root
/
openssl
mirror of https://github.com/openssl/openssl.git
1
0
Fork 0
Code Issues Actions 7 Packages Projects Releases Wiki Activity
openssl/test/evp_extra_test.c

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

6734 lines
242 KiB
C
Raw Normal View History

Copyright consolidation 02/10 Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-05-18 02:20:24 +08:00
/*
Copyright year updates Reviewed-by: Neil Horman <nhorman@openssl.org> Release: yes (cherry picked from commit 0ce7d1f355c1240653e320a3f6f8109c1f05f8c0) Reviewed-by: Hugo Landau <hlandau@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/24034)
2024-03-20 20:07:54 +08:00
* Copyright 2015-2024 The OpenSSL Project Authors. All Rights Reserved.
Import evp_test.c from BoringSSL. Unfortunately we already have a file called evp_test.c, so I have called this one evp_extra_test.c Reviewed-by: Emilia Käsper <emilia@openssl.org>
2015-02-09 17:45:35 +08:00
*
Following the license change, modify the boilerplates in test/ Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7767)
2018-12-06 20:05:25 +08:00
* Licensed under the Apache License 2.0 (the "License"). You may not use
Copyright consolidation 02/10 Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-05-18 02:20:24 +08:00
* this file except in compliance with the License. You can obtain a copy
* in the file LICENSE in the source distribution or at
* https://www.openssl.org/source/license.html
Import evp_test.c from BoringSSL. Unfortunately we already have a file called evp_test.c, so I have called this one evp_extra_test.c Reviewed-by: Emilia Käsper <emilia@openssl.org>
2015-02-09 17:45:35 +08:00
*/
EVP: deprecate the EVP_X_meth_ functions. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/11082)
2020-02-13 09:00:57 +08:00
/* We need to use some deprecated APIs */
#define OPENSSL_SUPPRESS_DEPRECATED
Import evp_test.c from BoringSSL. Unfortunately we already have a file called evp_test.c, so I have called this one evp_extra_test.c Reviewed-by: Emilia Käsper <emilia@openssl.org>
2015-02-09 17:45:35 +08:00
#include <stdio.h>
#include <stdlib.h>
Add a test for the raw private/public key getters Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6394)
2018-06-01 22:06:52 +08:00
#include <string.h>
Import evp_test.c from BoringSSL. Unfortunately we already have a file called evp_test.c, so I have called this one evp_extra_test.c Reviewed-by: Emilia Käsper <emilia@openssl.org>
2015-02-09 17:45:35 +08:00
#include <openssl/bio.h>
Add fips module integrity check Add environment variable for setting CONF .include path Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/9769)
2019-09-15 17:55:10 +08:00
#include <openssl/conf.h>
Import evp_test.c from BoringSSL. Unfortunately we already have a file called evp_test.c, so I have called this one evp_extra_test.c Reviewed-by: Emilia Käsper <emilia@openssl.org>
2015-02-09 17:45:35 +08:00
#include <openssl/crypto.h>
#include <openssl/err.h>
#include <openssl/evp.h>
#include <openssl/x509.h>
Support setting SM2 ID zero-length ID is allowed, but it's not allowed to skip the ID. Fixes: #6534 Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7113)
2018-09-04 17:21:10 +08:00
#include <openssl/pem.h>
Add a test where we reuse the EVP_PKEY_CTX for two HKDF test runs Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7501)
2018-10-26 19:45:27 +08:00
#include <openssl/kdf.h>
Add a test for EVP_MD_fetch Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/8513)
2019-03-18 22:36:41 +08:00
#include <openssl/provider.h>
Revise EVP_PKEY param handling We add new functions for getting parameters and discovering the gettable and settable parameters. We also make EVP_PKEY_CTX_get_signature_md() a function and implement it in terms of the new functions. This enables applications to discover the set of parameters that are supported for a given algorithm implementation. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/9753)
2019-09-04 19:46:02 +08:00
#include <openssl/core_names.h>
test/evp_extra_test.c: adapt for RSA signature tests Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/10557)
2019-12-02 18:26:15 +08:00
#include <openssl/params.h>
Deprecate RSA harder This deprecates all functions that deal with the types RSA and RSA_METHOD Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/13096)
2020-10-04 22:34:31 +08:00
#include <openssl/param_build.h>
Revise EVP_PKEY param handling We add new functions for getting parameters and discovering the gettable and settable parameters. We also make EVP_PKEY_CTX_get_signature_md() a function and implement it in terms of the new functions. This enables applications to discover the set of parameters that are supported for a given algorithm implementation. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/9753)
2019-09-04 19:46:02 +08:00
#include <openssl/dsa.h>
Test that EVP_PKEY_set1_DH() correctly identifies the DH type Provide a test to check tat when we assign a DH object we know whether we are dealing with PKCS#3 or X9.42 DH keys. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10593)
2019-12-09 20:03:02 +08:00
#include <openssl/dh.h>
Extend test_CMAC_keygen in evp_extra_test The test only setup the generation of a key. It did not complete that generation. We extend the test to complete the generation and to use the generated key. We also compare the result with a key generated in a different way. Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/12637)
2020-08-12 21:41:12 +08:00
#include <openssl/aes.h>
Deprecate RSA harder This deprecates all functions that deal with the types RSA and RSA_METHOD Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/13096)
2020-10-04 22:34:31 +08:00
#include <openssl/decoder.h>
EVP: fix keygen for EVP_PKEY_RSA_PSS This commit attempts to fix the an issue when generating a key of type EVP_PKEY_RSA_PSS. Currently, EVP_PKEY_CTX_set_rsa_keygen_bits will return -1 if the key id is not of type EVP_PKEY_RSA. This commit adds EVP_PKEY_RSA_PSS to also be accepted. The macro EVP_PKEY_CTX_set_rsa_pss_keygen_md si converted into a function and it is now called in legacy_ctrl_to_param. Fixes #12384 Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13099)
2020-10-09 12:07:43 +08:00
#include <openssl/rsa.h>
Add tests for ENGINE problems Add some tests which would have caught the issues fixed in the previous 3 commits related to engine handling. Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16846)
2021-10-15 23:30:45 +08:00
#include <openssl/engine.h>
Check for private key existence before calling eddsa sign functions Fixes #19524 Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/19525)
2022-10-28 11:47:50 +08:00
#include <openssl/proverr.h>
Multi-variant ML-KEM This introduces support for ML-KEM-512 and ML-KEM-1024 using the same underlying implementation parameterised by a few macros for the associated types and constants. KAT tests are added for ML-KEM 512 and 1024, to complement the previous tests for ML-KEM-768. MLKEM{512,768,1024} TLS "group" codepoints are updated to match the final IANA assigments and to make the additional KEMs known to the TLS layer. The pure-QC MLKEMs are not in the default list of supported groups, and need to be explicitly enabled by the application. Future work will introduce support for hybrids, and for more fine-grained policy of which keyshares a client should send by default, and when a server should request (HRR) a new mutually-supported group that was not sent. Tests for ML-KEM key exchange added to sslapitest to make sure that our TLS client MLKEM{512,768,1024} implementations interoperate with our TLS server, and that MLKEM* are not negotiated in TLS 1.2. Tests also added to excercise non-derandomised ML-KEM APIs, both directly (bypassing the provider layer), and through the generic EVP KEM API (exercising the provider). These make sure that RNG input is used correctly (KAT tests bypass the RNG by specifying seeds). The API interface to the provider takes an "const ML_KEM_VINFO" pointer, (obtained from ossl_ml_kem_get_vinfo()). This checks input and output buffer sizes before passing control to internal code that assumes correctly