Richard Levitte [Tue, 21 Apr 2020 10:24:44 +0000 (12:24 +0200)]
Revert "TEST: make and use a fipsinstall script"
Unfortunately, this won't work on MacOS because of system integrity
measures on that platform, which clears DYLD_LIBRARY_PATH before
starting a sub-process executable.
Benjamin Kaduk [Fri, 10 Apr 2020 19:27:28 +0000 (12:27 -0700)]
Fix NULL dereference in SSL_check_chain() for TLS 1.3
In the tls1_check_sig_alg() helper function, we loop through the list of
"signature_algorithms_cert" values received from the client and attempt
to look up each one in turn in our internal table that maps wire
codepoint to string-form name, digest and/or signature NID, etc., in
order to compare the signature scheme from the peer's list against what
is used to sign the certificates in the certificate chain we're
checking. Unfortunately, when the peer sends a value that we don't
support, the lookup returns NULL, but we unconditionally dereference the
lookup result for the comparison, leading to an application crash
triggerable by an unauthenticated client.
Since we will not be able to say anything about algorithms we don't
recognize, treat NULL return from lookup as "does not match".
We currently only apply the "signature_algorithm_cert" checks on TLS 1.3
connections, so previous TLS versions are unaffected. SSL_check_chain()
is not called directly from libssl, but may be used by the application
inside a callback (e.g., client_hello or cert callback) to verify that a
candidate certificate chain will be acceptable to the client.
Richard Levitte [Mon, 6 Apr 2020 21:58:24 +0000 (23:58 +0200)]
Developer scripts: Release script
The old release script that exists in another repository has aged, and
risks becoming messy beyond maintainability if it's made to deal with
multiple OpenSSL version schemes.
A solution, which has been seen in other projects, is to have the
release script as part of the versioned source tree, and ensure it's
adapted for the ongoing version scheme in that source tree.
This introduces dev/, a directory of OpenSSL developer "stuff". We
may expand it with other practical scripts to easy development setup
and other similar things that developers may need. For now, it's the
release script dev/release.sh, with auxilliary files in dev/release-aux/.
The script is self describing, the manual is available by running the
command `./dev/release.sh --manual`.
The dev/ directory shall never appear in a source distribution.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11516)
Fix test_cmp_vfy failures with FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION
[extended tests]
Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com>
(Merged from https://github.com/openssl/openssl/pull/11585)
Matt Caswell [Sat, 18 Apr 2020 10:54:23 +0000 (11:54 +0100)]
Use a fetched MD if appropriate in ssl3_cbc_digest_record
HMACs used via the legacy EVP_DigestSign interface are strange in
that they use legacy codepath's which eventually (under the covers)
transform the operation into a new style EVP_MAC. This can mean the
digest in use can be a legacy one, so we need to be careful with any
digest we extract from the ctx.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11511)
Matt Caswell [Tue, 7 Apr 2020 15:22:49 +0000 (16:22 +0100)]
Teach ssl_test_new to have different tests for different loaded providers
We now run the tests twice: Once with no specific providers loaded and
just using the default libctx, and a second time with a non-default libctx
and the default provider.
In the second run we disable tests which use a PSS cert/key because we
don't yet have support for that.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11511)
Fix descriptions of credentials and verification options for various apps
fix doc of s_client and s_server credentials and verification options
fix doc of verification options also for s_time, x509, crl, req, ts, and verify
correcting and extending texts regarding untrusted and trusted certs,
making the order of options in the docs and help texts more consistent,
etc.
Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com> Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11273)
Fix DH_get_nid() so that it does not cache values.
DH_set0_pqg() is now responsible for caching the nid, q and length.
DH with or without named safe prime groups now default to using the maximum private key length (BN_num_bits(q) - 1)
when generating a DH private key. The code is now shared between fips and non fips mode for DH key generation.
The OSSL_PKEY_PARAM_DH_PRIV_LEN parameter can be used during keygen to override the maximum private key length to be
in the range (2 * strength ... bits(q) - 1). Where the strength depends on the length of p.
Added q = (p - 1) / 2 safe prime BIGNUMS so that the code is data driven (To simplify adding new names).
The BIGNUMS were code generated.
Fix error in documented return value for DH_get_nid
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11562)
Pauli [Thu, 12 Mar 2020 23:06:04 +0000 (09:06 +1000)]
CHANGES: note which command line utilities are marked for deprecation but still available.
Some of the utilities are much easier to use than their pkey alternatives.
