Yang <kaishen.yy@antfin.com>
(Merged from https://github.com/openssl/openssl/pull/9539)
We run the cipher and digest evp_test test files in FIPS mode. Some
ciphers/digests aren't available in FIPS mode so we mark those as
only being available in the default provider.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9531)
If an implict EVP_CIPHER_fetch fails then ctx->cipher should not be set
otherwise strange things will happen when trying to free the ctx.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9531)
They incorrectly said that i2d_ECDSA_SIG returns 0 on error. In fact it
returns a negative value on error.
We fix this by moving the i2d_ECDSA_SIG/d2i_ECDSA_SIG docs onto the same
page as all the other d2i/i2d docs.
Fixes#9517
Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/9533)
The macro TLS_MD_MASTER_SECRET_CONST is supposed to hold the ascii string
"extended master secret". On EBCDIC machines it actually contained the
value "extecded master secret"
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9430)
At some point in the past do_ssl3_write() used to return the number of
bytes written, or a value <= 0 on error. It now just returns a success/
error code and writes the number of bytes written to |tmpwrit|.
The SSL_MODE_RELEASE_BUFFERS code was still looking at the return code
for the number of bytes written rather than |tmpwrit|. This has the effect
that the buffers are not released when they are supposed to be.
Fixes#9490
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9505)
The krb5 external test relies on legacy algorithms. Therefore we make
use of the capability to load a config file by default, and ensure that
the config file in use by the krb5 tests loads both the legacy and default
providers.
[extended tests]
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9492)
Previously we only loaded the config file by default for libssl. Now we do
it for libcrypto too.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9492)
Running evp_test with the FIPS module has never worked because the
config file was never loaded by default. Actually loading the FIPS module
reveals lots of failures in evp_test. The following commits will enable
loading the config file by default and so we temporarily disable running
the evp_test with the FIPS module until the tests can be fixed.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9492)
The "Availablein" keyword is supposed to indicate which providers are
required in evp_test in order for a particular test to pass. Unfortunately
this didn't work. If the provider was available then the test failed.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9492)
The problem being that the "requires flush" flag was being cleared after the
the flush. The fix is to clear it before. This is a problem because the
cache flushing called RAND_bytes and if the DRBG hadn't been created yet, it
would be queried and added to the cache causing the flush code to repeat.
Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/9477)
The alternative is to use a fast and small xorshift
random number generator. The stochastic flushing doesn't require good
random numbers, just enough variety to avoid causing problems.
Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/9477)
Fix: crypto\whrlpool\wp_block.c(90) : warning C4164: '_rotl64' : intrinsic function not declared.
Fixes#9487
Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9488)
The code has been modularized so that it can be shared by algorithms.
A fixed size IV is now used instead of being allocated.
The IV is not set into the low level struct now until the update (it uses an
iv_state for this purpose).
Hardware specific methods have been added to a PROV_GCM_HW object.
The S390 code has been changed to just contain methods that can be accessed in
a modular way. There are equivalent generic methods also for the other
platforms.
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Patrick Steuer <patrick.steuer@de.ibm.com>
(Merged from https://github.com/openssl/openssl/pull/9231)
[extended tests]
This is a temporary workaround for issue #9251, which contains a full
discussion of the real problem.
As a temporary workaround, we test `EC_GROUP_new_from_ecparameters()`
against a curve that does not currently have alternative
implementations.
The proper fix is dependant on resolution of issue #8615
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9474)
In ERR_add_error_vdata(), the size of err_data had 1 added to it in
some spots, which could lead to buffer overflow.
In ERR_vset_error(), ERR_MAX_DATA_SIZE was used instead of buf_size in
the BIO_vsnprintf() call, which would lead to a buffer overflow if
such a large buffer couldn't be allocated.
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9491)
Also added EVP_CTRL_RET_UNSUPPORTED define (so magic numbers can be removed)
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9464)
CAdES : rework CAdES signing API.
Make it private, as it is unused outside library bounds.
Fix varous doc-nits.
Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
If compiled with 'no-deprecated', ERR_put_error() is undefined. We
had one spot where we were using it directly, because the file and
line information was passed from elsewhere.
Fortunately, it's possible to use ERR_raise() for that situation, and
call ERR_set_debug() immediately after and thereby override the
information that ERR_raise() stored in the error record.
util/mkerr.pl needed a small adjustment to not generate code that
won't compile in a 'no-deprecated' configuration.
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9452)
The core now supplies its own versions of ERR_new(), ERR_set_debug()
and ERR_vset_error(). This should suffice for a provider to have any
OpenSSL compatible functionlity it desires.
The main difference between the ERR functions and the core
counterparts is that the core counterparts take an OSSL_PROVIDER
parameter instead of the library number. That way, providers do not
need to know what number they have been assigned, that information
stays in the core.
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9452)
The ERR_raise() macro uses a trick in C. The following is permitted:
#include <stdio.h>
void first(void)
{
printf("Hello! ");
}
void foo(const char *bar)
{
printf("%s", bar);
}
int main()
{
/* This */
(first(),foo)("cookie");
}
ERR_raise_data() can be used to implement FUNCerr() as well, which
takes away the need for the special function ERR_put_func_error().
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9452)
The new building block are ERR_new(), ERR_set_debug(),
ERR_set_error(), ERR_vset_error(), which allocate a new error record
and set the diverse data in them. They are designed in such a way
that it's reasonably easy to create macros that use all of them but
then rely completely on the function signature of ERR_set_error() or
ERR_vset_error().
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9452)
The useful inner macros are now static inline functions. That will
make them easier to debug in the future.
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9452)
New header file, include/openssl/macros.h, which contains diverse
useful macros that we use elsewhere.
We also add the new macro OPENSSL_FUNC, which is an alias for
__FUNC__, __FUNCTION__, __FUNCSIG or __func__, depending on what the
compiler supports. In the worst case, it's an alias for the string
"(unknown function)".
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9452)