sized (for each variant) buffers. The original BoringSSL API was refactored to eliminate the opaque public/private key structure wrappers, since these structures are an internal detail between libcrypto and the provider, they are not part of the public (EVP) API. New "clangover" counter-measures added, refined with much appreciated input from David Benjamin (Chromium). The internal steps of "encrypt_cpa" were reordered to reduce the working-set size of the algorithm, now needs space for just two temporary "vectors" rather than three. The "decap" function now process the decrypted message in one call, rather than three separate calls to scalar_decode_1, scalar_decompress and scalar_add. Some loops were unrolled, improving performance of en/decapsulate (pre-expanded vectors and matrix) by around 5%. To handle, however unlikely, the SHA3 primitives not behaving like "pure" functions and failing, the implementation of `decap` was modifed: - To use the KDF to compute the Fujisaki-Okamoto (FO) failure secret first thing, and if that fails, bail out returning an error, a shared secret is still returned at random from the RNG, but it is OK for the caller to not use it. - If any of the subsequently used hash primitives fail, use the computed FO failure secret (OK, despite no longer constant-time) and return success (otherwise the RNG would replace the result). - We quite reasonably assume that chosen-ciphertext attacks (of the correct length) cannot cause hash functions to fail in a manner the depends on the private key content. Support for ML-KEM-512 required adding a centered binomial distribution helper function to deal with η_1 == 3 in just that variant. Some additional comments were added to highlight how the code relates to the ML-KEM specification in FIPS 203. Reviewed-by: Paul Dale <ppzgs1@gmail.com> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26172)
2024-11-30 22:20:58 +08:00
#include <openssl/rand.h>
# include <crypto/ml_kem.h>
Convert more tests ct_test,evp_extra_test,wpackettest,packettest Add strncmp TEST wrappers And make some style/consistency fixes to ct_test Silence travis; gcc bug? Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/3234)
2017-04-19 04:33:15 +08:00
#include "testutil.h"
Use "" not <> for internal/ includes Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/4217)
2017-08-22 20:35:43 +08:00
#include "internal/nelem.h"
Don't pass a digest-size to signature implementations It turns out this was never necessary, as the implementation should always check the default digest size anyway. Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/10947)
2020-02-02 19:55:05 +08:00
#include "internal/sizes.h"
Reorganize private crypto header files Currently, there are two different directories which contain internal header files of libcrypto which are meant to be shared internally: While header files in 'include/internal' are intended to be shared between libcrypto and libssl, the files in 'crypto/include/internal' are intended to be shared inside libcrypto only. To make things complicated, the include search path is set up in such a way that the directive #include "internal/file.h" could refer to a file in either of these two directoroes. This makes it necessary in some cases to add a '_int.h' suffix to some files to resolve this ambiguity: #include "internal/file.h" # located in 'include/internal' #include "internal/file_int.h" # located in 'crypto/include/internal' This commit moves the private crypto headers from 'crypto/include/internal' to 'include/crypto' As a result, the include directives become unambiguous #include "internal/file.h" # located in 'include/internal' #include "crypto/file.h" # located in 'include/crypto' hence the superfluous '_int.h' suffixes can be stripped. The files 'store_int.h' and 'store.h' need to be treated specially; they are joined into a single file. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/9333)
2019-09-28 06:45:33 +08:00
#include "crypto/evp.h"
Add test for context duplication failure This checks that the first operation successfully completes even if context duplication fails. But follwing operations get errors as if the context was finlised. Signed-off-by: Simo Sorce <simo@redhat.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com> (Merged from https://github.com/openssl/openssl/pull/20375)
2023-03-07 07:38:00 +08:00
#include "fake_rsaprov.h"
test: add evp_extra_test case for cipher pipeline API with fake pipeline provider Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/24636)
2024-08-31 20:03:49 +08:00
#include "fake_pipelineprov.h"
Import evp_test.c from BoringSSL. Unfortunately we already have a file called evp_test.c, so I have called this one evp_extra_test.c Reviewed-by: Emilia Käsper <emilia@openssl.org>
2015-02-09 17:45:35 +08:00
Statically link legacy provider to evp_extra_test Like in #17345, evp_extra_test links libcrypto statically, but also has a dynamic/shared load via the legacy provider, which leads to ambiguous behavior in evp_extra_test on some platforms, usually a crash (SIGSEGV) on exit via the atexit handlers. Statically link the legacy provider to avoid this. Fixes #22819 Helped-by: Neil Horman <nhorman@openssl.org> Helped-by: Tomas Mraz <tomas@openssl.org> Signed-off-by: Randall S. Becker <randall.becker@nexbridge.ca> Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/22905)
2023-12-01 00:20:34 +08:00
#ifdef STATIC_LEGACY
OSSL_provider_init_fn ossl_legacy_provider_init;
#endif
Rename OPENSSL_CTX prefix to OSSL_LIB_CTX Many of the new types introduced by OpenSSL 3.0 have an OSSL_ prefix, e.g., OSSL_CALLBACK, OSSL_PARAM, OSSL_ALGORITHM, OSSL_SERIALIZER. The OPENSSL_CTX type stands out a little by using a different prefix. For consistency reasons, this type is renamed to OSSL_LIB_CTX. Reviewed-by: Paul Dale <paul.dale@oracle.com> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/12621)
2020-10-15 17:55:50 +08:00
static OSSL_LIB_CTX *testctx = NULL;
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
static char *testpropq = NULL;
static OSSL_PROVIDER *nullprov = NULL;
static OSSL_PROVIDER *deflprov = NULL;
static OSSL_PROVIDER *lgcyprov = NULL;
Add some tests for the newly added raw private/public key functions We already had soem tests for the older raw private/public key functions so we expand those to call the new versions as well and pass in a libctx. Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/11635)
2020-04-28 00:17:05 +08:00
Import evp_test.c from BoringSSL. Unfortunately we already have a file called evp_test.c, so I have called this one evp_extra_test.c Reviewed-by: Emilia Käsper <emilia@openssl.org>
2015-02-09 17:45:35 +08:00
/*
* kExampleRSAKeyDER is an RSA private key in ASN.1, DER format. Of course, you
* should never use this key anywhere but in an example.
*/
static const unsigned char kExampleRSAKeyDER[] = {
0x30, 0x82, 0x02, 0x5c, 0x02, 0x01, 0x00, 0x02, 0x81, 0x81, 0x00, 0xf8,
0xb8, 0x6c, 0x83, 0xb4, 0xbc, 0xd9, 0xa8, 0x57, 0xc0, 0xa5, 0xb4, 0x59,
0x76, 0x8c, 0x54, 0x1d, 0x79, 0xeb, 0x22, 0x52, 0x04, 0x7e, 0xd3, 0x37,
0xeb, 0x41, 0xfd, 0x83, 0xf9, 0xf0, 0xa6, 0x85, 0x15, 0x34, 0x75, 0x71,
0x5a, 0x84, 0xa8, 0x3c, 0xd2, 0xef, 0x5a, 0x4e, 0xd3, 0xde, 0x97, 0x8a,
0xdd, 0xff, 0xbb, 0xcf, 0x0a, 0xaa, 0x86, 0x92, 0xbe, 0xb8, 0x50, 0xe4,
0xcd, 0x6f, 0x80, 0x33, 0x30, 0x76, 0x13, 0x8f, 0xca, 0x7b, 0xdc, 0xec,
0x5a, 0xca, 0x63, 0xc7, 0x03, 0x25, 0xef, 0xa8, 0x8a, 0x83, 0x58, 0x76,
0x20, 0xfa, 0x16, 0x77, 0xd7, 0x79, 0x92, 0x63, 0x01, 0x48, 0x1a, 0xd8,
0x7b, 0x67, 0xf1, 0x52, 0x55, 0x49, 0x4e, 0xd6, 0x6e, 0x4a, 0x5c, 0xd7,
0x7a, 0x37, 0x36, 0x0c, 0xde, 0xdd, 0x8f, 0x44, 0xe8, 0xc2, 0xa7, 0x2c,
0x2b, 0xb5, 0xaf, 0x64, 0x4b, 0x61, 0x07, 0x02, 0x03, 0x01, 0x00, 0x01,
0x02, 0x81, 0x80, 0x74, 0x88, 0x64, 0x3f, 0x69, 0x45, 0x3a, 0x6d, 0xc7,
0x7f, 0xb9, 0xa3, 0xc0, 0x6e, 0xec, 0xdc, 0xd4, 0x5a, 0xb5, 0x32, 0x85,
0x5f, 0x19, 0xd4, 0xf8, 0xd4, 0x3f, 0x3c, 0xfa, 0xc2, 0xf6, 0x5f, 0xee,
0xe6, 0xba, 0x87, 0x74, 0x2e, 0xc7, 0x0c, 0xd4, 0x42, 0xb8, 0x66, 0x85,
0x9c, 0x7b, 0x24, 0x61, 0xaa, 0x16, 0x11, 0xf6, 0xb5, 0xb6, 0xa4, 0x0a,
0xc9, 0x55, 0x2e, 0x81, 0xa5, 0x47, 0x61, 0xcb, 0x25, 0x8f, 0xc2, 0x15,