These have been modified to use the PKEY APIs but still note that they are
deprecated.
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/11225)
apps/speed: fix invalid final report when run SM2 benchmarks in parallel
Fixed #11523
CLA: trivial
Signed-off-by: å¼ ç››è±ª <zhangshenghao1995@163.com> Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Paul Yang <kaishen.yy@antfin.com>
(Merged from https://github.com/openssl/openssl/pull/11557)
The erroneously introduced names grasshopper-* replaced with
kuznyechik-* according to official algorithm name translation.
Too long symbolic names replaced with human-enterable ones.
Also the mechanism of deprecating names in objects.txt is implemented
Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/11440)
Richard Levitte [Thu, 16 Apr 2020 07:49:47 +0000 (09:49 +0200)]
INSTALL: document 'no-ui-console' rather than 'no-ui'
The UI interface itself is never disabled, but the console backend may
be. 'no-ui' is a deprecated backward compatibility alias for
'no-ui-console'.
Fixes #11551
Reviewed-by: Paul Dale <paul.dale@oracle.com> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/11553)
Matt Caswell [Fri, 10 Apr 2020 17:28:24 +0000 (18:28 +0100)]
When calling the import_to function pass the libctx too
Previously import_to just took an EVP_PKEY as the argument. However we
need to some additional context data as well - specifically the libctx.
Therefore we pass an EVP_PKEY_CTX instead to hold the combination of
both of these things.
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11536)
Kochise [Mon, 16 Mar 2020 19:04:04 +0000 (20:04 +0100)]
Windows: Add type casting in CRYPTO_atomic_add to remove warning
CLA: trivial
Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/11337)
Richard Levitte [Wed, 15 Apr 2020 11:36:19 +0000 (13:36 +0200)]
TEST: Add a test of keygen with an empty template in test/evp_extra_test.c
We do it with RSA, which may seem strange. However, an RSA "template"
is generally ignored, so this is safe. This is modelled after the test
code given in github issue #11549.
Reviewed-by: Paul Dale <paul.dale@oracle.com> Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11550)
Richard Levitte [Wed, 15 Apr 2020 07:54:11 +0000 (09:54 +0200)]
EVP: Fix calls to evp_pkey_export_to_provider()
The calls weren't quite right, as this function has changed its behaviour.
We also change the internal documentation of this function, and document
evp_pkey_downgrade().
Fixes #11549
Reviewed-by: Paul Dale <paul.dale@oracle.com> Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11550)
Rich Salz [Thu, 5 Mar 2020 17:58:00 +0000 (12:58 -0500)]
Put sys-specific files in build.info
Don't wrap whole files in if[n]def, test in build.info if they
should be compiled. rand_win isn't done as there are multiple
ways to say "this is windows."
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> Reviewed-by: Paul Dale <paul.dale@oracle.com> Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11263)
Rich Salz [Thu, 5 Mar 2020 17:50:31 +0000 (12:50 -0500)]
Use build.info, not ifdef for crypto modules
Don't wrap conditionally-compiled files in global ifndef tests.
Instead, test if the feature is disabled and, if so, do not
compile it.
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> Reviewed-by: Paul Dale <paul.dale@oracle.com> Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11263)
Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11332)
Matt Caswell [Fri, 10 Apr 2020 17:27:11 +0000 (18:27 +0100)]
When calling EC_POINT_point2buf we must use a libctx
In a similar way to commit 76e23fc5 we must ensure that we use a libctx
whenever we call EC_POINT_point2buf because it can end up using crypto
algorithms.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11535)
Matt Caswell [Mon, 6 Apr 2020 17:18:18 +0000 (18:18 +0100)]
Teach d2i_PrivateKey et al about libctx
The Ed448 private key decoding makes algorithm fetches. Therefore we teach
d2i_PrivateKey et al about libctx and make sure it is passed through the
layers.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11494)
Richard Levitte [Thu, 19 Mar 2020 13:02:51 +0000 (14:02 +0100)]
EVP: Add a temporary SM2 hack to key generation
The reason to do this is many-fold. We need EC key generation for
other work. However, SM2 are currently closely related to EC keys
with legacy methods, but not with provider methods.
To avoid having to wait on provider support for SM2, we temporarly
do an extra check for what the legacy methods identify as SM2 keys
(either the EVP_PKEY_SM2 pkey id was used, or the SM2 curve), and
redirect to legacy code in one case, and in the other case, we
forcedly downgrade provider side EC keys with SM2 curves to legacy
SM2 keys, using available tools.
Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/11328)
Fix Dynamic engine loading so that the call to ENGINE_load_builtin_engines() is performed.
Fixes #11510
PR #11240 Added support for passing the libctx to the config loader.
As part of this work the call to OPENSSL_load_builtin_modules() + ENGINE_load_builtin_engines() was deferred until module_run() is called.
The call to ENGINE_load_builtin_engines() has been added to ENGINE_by_id().
Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/11543)
mkerr: remove legacy guards from generated error headers
In pull request #9333, legacy guards were added to the generated
error headers, but the mkerr.pl script was not adjusted accordingly.
So the legacy guards were removed by subsequent `make update` calls.
Fixing the mkerr.pl script properly was disproportionately complicated
by the fact that adding legacy guards only made sense for files which
already existed in version 1.1.1. To keep things simple, it was decided
to drop the legacy guards from the generated headers entirely.
Fixes #10569
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11541)
scott [Thu, 9 Apr 2020 11:36:37 +0000 (12:36 +0100)]
BIO_do_accept: correct error return value
`BIO_do_accept` was returning incorrect values when unable to bind to a port.
Fixes #7717
CLA: trivial
Reviewed-by: Viktor Dukhovni <viktor@openssl.org> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/11505)
Richard Levitte [Mon, 6 Apr 2020 11:58:41 +0000 (13:58 +0200)]
Initialize files that declare internal symbols
util/other-internal.syms is like util/other.syms, but for internal symbols.
Likewise, util/missingcrypto-internal.txt and util/missingssl-internal.txt
are like util/missingcrypto.txt and util/missingssl.txt
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11476)
Richard Levitte [Mon, 6 Apr 2020 11:51:36 +0000 (13:51 +0200)]
Rework util/find-doc-nits to distinguish internal documentation
We didn't really distinguish internal and public documentation, or
matched that with the state of the documented symbols. we therefore
needed to rework the logic to account for the state of each symbol.
To simplify things, and make them consistent, we load all of
util/*.num, util/*.syms and util/missing*.txt unconditionally.
Also, we rework the reading of the manuals to happen only once (or
well, not quite, Pod::Checker reads from file too, but at the very
least, our script isn't reading the same file multiple times).
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11476)
Patrick Steuer [Thu, 9 Apr 2020 17:58:02 +0000 (19:58 +0200)]
Fix EVP_DigestSign interface when used with DES CMAC
DES implementations were missing the dup/copy ctx routines
required by CMAC implementation. A regression test is added.
Signed-off-by: Patrick Steuer <patrick.steuer@de.ibm.com> Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11498)
Richard Levitte [Thu, 9 Apr 2020 04:07:54 +0000 (06:07 +0200)]
EVP: legacy_ctrl_to_param() to handle provider side keys
There was one spot where this function would look at ctx->pmeth
directly to determine if it's for RSASSA-PSS, which fails when
presented with an EVP_PKEY_CTX holding a provider side key.
Switching to use EVP_PKEY_is_a() should make things better.
Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/11501)
Richard Levitte [Wed, 8 Apr 2020 13:41:05 +0000 (15:41 +0200)]
CMS KARI: Temporarly downgrade newly generated EVP_PKEYs to legacy
The EVP_PKEY_ASN1_METHOD code used by CMS_RecipientInfo_kari_decrypt()
and cms_RecipientInfo_kari_encrypt() is quite complex and needs more
careful thought to work with provider side keys. Unfortunately, we
need to get key generation in place, among others for ECC keys, so we
add a temporary hack, similar to what's already done in TLS code, that
downgrades a provider side EVP_PKEY to become EVP_PKEY_ASN1_METHOD /
EVP_PKEY_METHOD based.
Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/11501)
Benjamin Kaduk [Wed, 8 Apr 2020 17:05:27 +0000 (10:05 -0700)]
Fix krb5 external test
Since commit c3845ceba84aab9ddeb43f043549238fd10de63b ("Build file
templates: don't set OPENSSL_{ENGINES,MODULES}") the krb5 external test
has been failing. This is because it relied on OPENSSL_MODULES already
being set -- even though it did assign to OPENSSL_MODULES itself (and
thus got skipped by the cleanup pass in that commit), it was doing so
only to canonicalize the existing value to an absolute path, not as a de
novo assignment.
Catch up to the rest of the tree and just set it directly as the
"providers" path from the build top (but still canonicalized to an
absolute path).
[extended tests]
Fixes: 11492 Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/11499)