0x7b, 0x0e, 0x7c, 0x36, 0x9f, 0x3a, 0xda, 0x58, 0x86, 0x1c, 0x5b, 0x83,
0x79, 0xe6, 0x2b, 0xcc, 0xe6, 0xfa, 0x2c, 0x61, 0xf2, 0x78, 0x80, 0x1b,
0xe2, 0xf3, 0x9d, 0x39, 0x2b, 0x65, 0x57, 0x91, 0x3d, 0x71, 0x99, 0x73,
0xa5, 0xc2, 0x79, 0x20, 0x8c, 0x07, 0x4f, 0xe5, 0xb4, 0x60, 0x1f, 0x99,
0xa2, 0xb1, 0x4f, 0x0c, 0xef, 0xbc, 0x59, 0x53, 0x00, 0x7d, 0xb1, 0x02,
0x41, 0x00, 0xfc, 0x7e, 0x23, 0x65, 0x70, 0xf8, 0xce, 0xd3, 0x40, 0x41,
0x80, 0x6a, 0x1d, 0x01, 0xd6, 0x01, 0xff, 0xb6, 0x1b, 0x3d, 0x3d, 0x59,
0x09, 0x33, 0x79, 0xc0, 0x4f, 0xde, 0x96, 0x27, 0x4b, 0x18, 0xc6, 0xd9,
0x78, 0xf1, 0xf4, 0x35, 0x46, 0xe9, 0x7c, 0x42, 0x7a, 0x5d, 0x9f, 0xef,
0x54, 0xb8, 0xf7, 0x9f, 0xc4, 0x33, 0x6c, 0xf3, 0x8c, 0x32, 0x46, 0x87,
0x67, 0x30, 0x7b, 0xa7, 0xac, 0xe3, 0x02, 0x41, 0x00, 0xfc, 0x2c, 0xdf,
0x0c, 0x0d, 0x88, 0xf5, 0xb1, 0x92, 0xa8, 0x93, 0x47, 0x63, 0x55, 0xf5,
0xca, 0x58, 0x43, 0xba, 0x1c, 0xe5, 0x9e, 0xb6, 0x95, 0x05, 0xcd, 0xb5,
0x82, 0xdf, 0xeb, 0x04, 0x53, 0x9d, 0xbd, 0xc2, 0x38, 0x16, 0xb3, 0x62,
0xdd, 0xa1, 0x46, 0xdb, 0x6d, 0x97, 0x93, 0x9f, 0x8a, 0xc3, 0x9b, 0x64,
0x7e, 0x42, 0xe3, 0x32, 0x57, 0x19, 0x1b, 0xd5, 0x6e, 0x85, 0xfa, 0xb8,
0x8d, 0x02, 0x41, 0x00, 0xbc, 0x3d, 0xde, 0x6d, 0xd6, 0x97, 0xe8, 0xba,
0x9e, 0x81, 0x37, 0x17, 0xe5, 0xa0, 0x64, 0xc9, 0x00, 0xb7, 0xe7, 0xfe,
0xf4, 0x29, 0xd9, 0x2e, 0x43, 0x6b, 0x19, 0x20, 0xbd, 0x99, 0x75, 0xe7,
0x76, 0xf8, 0xd3, 0xae, 0xaf, 0x7e, 0xb8, 0xeb, 0x81, 0xf4, 0x9d, 0xfe,
0x07, 0x2b, 0x0b, 0x63, 0x0b, 0x5a, 0x55, 0x90, 0x71, 0x7d, 0xf1, 0xdb,
0xd9, 0xb1, 0x41, 0x41, 0x68, 0x2f, 0x4e, 0x39, 0x02, 0x40, 0x5a, 0x34,
0x66, 0xd8, 0xf5, 0xe2, 0x7f, 0x18, 0xb5, 0x00, 0x6e, 0x26, 0x84, 0x27,
0x14, 0x93, 0xfb, 0xfc, 0xc6, 0x0f, 0x5e, 0x27, 0xe6, 0xe1, 0xe9, 0xc0,
0x8a, 0xe4, 0x34, 0xda, 0xe9, 0xa2, 0x4b, 0x73, 0xbc, 0x8c, 0xb9, 0xba,
0x13, 0x6c, 0x7a, 0x2b, 0x51, 0x84, 0xa3, 0x4a, 0xe0, 0x30, 0x10, 0x06,
0x7e, 0xed, 0x17, 0x5a, 0x14, 0x00, 0xc9, 0xef, 0x85, 0xea, 0x52, 0x2c,
0xbc, 0x65, 0x02, 0x40, 0x51, 0xe3, 0xf2, 0x83, 0x19, 0x9b, 0xc4, 0x1e,
0x2f, 0x50, 0x3d, 0xdf, 0x5a, 0xa2, 0x18, 0xca, 0x5f, 0x2e, 0x49, 0xaf,
0x6f, 0xcc, 0xfa, 0x65, 0x77, 0x94, 0xb5, 0xa1, 0x0a, 0xa9, 0xd1, 0x8a,
0x39, 0x37, 0xf4, 0x0b, 0xa0, 0xd7, 0x82, 0x27, 0x5e, 0xae, 0x17, 0x17,
0xa1, 0x1e, 0x54, 0x34, 0xbf, 0x6e, 0xc4, 0x8e, 0x99, 0x5d, 0x08, 0xf1,
0x2d, 0x86, 0x9d, 0xa5, 0x20, 0x1b, 0xe5, 0xdf,
};
RSA decoder should check also sanity of p, q, e, d ... with respect to n This issue has been discovered by osss-fuzzer [1]. The test function decodes RSA key created by fuzzer and calls EVP_PKEY_pairwise_check() which proceeds to ossl_bn_miller_rabin_is_prime() check which takes too long exceeding timeout (45secs). The idea is to fix OSSL_DECODER_from_data() code path so invalid RSA keys will be refused. [1] https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=69134 Test case generated by the fuzzer is added. Reviewed-by: Neil Horman <nhorman@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/25190)
2024-08-15 02:07:29 +08:00
/* An invalid key whose prime factors exceed modulus n. */
static const unsigned char kInvalidRSAKeyDER[] = {
0x30, 0x80, 0x02, 0x00, 0x02, 0x02, 0xb6, 0x00, 0x02, 0x02, 0x04, 0x80,
0x02, 0x00, 0x02, 0x82, 0x08, 0x01, 0x00, 0x00, 0x00, 0x13, 0x00, 0x00,
0x00, 0x00, 0x00, 0x2a, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x02, 0x00, 0x00, 0x00, 0x87, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
0xff, 0x01, 0x04, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06,
0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06,
0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0xaa, 0xaa, 0xaa, 0xaa,
0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa,
0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0x2b,
0x31, 0xff, 0x44, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa,
0xaa, 0xaa, 0xaa, 0x29, 0xaa, 0xaa, 0xaa, 0xd9, 0xd9, 0xbf, 0x02, 0x01,
0xc8, 0x02, 0x00, 0x02, 0x02, 0x00, 0x00, 0x00, 0x88, 0x88, 0x88, 0xee,
0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07,
0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x15, 0x15, 0x15, 0x15,
0x00, 0x02, 0x00, 0x02, 0x00, 0x02, 0x00, 0x02, 0x00, 0x02, 0x15, 0x07,
0x07, 0x07, 0x07, 0x07, 0x29, 0x0f, 0x07, 0x07, 0x4d, 0x00, 0x07, 0x07,
0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07,
0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x15, 0x14, 0x15, 0x15,
0xec, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15,
0x15, 0xe6, 0x15, 0x15, 0x15, 0x15, 0xff, 0x03, 0x00, 0x0a, 0x00, 0x00,
0x00, 0x00, 0x55, 0x15, 0x15, 0x15, 0x15, 0x11, 0x05, 0x15, 0x15, 0x15,
0x07, 0x07, 0x07, 0x07, 0x07, 0x00, 0x00, 0x00, 0x4d, 0xf9, 0xf8, 0xf9,
0x02, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x0b, 0x07, 0x07, 0x07,
0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07,
0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x15, 0x15, 0x15, 0x15,
0x00, 0x02, 0x00, 0x6d, 0x61, 0x78, 0x00, 0x02, 0x00, 0x02, 0x15, 0x59,
0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59,
0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59,
0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59,
0x59, 0x59, 0x51, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0xa5, 0x59, 0x59,
0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59,
0x5d, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59,
0x59, 0x59, 0x59, 0x59, 0x59, 0x07, 0x07, 0x07, 0x07, 0x07, 0x15, 0x15,
0x15, 0x15, 0x00, 0x02, 0x01, 0x02, 0x00, 0x02, 0x00, 0x02, 0x00, 0x02,
0x00, 0x22, 0x00, 0x15, 0x15, 0x15, 0xeb, 0xea, 0xea, 0xea, 0xea, 0xea,
0xea, 0xf1, 0x15, 0x15, 0x15, 0x15, 0x15, 0x40, 0x55, 0x15, 0x15, 0x15,
0x15, 0x15, 0x05, 0x15, 0x15, 0x30, 0x00, 0x07, 0x07, 0x07, 0x07, 0x07,
0x00, 0x00, 0x00, 0x4d, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07,
0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x11, 0x07, 0x07, 0x07, 0x07,
0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07,
0x07, 0x07, 0x07, 0x15, 0x15, 0x15, 0x15, 0x00, 0x02, 0x00, 0x02, 0x00,
0x02, 0x00, 0x02, 0x00, 0x02, 0x15, 0x07, 0x07, 0x07, 0x07, 0x07, 0x29,
0x07, 0x07, 0x07, 0x4d, 0x00, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07,
0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07,
0x07, 0x07, 0x07, 0x15, 0x14, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15,
0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15,
0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x55, 0x15, 0x59,
0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59,
0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59,
0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59,
0x59, 0x59, 0x59, 0x02, 0x02, 0xb6, 0x00, 0x02, 0x02, 0x04, 0x80, 0x02,
0x00, 0x02, 0x82, 0x08, 0x01, 0x00, 0x00, 0x00, 0x13, 0x00, 0x00, 0x00,
0x00, 0x00, 0x2a, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02,
0x00, 0x00, 0x00, 0x87, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
0x01, 0x04, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06,
0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06,
0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa,
0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa,
0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0x2b, 0x31,
0xff, 0x44, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa,
0xaa, 0xaa, 0x29, 0xaa, 0xaa, 0xaa, 0xd9, 0xd9, 0xbf, 0x02, 0x01, 0xc8,
0x02, 0x00, 0x02, 0x02, 0x00, 0x00, 0x00, 0x88, 0x88, 0x88, 0xee, 0x07,
0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07,
0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x15, 0x15, 0x15, 0x15, 0x00,
0x02, 0x00, 0x02, 0x00, 0x02, 0x00, 0x02, 0x00, 0x02, 0x15, 0x07, 0x07,
0x07, 0x07, 0x07, 0x29, 0x0f, 0x07, 0x07, 0x4d, 0x00, 0x07, 0x07, 0x07,
0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07,
0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x15, 0x14, 0x15, 0x15, 0xec,
0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15,
0xe6, 0x15, 0x15, 0x15, 0x15, 0xff, 0x03, 0x00, 0x0a, 0x00, 0x00, 0x00,
0x00, 0x55, 0x15, 0x15, 0x15, 0x15, 0x11, 0x05, 0x15, 0x15, 0x15, 0x07,
0x07, 0x07, 0x07, 0x07, 0x00, 0x00, 0x00, 0x4d, 0xf9, 0xf8, 0xf9, 0x02,
0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x0b, 0x07, 0x07, 0x07, 0x07,
0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07,
0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x15, 0x15, 0x15, 0x15, 0x00,
0x02, 0x00, 0x6d, 0x61, 0x78, 0x00, 0x02, 0x00, 0x02, 0x15, 0x59, 0x59,
0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59,
0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59,
0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59,
0x59, 0x51, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0xa5, 0x59, 0x59, 0x59,
0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x5d,
0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59,
0x59, 0x59, 0x59, 0x59, 0x07, 0x07, 0x07, 0x07, 0x07, 0x15, 0x15, 0x15,
0x15, 0x00, 0x02, 0x01, 0x02, 0x00, 0x02, 0x00, 0x02, 0x00, 0x02, 0x00,
0x22, 0x00, 0x15, 0x15, 0x15, 0xeb, 0xea, 0xea, 0xea, 0xea, 0xea, 0xea,
0xf1, 0x15, 0x15, 0x15, 0x15, 0x15, 0x40, 0x55, 0x15, 0x15, 0x15, 0x15,
0x15, 0x05, 0x15, 0x15, 0x30, 0x00, 0x07, 0x07, 0x07, 0x07, 0x07, 0x00,
0x00, 0x00, 0x4d, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0xff,
0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x11, 0x07, 0x07, 0x07, 0x07, 0x07,
0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07,
0x07, 0x07, 0x15, 0x15, 0x15, 0x15, 0x00, 0x02, 0x00, 0x02, 0x00, 0x02,
0x00, 0x02, 0x00, 0x02, 0x15, 0x07, 0x07, 0x07, 0x07, 0x07, 0x29, 0x07,
0x07, 0x07, 0x4d, 0x00, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07,
0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07,
0x07, 0x07, 0x15, 0x14, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15,
0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15,
0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x55, 0x15, 0x59, 0x59,
0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59,
0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59,
0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59,
0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59,
0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59,
0x59, 0x59, 0x59, 0x06, 0xce, 0x15, 0x00, 0xfe, 0xf7, 0x52, 0x53, 0x41,
0x31, 0x01, 0x00, 0x00, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
0xff, 0xff, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa,
0x2b, 0x31, 0xff, 0x44, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa,
0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xd9, 0xd9, 0xbf, 0x02,
0x01, 0xc8, 0x02, 0x00, 0x02, 0x02, 0x00, 0x00, 0x00, 0x88, 0x88, 0x88,
0xee, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07,
0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x15, 0x15, 0x15,
0x15, 0x00, 0x02, 0x00, 0x02, 0x00, 0x02, 0x00, 0x02, 0x00, 0x02, 0x15,
0x07, 0x07, 0x07, 0x07, 0x07, 0x29, 0x07, 0x07, 0x07, 0x4d, 0x00, 0x07,
0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07,
0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x15, 0x14, 0x15,
0x15, 0xec, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15,
0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15,
0x15, 0x15, 0x15, 0x55, 0x15, 0x15, 0x15, 0x15, 0x15, 0x05, 0x15, 0x15,
0x15, 0x07, 0x07, 0x07, 0x07, 0x07, 0x00, 0x00, 0x00, 0x4d, 0x07, 0x07,
0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x0b, 0x07, 0x07,
0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07,
0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x15, 0x15, 0x15,
0x15, 0x00, 0x02, 0x00, 0x02, 0x00, 0x02, 0x00, 0x02, 0x00, 0x02, 0x15,
0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59,
0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x59, 0x8f,
0x8f, 0x8f, 0x8f, 0x8f, 0x8f, 0x8f, 0x8f, 0x8f, 0x8f, 0x8f, 0x8f, 0x8f,
0x8f, 0x8f, 0x8f, 0x8f, 0x8f, 0x8f, 0x8f, 0x8f, 0x8f, 0x8f, 0x8f, 0x8f,
0x8f, 0x8f, 0x8f, 0x8f, 0x8f, 0x8f, 0x8f, 0x8f, 0x8f, 0x8f, 0x8f, 0x8f,
0x8f, 0x8f, 0x8f, 0x8f, 0x59, 0x59, 0x59, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xc1, 0x3d, 0xc1, 0xc1,
0xc1, 0xc1, 0xc1, 0xc1, 0xc1, 0xc1, 0xc1, 0xc1, 0xc1, 0xc1, 0xc1, 0xc1,
0xc1, 0xc1, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x80, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x42, 0xa5, 0x02, 0x02, 0x42, 0x02,
0x02, 0x51, 0x01, 0x02, 0x02, 0xd2, 0x42, 0x02, 0xe8, 0xe8, 0xe8, 0xe8,
0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0xe8,
0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0xe8,
0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0xe8,
0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0xe8,
0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0xe8,
0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0xe8, 0x02,
0x02, 0x42, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02,
0x21, 0x2b, 0x02, 0x02, 0x02, 0x02, 0x02, 0x21, 0x02, 0x02, 0x32, 0x80,
0x02, 0x02, 0x7f, 0x1b, 0x02, 0x00, 0x1f, 0x04, 0xff, 0x80, 0x02, 0x02,
0x02, 0x02, 0x42, 0x02, 0x12, 0x02, 0x42, 0x02, 0x79, 0x70, 0x65, 0x36,
0x28, 0xc8, 0x02, 0x01, 0x81, 0x08, 0xfe, 0x00, 0xf9, 0x02, 0x42, 0x10,
0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10,
0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10,
0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10,
0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10,
0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0xa5, 0x02,
0x02, 0x42, 0x02, 0x02, 0x51, 0x01, 0x02, 0x02, 0xd2, 0x42, 0x02, 0x02,
0x02, 0x42, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02,
0x21, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x32, 0x80, 0x02, 0x02,
0x7f, 0x1b, 0x02, 0x00, 0x1f, 0x04, 0xff, 0x80, 0x02, 0x02, 0x02, 0x02,
0x42, 0x02, 0x12, 0x02, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x47,
0x4f, 0x53, 0x54, 0x20, 0x52, 0x20, 0x33, 0x34, 0x2e, 0x31, 0x31, 0x2d,
0x32, 0x30, 0x31, 0x32, 0x20, 0x77, 0x69, 0x74, 0x68, 0x30, 0x80, 0x60,
0x02, 0x82, 0x24, 0x02, 0x02, 0x41, 0x52, 0x49, 0x41, 0x2d, 0x31, 0x32,
0x38, 0x2d, 0x43, 0x46, 0x42, 0x38, 0xff, 0xff, 0xff, 0x3a, 0x5b, 0xff,
0xff, 0x7f, 0x49, 0x74, 0x84, 0x00, 0x00, 0x70, 0x65, 0x00, 0x00, 0x30,
0x80, 0x60, 0x02, 0x82, 0x24, 0x02, 0x02, 0x41, 0x52, 0x49, 0x41, 0x2d,
0x31, 0x32, 0x38, 0x2d, 0x43, 0x46, 0x42, 0x38, 0xff, 0xff, 0xff, 0x3a,
0x5b, 0xff, 0xff, 0x7f, 0x49, 0x74, 0x84, 0x00, 0x00, 0x70, 0x65, 0x33,
0x28, 0xc8, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15,
0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x55, 0x15, 0x15, 0x15, 0x15,
0x15, 0x05, 0x15, 0x95, 0x15, 0x07, 0x07, 0x07, 0x07, 0x07, 0x00, 0x00,
0x00, 0x4d, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07,
0x07, 0x0b, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07,
0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07,
0x07, 0x15, 0x15, 0x15, 0x15, 0x00, 0x02, 0x00, 0x02, 0x00, 0x02, 0x00,
0x02, 0x00, 0x25, 0x02, 0x02, 0x22, 0x3a, 0x02, 0x02, 0x02, 0x42, 0x02,
0x02, 0x02, 0x02, 0x02, 0x02, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00,
0xe2, 0x80,
};
Add a test for EVP_Digest[Sign|Verify]* with a BIO MD If an EVP_MD_CTX is initialised with EVP_DigestSignInit_ex() or EVP_DigestVerifyInit_ex() it should work with an MD BIO to provide the "Updates". Test that this is the case. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10116)
2019-10-08 01:21:39 +08:00
/*
RSA decoder should check also sanity of p, q, e, d ... with respect to n This issue has been discovered by osss-fuzzer [1]. The test function decodes RSA key created by fuzzer and calls EVP_PKEY_pairwise_check() which proceeds to ossl_bn_miller_rabin_is_prime() check which takes too long exceeding timeout (45secs). The idea is to fix OSSL_DECODER_from_data() code path so invalid RSA keys will be refused. [1] https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=69134 Test case generated by the fuzzer is added. Reviewed-by: Neil Horman <nhorman@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/25190)
2024-08-15 02:07:29 +08:00
* kExampleDSAKeyDER is a DSA private key in ASN.1, DER format. Of course, you
Add a test for EVP_Digest[Sign|Verify]* with a BIO MD If an EVP_MD_CTX is initialised with EVP_DigestSignInit_ex() or EVP_DigestVerifyInit_ex() it should work with an MD BIO to provide the "Updates". Test that this is the case. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10116)
2019-10-08 01:21:39 +08:00
* should never use this key anywhere but in an example.
*/
Fix no-dsa Make sure we don't try and load a DSA key in the tests if DSA has been disabled. Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10375)
2019-11-07 17:24:18 +08:00
#ifndef OPENSSL_NO_DSA
Add a test for EVP_Digest[Sign|Verify]* with a BIO MD If an EVP_MD_CTX is initialised with EVP_DigestSignInit_ex() or EVP_DigestVerifyInit_ex() it should work with an MD BIO to provide the "Updates". Test that this is the case. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10116)
2019-10-08 01:21:39 +08:00
static const unsigned char kExampleDSAKeyDER[] = {
0x30, 0x82, 0x01, 0xba, 0x02, 0x01, 0x00, 0x02, 0x81, 0x81, 0x00, 0x9a,
0x05, 0x6d, 0x33, 0xcd, 0x5d, 0x78, 0xa1, 0xbb, 0xcb, 0x7d, 0x5b, 0x8d,
0xb4, 0xcc, 0xbf, 0x03, 0x99, 0x64, 0xde, 0x38, 0x78, 0x06, 0x15, 0x2f,
0x86, 0x26, 0x77, 0xf3, 0xb1, 0x85, 0x00, 0xed, 0xfc, 0x28, 0x3a, 0x42,
0x4d, 0xab, 0xab, 0xdf, 0xbc, 0x9c, 0x16, 0xd0, 0x22, 0x50, 0xd1, 0x38,
0xdd, 0x3f, 0x64, 0x05, 0x9e, 0x68, 0x7a, 0x1e, 0xf1, 0x56, 0xbf, 0x1e,
0x2c, 0xc5, 0x97, 0x2a, 0xfe, 0x7a, 0x22, 0xdc, 0x6c, 0x68, 0xb8, 0x2e,
0x06, 0xdb, 0x41, 0xca, 0x98, 0xd8, 0x54, 0xc7, 0x64, 0x48, 0x24, 0x04,
0x20, 0xbc, 0x59, 0xe3, 0x6b, 0xea, 0x7e, 0xfc, 0x7e, 0xc5, 0x4e, 0xd4,
0xd8, 0x3a, 0xed, 0xcd, 0x5d, 0x99, 0xb8, 0x5c, 0xa2, 0x8b, 0xbb, 0x0b,
0xac, 0xe6, 0x8e, 0x25, 0x56, 0x22, 0x3a, 0x2d, 0x3a, 0x56, 0x41, 0x14,
0x1f, 0x1c, 0x8f, 0x53, 0x46, 0x13, 0x85, 0x02, 0x15, 0x00, 0x98, 0x7e,
0x92, 0x81, 0x88, 0xc7, 0x3f, 0x70, 0x49, 0x54, 0xf6, 0x76, 0xb4, 0xa3,
0x9e, 0x1d, 0x45, 0x98, 0x32, 0x7f, 0x02, 0x81, 0x80, 0x69, 0x4d, 0xef,
0x55, 0xff, 0x4d, 0x59, 0x2c, 0x01, 0xfa, 0x6a, 0x38, 0xe0, 0x70, 0x9f,
0x9e, 0x66, 0x8e, 0x3e, 0x8c, 0x52, 0x22, 0x9d, 0x15, 0x7e, 0x3c, 0xef,
0x4c, 0x7a, 0x61, 0x26, 0xe0, 0x2b, 0x81, 0x3f, 0xeb, 0xaf, 0x35, 0x38,
0x8d, 0xfe, 0xed, 0x46, 0xff, 0x5f, 0x03, 0x9b, 0x81, 0x92, 0xe7, 0x6f,
0x76, 0x4f, 0x1d, 0xd9, 0xbb, 0x89, 0xc9, 0x3e, 0xd9, 0x0b, 0xf9, 0xf4,
0x78, 0x11, 0x59, 0xc0, 0x1d, 0xcd, 0x0e, 0xa1, 0x6f, 0x15, 0xf1, 0x4d,
0xc1, 0xc9, 0x22, 0xed, 0x8d, 0xad, 0x67, 0xc5, 0x4b, 0x95, 0x93, 0x86,
0xa6, 0xaf, 0x8a, 0xee, 0x06, 0x89, 0x2f, 0x37, 0x7e, 0x64, 0xaa, 0xf6,
0xe7, 0xb1, 0x5a, 0x0a, 0x93, 0x95, 0x5d, 0x3e, 0x53, 0x9a, 0xde, 0x8a,
0xc2, 0x95, 0x45, 0x81, 0xbe, 0x5c, 0x2f, 0xc2, 0xb2, 0x92, 0x58, 0x19,
0x72, 0x80, 0xe9, 0x79, 0xa1, 0x02, 0x81, 0x80, 0x07, 0xd7, 0x62, 0xff,
0xdf, 0x1a, 0x3f, 0xed, 0x32, 0xd4, 0xd4, 0x88, 0x7b, 0x2c, 0x63, 0x7f,
0x97, 0xdc, 0x44, 0xd4, 0x84, 0xa2, 0xdd, 0x17, 0x16, 0x85, 0x13, 0xe0,
0xac, 0x51, 0x8d, 0x29, 0x1b, 0x75, 0x9a, 0xe4, 0xe3, 0x8a, 0x92, 0x69,
0x09, 0x03, 0xc5, 0x68, 0xae, 0x5e, 0x94, 0xfe, 0xc9, 0x92, 0x6c, 0x07,
0xb4, 0x1e, 0x64, 0x62, 0x87, 0xc6, 0xa4, 0xfd, 0x0d, 0x5f, 0xe5, 0xf9,
0x1b, 0x4f, 0x85, 0x5f, 0xae, 0xf3, 0x11, 0xe5, 0x18, 0xd4, 0x4d, 0x79,
0x9f, 0xc4, 0x79, 0x26, 0x04, 0x27, 0xf0, 0x0b, 0xee, 0x2b, 0x86, 0x9f,
0x86, 0x61, 0xe6, 0x51, 0xce, 0x04, 0x9b, 0x5d, 0x6b, 0x34, 0x43, 0x8c,
0x85, 0x3c, 0xf1, 0x51, 0x9b, 0x08, 0x23, 0x1b, 0xf5, 0x7e, 0x33, 0x12,
0xea, 0xab, 0x1f, 0xb7, 0x2d, 0xe2, 0x5f, 0xe6, 0x97, 0x99, 0xb5, 0x45,
0x16, 0x5b, 0xc3, 0x41, 0x02, 0x14, 0x61, 0xbf, 0x51, 0x60, 0xcf, 0xc8,
0xf1, 0x8c, 0x82, 0x97, 0xf2, 0xf4, 0x19, 0xba, 0x2b, 0xf3, 0x16, 0xbe,
0x40, 0x48
};
Fix no-dsa Make sure we don't try and load a DSA key in the tests if DSA has been disabled. Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10375)
2019-11-07 17:24:18 +08:00
#endif
Add a test for EVP_Digest[Sign|Verify]* with a BIO MD If an EVP_MD_CTX is initialised with EVP_DigestSignInit_ex() or EVP_DigestVerifyInit_ex() it should work with an MD BIO to provide the "Updates". Test that this is the case. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10116)
2019-10-08 01:21:39 +08:00
Support key check in EVP interface A new method is added to EVP_PKEY_METH as: int (*check) (EVP_PKEY_CTX *ctx); and to EVP_PKEY_ASN1_METHOD as: int (*pkey_check) (EVP_PKEY_CTX *ctx); This is used to check the validity of a specific key. The order of calls is: EVP_PKEY_check -> pmeth.check -> ameth.pkey_check. Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/4337)
2017-09-04 22:02:59 +08:00
/*
* kExampleBadRSAKeyDER is an RSA private key in ASN.1, DER format. The private
* components are not correct.
*/
static const unsigned char kExampleBadRSAKeyDER[] = {
0x30, 0x82, 0x04, 0x27, 0x02, 0x01, 0x00, 0x02, 0x82, 0x01, 0x01, 0x00,
0xa6, 0x1a, 0x1e, 0x6e, 0x7b, 0xee, 0xc6, 0x89, 0x66, 0xe7, 0x93, 0xef,
0x54, 0x12, 0x68, 0xea, 0xbf, 0x86, 0x2f, 0xdd, 0xd2, 0x79, 0xb8, 0xa9,
0x6e, 0x03, 0xc2, 0xa3, 0xb9, 0xa3, 0xe1, 0x4b, 0x2a, 0xb3, 0xf8, 0xb4,
0xcd, 0xea, 0xbe, 0x24, 0xa6, 0x57, 0x5b, 0x83, 0x1f, 0x0f, 0xf2, 0xd3,
0xb7, 0xac, 0x7e, 0xd6, 0x8e, 0x6e, 0x1e, 0xbf, 0xb8, 0x73, 0x8c, 0x05,
0x56, 0xe6, 0x35, 0x1f, 0xe9, 0x04, 0x0b, 0x09, 0x86, 0x7d, 0xf1, 0x26,
0x08, 0x99, 0xad, 0x7b, 0xc8, 0x4d, 0x94, 0xb0, 0x0b, 0x8b, 0x38, 0xa0,
0x5c, 0x62, 0xa0, 0xab, 0xd3, 0x8f, 0xd4, 0x09, 0x60, 0x72, 0x1e, 0x33,
0x50, 0x80, 0x6e, 0x22, 0xa6, 0x77, 0x57, 0x6b, 0x9a, 0x33, 0x21, 0x66,
0x87, 0x6e, 0x21, 0x7b, 0xc7, 0x24, 0x0e, 0xd8, 0x13, 0xdf, 0x83, 0xde,
0xcd, 0x40, 0x58, 0x1d, 0x84, 0x86, 0xeb, 0xb8, 0x12, 0x4e, 0xd2, 0xfa,
0x80, 0x1f, 0xe4, 0xe7, 0x96, 0x29, 0xb8, 0xcc, 0xce, 0x66, 0x6d, 0x53,
0xca, 0xb9, 0x5a, 0xd7, 0xf6, 0x84, 0x6c, 0x2d, 0x9a, 0x1a, 0x14, 0x1c,
0x4e, 0x93, 0x39, 0xba, 0x74, 0xed, 0xed, 0x87, 0x87, 0x5e, 0x48, 0x75,
0x36, 0xf0, 0xbc, 0x34, 0xfb, 0x29, 0xf9, 0x9f, 0x96, 0x5b, 0x0b, 0xa7,
0x54, 0x30, 0x51, 0x29, 0x18, 0x5b, 0x7d, 0xac, 0x0f, 0xd6, 0x5f, 0x7c,
0xf8, 0x98, 0x8c, 0xd8, 0x86, 0x62, 0xb3, 0xdc, 0xff, 0x0f, 0xff, 0x7a,
0xaf, 0x5c, 0x4c, 0x61, 0x49, 0x2e, 0xc8, 0x95, 0x86, 0xc4, 0x0e, 0x87,
0xfc, 0x1d, 0xcf, 0x8b, 0x7c, 0x61, 0xf6, 0xd8, 0xd0, 0x69, 0xf6, 0xcd,
0x8a, 0x8c, 0xf6, 0x62, 0xa2, 0x56, 0xa9, 0xe3, 0xd1, 0xcf, 0x4d, 0xa0,
0xf6, 0x2d, 0x20, 0x0a, 0x04, 0xb7, 0xa2, 0xf7, 0xb5, 0x99, 0x47, 0x18,
0x56, 0x85, 0x87, 0xc7, 0x02, 0x03, 0x01, 0x00, 0x01, 0x02, 0x82, 0x01,
0x01, 0x00, 0x99, 0x41, 0x38, 0x1a, 0xd0, 0x96, 0x7a, 0xf0, 0x83, 0xd5,
0xdf, 0x94, 0xce, 0x89, 0x3d, 0xec, 0x7a, 0x52, 0x21, 0x10, 0x16, 0x06,
0xe0, 0xee, 0xd2, 0xe6, 0xfd, 0x4b, 0x7b, 0x19, 0x4d, 0xe1, 0xc0, 0xc0,
0xd5, 0x14, 0x5d, 0x79, 0xdd, 0x7e, 0x8b, 0x4b, 0xc6, 0xcf, 0xb0, 0x75,
0x52, 0xa3, 0x2d, 0xb1, 0x26, 0x46, 0x68, 0x9c, 0x0a, 0x1a, 0xf2, 0xe1,
0x09, 0xac, 0x53, 0x85, 0x8c, 0x36, 0xa9, 0x14, 0x65, 0xea, 0xa0, 0x00,
0xcb, 0xe3, 0x3f, 0xc4, 0x2b, 0x61, 0x2e, 0x6b, 0x06, 0x69, 0x77, 0xfd,
0x38, 0x7e, 0x1d, 0x3f, 0x92, 0xe7, 0x77, 0x08, 0x19, 0xa7, 0x9d, 0x29,
0x2d, 0xdc, 0x42, 0xc6, 0x7c, 0xd7, 0xd3, 0xa8, 0x01, 0x2c, 0xf2, 0xd5,
0x82, 0x57, 0xcb, 0x55, 0x3d, 0xe7, 0xaa, 0xd2, 0x06, 0x30, 0x30, 0x05,
0xe6, 0xf2, 0x47, 0x86, 0xba, 0xc6, 0x61, 0x64, 0xeb, 0x4f, 0x2a, 0x5e,
0x07, 0x29, 0xe0, 0x96, 0xb2, 0x43, 0xff, 0x5f, 0x1a, 0x54, 0x16, 0xcf,
0xb5, 0x56, 0x5c, 0xa0, 0x9b, 0x0c, 0xfd, 0xb3, 0xd2, 0xe3, 0x79, 0x1d,
0x21, 0xe2, 0xd6, 0x13, 0xc4, 0x74, 0xa6, 0xf5, 0x8e, 0x8e, 0x81, 0xbb,
0xb4, 0xad, 0x8a, 0xf0, 0x93, 0x0a, 0xd8, 0x0a, 0x42, 0x36, 0xbc, 0xe5,
0x26, 0x2a, 0x0d, 0x5d, 0x57, 0x13, 0xc5, 0x4e, 0x2f, 0x12, 0x0e, 0xef,
0xa7, 0x81, 0x1e, 0xc3, 0xa5, 0xdb, 0xc9, 0x24, 0xeb, 0x1a, 0xa1, 0xf9,
0xf6, 0xa1, 0x78, 0x98, 0x93, 0x77, 0x42, 0x45, 0x03, 0xe2, 0xc9, 0xa2,
0xfe, 0x2d, 0x77, 0xc8, 0xc6, 0xac, 0x9b, 0x98, 0x89, 0x6d, 0x9a, 0xe7,
0x61, 0x63, 0xb7, 0xf2, 0xec, 0xd6, 0xb1, 0xa1, 0x6e, 0x0a, 0x1a, 0xff,
0xfd, 0x43, 0x28, 0xc3, 0x0c, 0xdc, 0xf2, 0x47, 0x4f, 0x27, 0xaa, 0x99,
0x04, 0x8e, 0xac, 0xe8, 0x7c, 0x01, 0x02, 0x04, 0x12, 0x34, 0x56, 0x78,
0x02, 0x81, 0x81, 0x00, 0xca, 0x69, 0xe5, 0xbb, 0x3a, 0x90, 0x82, 0xcb,
0x82, 0x50, 0x2f, 0x29, 0xe2, 0x76, 0x6a, 0x57, 0x55, 0x45, 0x4e, 0x35,
0x18, 0x61, 0xe0, 0x12, 0x70, 0xc0, 0xab, 0xc7, 0x80, 0xa2, 0xd4, 0x46,
0x34, 0x03, 0xa0, 0x19, 0x26, 0x23, 0x9e, 0xef, 0x1a, 0xcb, 0x75, 0xd6,
0xba, 0x81, 0xf4, 0x7e, 0x52, 0xe5, 0x2a, 0xe8, 0xf1, 0x49, 0x6c, 0x0f,
0x1a, 0xa0, 0xf9, 0xc6, 0xe7, 0xec, 0x60, 0xe4, 0xcb, 0x2a, 0xb5, 0x56,
0xe9, 0x9c, 0xcd, 0x19, 0x75, 0x92, 0xb1, 0x66, 0xce, 0xc3, 0xd9, 0x3d,
0x11, 0xcb, 0xc4, 0x09, 0xce, 0x1e, 0x30, 0xba, 0x2f, 0x60, 0x60, 0x55,
0x8d, 0x02, 0xdc, 0x5d, 0xaf, 0xf7, 0x52, 0x31, 0x17, 0x07, 0x53, 0x20,
0x33, 0xad, 0x8c, 0xd5, 0x2f, 0x5a, 0xd0, 0x57, 0xd7, 0xd1, 0x80, 0xd6,
0x3a, 0x9b, 0x04, 0x4f, 0x35, 0xbf, 0xe7, 0xd5, 0xbc, 0x8f, 0xd4, 0x81,
0x02, 0x81, 0x81, 0x00, 0xc0, 0x9f, 0xf8, 0xcd, 0xf7, 0x3f, 0x26, 0x8a,
0x3d, 0x4d, 0x2b, 0x0c, 0x01, 0xd0, 0xa2, 0xb4, 0x18, 0xfe, 0xf7, 0x5e,
0x2f, 0x06, 0x13, 0xcd, 0x63, 0xaa, 0x12, 0xa9, 0x24, 0x86, 0xe3, 0xf3,
0x7b, 0xda, 0x1a, 0x3c, 0xb1, 0x38, 0x80, 0x80, 0xef, 0x64, 0x64, 0xa1,
0x9b, 0xfe, 0x76, 0x63, 0x8e, 0x83, 0xd2, 0xd9, 0xb9, 0x86, 0xb0, 0xe6,
0xa6, 0x0c, 0x7e, 0xa8, 0x84, 0x90, 0x98, 0x0c, 0x1e, 0xf3, 0x14, 0x77,
0xe0, 0x5f, 0x81, 0x08, 0x11, 0x8f, 0xa6, 0x23, 0xc4, 0xba, 0xc0, 0x8a,
0xe4, 0xc6, 0xe3, 0x5c, 0xbe, 0xc5, 0xec, 0x2c, 0xb9, 0xd8, 0x8c, 0x4d,
0x1a, 0x9d, 0xe7, 0x7c, 0x85, 0x4c, 0x0d, 0x71, 0x4e, 0x72, 0x33, 0x1b,
0xfe, 0xa9, 0x17, 0x72, 0x76, 0x56, 0x9d, 0x74, 0x7e, 0x52, 0x67, 0x9a,
0x87, 0x9a, 0xdb, 0x30, 0xde, 0xe4, 0x49, 0x28, 0x3b, 0xd2, 0x67, 0xaf,
0x02, 0x81, 0x81, 0x00, 0x89, 0x74, 0x9a, 0x8e, 0xa7, 0xb9, 0xa5, 0x28,
0xc0, 0x68, 0xe5, 0x6e, 0x63, 0x1c, 0x99, 0x20, 0x8f, 0x86, 0x8e, 0x12,
0x9e, 0x69, 0x30, 0xfa, 0x34, 0xd9, 0x92, 0x8d, 0xdb, 0x7c, 0x37, 0xfd,
0x28, 0xab, 0x61, 0x98, 0x52, 0x7f, 0x14, 0x1a, 0x39, 0xae, 0xfb, 0x6a,
0x03, 0xa3, 0xe6, 0xbd, 0xb6, 0x5b, 0x6b, 0xe5, 0x5e, 0x9d, 0xc6, 0xa5,
0x07, 0x27, 0x54, 0x17, 0xd0, 0x3d, 0x84, 0x9b, 0x3a, 0xa0, 0xd9, 0x1e,
0x99, 0x6c, 0x63, 0x17, 0xab, 0xf1, 0x1f, 0x49, 0xba, 0x95, 0xe3, 0x3b,
0x86, 0x8f, 0x42, 0xa4, 0x89, 0xf5, 0x94, 0x8f, 0x8b, 0x46, 0xbe, 0x84,
0xba, 0x4a, 0xbc, 0x0d, 0x5f, 0x46, 0xeb, 0xe8, 0xec, 0x43, 0x8c, 0x1e,
0xad, 0x19, 0x69, 0x2f, 0x08, 0x86, 0x7a, 0x3f, 0x7d, 0x0f, 0x07, 0x97,
0xf3, 0x9a, 0x7b, 0xb5, 0xb2, 0xc1, 0x8c, 0x95, 0x68, 0x04, 0xa0, 0x81,
0x02, 0x81, 0x80, 0x4e, 0xbf, 0x7e, 0x1b, 0xcb, 0x13, 0x61, 0x75, 0x3b,
0xdb, 0x59, 0x5f, 0xb1, 0xd4, 0xb8, 0xeb, 0x9e, 0x73, 0xb5, 0xe7, 0xf6,
0x89, 0x3d, 0x1c, 0xda, 0xf0, 0x36, 0xff, 0x35, 0xbd, 0x1e, 0x0b, 0x74,
0xe3, 0x9e, 0xf0, 0xf2, 0xf7, 0xd7, 0x82, 0xb7, 0x7b, 0x6a, 0x1b, 0x0e,
0x30, 0x4a, 0x98, 0x0e, 0xb4, 0xf9, 0x81, 0x07, 0xe4, 0x75, 0x39, 0xe9,
0x53, 0xca, 0xbb, 0x5c, 0xaa, 0x93, 0x07, 0x0e, 0xa8, 0x2f, 0xba, 0x98,
0x49, 0x30, 0xa7, 0xcc, 0x1a, 0x3c, 0x68, 0x0c, 0xe1, 0xa4, 0xb1, 0x05,
0xe6, 0xe0, 0x25, 0x78, 0x58, 0x14, 0x37, 0xf5, 0x1f, 0xe3, 0x22, 0xef,
0xa8, 0x0e, 0x22, 0xa0, 0x94, 0x3a, 0xf6, 0xc9, 0x13, 0xe6, 0x06, 0xbf,
0x7f, 0x99, 0xc6, 0xcc, 0xd8, 0xc6, 0xbe, 0xd9, 0x2e, 0x24, 0xc7, 0x69,
0x8c, 0x95, 0xba, 0xf6, 0x04, 0xb3, 0x0a, 0xf4, 0xcb, 0xf0, 0xce,
};
Add a test for an all 0 RSA key Testing a key with all values set to 0. Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/18651)
2022-06-24 18:02:22 +08:00
/*
* kExampleBad2RSAKeyDER is an RSA private key in ASN.1, DER format. All
* values are 0.
*/
static const unsigned char kExampleBad2RSAKeyDER[] = {
0x30, 0x1b, 0x02, 0x01, 0x00, 0x02, 0x01, 0x00, 0x02, 0x01, 0x00, 0x02,
0x01, 0x00, 0x02, 0x01, 0x00, 0x02, 0x01, 0x00, 0x02, 0x01, 0x00, 0x02,
0x01, 0x00, 0x02, 0x01, 0x00
};
Import evp_test.c from BoringSSL. Unfortunately we already have a file called evp_test.c, so I have called this one evp_extra_test.c Reviewed-by: Emilia Käsper <emilia@openssl.org>
2015-02-09 17:45:35 +08:00
static const unsigned char kMsg[] = { 1, 2, 3, 4 };
static const unsigned char kSignature[] = {
0xa5, 0xf0, 0x8a, 0x47, 0x5d, 0x3c, 0xb3, 0xcc, 0xa9, 0x79, 0xaf, 0x4d,
0x8c, 0xae, 0x4c, 0x14, 0xef, 0xc2, 0x0b, 0x34, 0x36, 0xde, 0xf4, 0x3e,
0x3d, 0xbb, 0x4a, 0x60, 0x5c, 0xc8, 0x91, 0x28, 0xda, 0xfb, 0x7e, 0x04,
0x96, 0x7e, 0x63, 0x13, 0x90, 0xce, 0xb9, 0xb4, 0x62, 0x7a, 0xfd, 0x09,
0x3d, 0xc7, 0x67, 0x78, 0x54, 0x04, 0xeb, 0x52, 0x62, 0x6e, 0x24, 0x67,
0xb4, 0x40, 0xfc, 0x57, 0x62, 0xc6, 0xf1, 0x67, 0xc1, 0x97, 0x8f, 0x6a,
0xa8, 0xae, 0x44, 0x46, 0x5e, 0xab, 0x67, 0x17, 0x53, 0x19, 0x3a, 0xda,
0x5a, 0xc8, 0x16, 0x3e, 0x86, 0xd5, 0xc5, 0x71, 0x2f, 0xfc, 0x23, 0x48,
0xd9, 0x0b, 0x13, 0xdd, 0x7b, 0x5a, 0x25, 0x79, 0xef, 0xa5, 0x7b, 0x04,
0xed, 0x44, 0xf6, 0x18, 0x55, 0xe4, 0x0a, 0xe9, 0x57, 0x79, 0x5d, 0xd7,
0x55, 0xa7, 0xab, 0x45, 0x02, 0x97, 0x60, 0x42,
};
/*
* kExampleRSAKeyPKCS8 is kExampleRSAKeyDER encoded in a PKCS #8
* PrivateKeyInfo.
*/
static const unsigned char kExampleRSAKeyPKCS8[] = {
0x30, 0x82, 0x02, 0x76, 0x02, 0x01, 0x00, 0x30, 0x0d, 0x06, 0x09, 0x2a,
0x86, 0x48, 0x86, 0xf7, 0x0d, 0x01, 0x01, 0x01, 0x05, 0x00, 0x04, 0x82,
0x02, 0x60, 0x30, 0x82, 0x02, 0x5c, 0x02, 0x01, 0x00, 0x02, 0x81, 0x81,
0x00, 0xf8, 0xb8, 0x6c, 0x83, 0xb4, 0xbc, 0xd9, 0xa8, 0x57, 0xc0, 0xa5,
0xb4, 0x59, 0x76, 0x8c, 0x54, 0x1d, 0x79, 0xeb, 0x22, 0x52, 0x04, 0x7e,
0xd3, 0x37, 0xeb, 0x41, 0xfd, 0x83, 0xf9, 0xf0, 0xa6, 0x85, 0x15, 0x34,
0x75, 0x71, 0x5a, 0x84, 0xa8, 0x3c, 0xd2, 0xef, 0x5a, 0x4e, 0xd3, 0xde,
0x97, 0x8a, 0xdd, 0xff, 0xbb, 0xcf, 0x0a, 0xaa, 0x86, 0x92, 0xbe, 0xb8,
0x50, 0xe4, 0xcd, 0x6f, 0x80, 0x33, 0x30, 0x76, 0x13, 0x8f, 0xca, 0x7b,
0xdc, 0xec, 0x5a, 0xca, 0x63, 0xc7, 0x03, 0x25, 0xef, 0xa8, 0x8a, 0x83,
0x58, 0x76, 0x20, 0xfa, 0x16, 0x77, 0xd7, 0x79, 0x92, 0x63, 0x01, 0x48,
0x1a, 0xd8, 0x7b, 0x67, 0xf1, 0x52, 0x55, 0x49, 0x4e, 0xd6, 0x6e, 0x4a,
0x5c, 0xd7, 0x7a, 0x37, 0x36, 0x0c, 0xde, 0xdd, 0x8f, 0x44, 0xe8, 0xc2,
0xa7, 0x2c, 0x2b, 0xb5, 0xaf, 0x64, 0x4b, 0x61, 0x07, 0x02, 0x03, 0x01,
0x00, 0x01, 0x02, 0x81, 0x80, 0x74, 0x88, 0x64, 0x3f, 0x69, 0x45, 0x3a,
0x6d, 0xc7, 0x7f, 0xb9, 0xa3, 0xc0, 0x6e, 0xec, 0xdc, 0xd4, 0x5a, 0xb5,
0x32, 0x85, 0x5f, 0x19, 0xd4, 0xf8, 0xd4, 0x3f, 0x3c, 0xfa, 0xc2, 0xf6,
0x5f, 0xee, 0xe6, 0xba, 0x87, 0x74, 0x2e, 0xc7, 0x0c, 0xd4, 0x42, 0xb8,
0x66, 0x85, 0x9c, 0x7b, 0x24, 0x61, 0xaa, 0x16, 0x11, 0xf6, 0xb5, 0xb6,
0xa4, 0x0a, 0xc9, 0x55, 0x2e, 0x81, 0xa5, 0x47, 0x61, 0xcb, 0x25, 0x8f,
0xc2, 0x15, 0x7b, 0x0e, 0x7c, 0x36, 0x9f, 0x3a, 0xda, 0x58, 0x86, 0x1c,
0x5b, 0x83, 0x79, 0xe6, 0x2b, 0xcc, 0xe6, 0xfa, 0x2c, 0x61, 0xf2, 0x78,
0x80, 0x1b, 0xe2, 0xf3, 0x9d, 0x39, 0x2b, 0x65, 0x57, 0x91, 0x3d, 0x71,
0x99, 0x73, 0xa5, 0xc2, 0x79, 0x20, 0x8c, 0x07, 0x4f, 0xe5, 0xb4, 0x60,
0x1f, 0x99, 0xa2, 0xb1, 0x4f, 0x0c, 0xef, 0xbc, 0x59, 0x53, 0x00, 0x7d,
0xb1, 0x02, 0x41, 0x00, 0xfc, 0x7e, 0x23, 0x65, 0x70, 0xf8, 0xce, 0xd3,
0x40, 0x41, 0x80, 0x6a, 0x1d, 0x01, 0xd6, 0x01, 0xff, 0xb6, 0x1b, 0x3d,
0x3d, 0x59, 0x09, 0x33, 0x79, 0xc0, 0x4f, 0xde, 0x96, 0x27, 0x4b, 0x18,
0xc6, 0xd9, 0x78, 0xf1, 0xf4, 0x35, 0x46, 0xe9, 0x7c, 0x42, 0x7a, 0x5d,
0x9f, 0xef, 0x54, 0xb8, 0xf7, 0x9f, 0xc4, 0x33, 0x6c, 0xf3, 0x8c, 0x32,
0x46, 0x87, 0x67, 0x30, 0x7b, 0xa7, 0xac, 0xe3, 0x02, 0x41, 0x00, 0xfc,
0x2c, 0xdf, 0x0c, 0x0d, 0x88, 0xf5, 0xb1, 0x92, 0xa8, 0x93, 0x47, 0x63,
0x55, 0xf5, 0xca, 0x58, 0x43, 0xba, 0x1c, 0xe5, 0x9e, 0xb6, 0x95, 0x05,
0xcd, 0xb5, 0x82, 0xdf, 0xeb, 0x04, 0x53, 0x9d, 0xbd, 0xc2, 0x38, 0x16,
0xb3, 0x62, 0xdd, 0xa1, 0x46, 0xdb, 0x6d, 0x97, 0x93, 0x9f, 0x8a, 0xc3,
0x9b, 0x64, 0x7e, 0x42, 0xe3, 0x32, 0x57, 0x19, 0x1b, 0xd5, 0x6e, 0x85,
0xfa, 0xb8, 0x8d, 0x02, 0x41, 0x00, 0xbc, 0x3d, 0xde, 0x6d, 0xd6, 0x97,
0xe8, 0xba, 0x9e, 0x81, 0x37, 0x17, 0xe5, 0xa0, 0x64, 0xc9, 0x00, 0xb7,
0xe7, 0xfe, 0xf4, 0x29, 0xd9, 0x2e, 0x43, 0x6b, 0x19, 0x20, 0xbd, 0x99,
0x75, 0xe7, 0x76, 0xf8, 0xd3, 0xae, 0xaf, 0x7e, 0xb8, 0xeb, 0x81, 0xf4,
0x9d, 0xfe, 0x07, 0x2b, 0x0b, 0x63, 0x0b, 0x5a, 0x55, 0x90, 0x71, 0x7d,
0xf1, 0xdb, 0xd9, 0xb1, 0x41, 0x41, 0x68, 0x2f, 0x4e, 0x39, 0x02, 0x40,
0x5a, 0x34, 0x66, 0xd8, 0xf5, 0xe2, 0x7f, 0x18, 0xb5, 0x00, 0x6e, 0x26,
0x84, 0x27, 0x14, 0x93, 0xfb, 0xfc, 0xc6, 0x0f, 0x5e, 0x27, 0xe6, 0xe1,
0xe9, 0xc0, 0x8a, 0xe4, 0x34, 0xda, 0xe9, 0xa2, 0x4b, 0x73, 0xbc, 0x8c,
0xb9, 0xba, 0x13, 0x6c, 0x7a, 0x2b, 0x51, 0x84, 0xa3, 0x4a, 0xe0, 0x30,
0x10, 0x06, 0x7e, 0xed, 0x17, 0x5a, 0x14, 0x00, 0xc9, 0xef, 0x85, 0xea,
0x52, 0x2c, 0xbc, 0x65, 0x02, 0x40, 0x51, 0xe3, 0xf2, 0x83, 0x19, 0x9b,
0xc4, 0x1e, 0x2f, 0x50, 0x3d, 0xdf, 0x5a, 0xa2, 0x18, 0xca, 0x5f, 0x2e,
0x49, 0xaf, 0x6f, 0xcc, 0xfa, 0x65, 0x77, 0x94, 0xb5, 0xa1, 0x0a, 0xa9,
0xd1, 0x8a, 0x39, 0x37, 0xf4, 0x0b, 0xa0, 0xd7, 0x82, 0x27, 0x5e, 0xae,
0x17, 0x17, 0xa1, 0x1e, 0x54, 0x34, 0xbf, 0x6e, 0xc4, 0x8e, 0x99, 0x5d,
0x08, 0xf1, 0x2d, 0x86, 0x9d, 0xa5, 0x20, 0x1b, 0xe5, 0xdf,
};
Fix evp_extra_test.c with no-ec When OpenSSL is configured with no-ec, then the new evp_extra_test fails to pass. This change adds appropriate OPENSSL_NO_EC guards around the code. Reviewed-by: Tim Hudson <tjh@openssl.org>
2015-02-26 18:35:50 +08:00
#ifndef OPENSSL_NO_EC
Import evp_test.c from BoringSSL. Unfortunately we already have a file called evp_test.c, so I have called this one evp_extra_test.c Reviewed-by: Emilia Käsper <emilia@openssl.org>
2015-02-09 17:45:35 +08:00
/*
* kExampleECKeyDER is a sample EC private key encoded as an ECPrivateKey
* structure.
*/
static const unsigned char kExampleECKeyDER[] = {
0x30, 0x77, 0x02, 0x01, 0x01, 0x04, 0x20, 0x07, 0x0f, 0x08, 0x72, 0x7a,
0xd4, 0xa0, 0x4a, 0x9c, 0xdd, 0x59, 0xc9, 0x4d, 0x89, 0x68, 0x77, 0x08,
0xb5, 0x6f, 0xc9, 0x5d, 0x30, 0x77, 0x0e, 0xe8, 0xd1, 0xc9, 0xce, 0x0a,
0x8b, 0xb4, 0x6a, 0xa0, 0x0a, 0x06, 0x08, 0x2a, 0x86, 0x48, 0xce, 0x3d,
0x03, 0x01, 0x07, 0xa1, 0x44, 0x03, 0x42, 0x00, 0x04, 0xe6, 0x2b, 0x69,
0xe2, 0xbf, 0x65, 0x9f, 0x97, 0xbe, 0x2f, 0x1e, 0x0d, 0x94, 0x8a, 0x4c,
0xd5, 0x97, 0x6b, 0xb7, 0xa9, 0x1e, 0x0d, 0x46, 0xfb, 0xdd, 0xa9, 0xa9,
0x1e, 0x9d, 0xdc, 0xba, 0x5a, 0x01, 0xe7, 0xd6, 0x97, 0xa8, 0x0a, 0x18,
0xf9, 0xc3, 0xc4, 0xa3, 0x1e, 0x56, 0xe2, 0x7c, 0x83, 0x48, 0xdb, 0x16,
0x1a, 0x1c, 0xf5, 0x1d, 0x7e, 0xf1, 0x94, 0x2d, 0x4b, 0xcf, 0x72, 0x22,
0xc1,
};
/*
* kExampleBadECKeyDER is a sample EC private key encoded as an ECPrivateKey
* structure. The private key is equal to the order and will fail to import
*/
static const unsigned char kExampleBadECKeyDER[] = {
0x30, 0x66, 0x02, 0x01, 0x00, 0x30, 0x13, 0x06, 0x07, 0x2A, 0x86, 0x48,
0xCE, 0x3D, 0x02, 0x01, 0x06, 0x08, 0x2A, 0x86, 0x48, 0xCE, 0x3D, 0x03,
0x01, 0x07, 0x04, 0x4C, 0x30, 0x4A, 0x02, 0x01, 0x01, 0x04, 0x20, 0xFF,
0xFF, 0xFF, 0xFF, 0x00, 0x00, 0x00, 0x00, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF,
0xFF, 0xFF, 0xFF, 0xBC, 0xE6, 0xFA, 0xAD, 0xA7, 0x17, 0x9E, 0x84, 0xF3,
0xB9, 0xCA, 0xC2, 0xFC, 0x63, 0x25, 0x51, 0xA1, 0x23, 0x03, 0x21, 0x00,
0x00, 0xFF, 0xFF, 0xFF, 0x00, 0x00, 0x00, 0x00, 0xFF, 0xFF, 0xFF, 0xFF,
0xFF, 0xFF, 0xFF, 0xFF, 0xBC, 0xE6, 0xFA, 0xAD, 0xA7, 0x17, 0x9E, 0x84,
0xF3, 0xB9, 0xCA, 0xC2, 0xFC, 0x63, 0x25, 0x51
};
Support public key and param check in EVP interface EVP_PKEY_public_check() and EVP_PKEY_param_check() Doc and test cases are added Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/4647)
2017-11-01 00:45:24 +08:00
/* prime256v1 */
static const unsigned char kExampleECPubKeyDER[] = {
0x30, 0x59, 0x30, 0x13, 0x06, 0x07, 0x2a, 0x86, 0x48, 0xce, 0x3d, 0x02,
0x01, 0x06, 0x08, 0x2a, 0x86, 0x48, 0xce, 0x3d, 0x03, 0x01, 0x07, 0x03,
0x42, 0x00, 0x04, 0xba, 0xeb, 0x83, 0xfb, 0x3b, 0xb2, 0xff, 0x30, 0x53,
0xdb, 0xce, 0x32, 0xf2, 0xac, 0xae, 0x44, 0x0d, 0x3d, 0x13, 0x53, 0xb8,
0xd1, 0x68, 0x55, 0xde, 0x44, 0x46, 0x05, 0xa6, 0xc9, 0xd2, 0x04, 0xb7,
0xe3, 0xa2, 0x96, 0xc8, 0xb2, 0x5e, 0x22, 0x03, 0xd7, 0x03, 0x7a, 0x8b,
0x13, 0x5c, 0x42, 0x49, 0xc2, 0xab, 0x86, 0xd6, 0xac, 0x6b, 0x93, 0x20,
0x56, 0x6a, 0xc6, 0xc8, 0xa5, 0x0b, 0xe5
};
Fix a crash in reuse of i2d_X509_PUBKEY If the second PUBKEY is malformed there is use after free. Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/8122)
2019-01-30 23:20:31 +08:00
/*
Fixed typo in code comment. [skip ci] Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/9102)
2019-06-07 08:45:51 +08:00
* kExampleBadECPubKeyDER is a sample EC public key with a wrong OID
Fix a crash in reuse of i2d_X509_PUBKEY If the second PUBKEY is malformed there is use after free. Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/8122)
2019-01-30 23:20:31 +08:00
* 1.2.840.10045.2.2 instead of 1.2.840.10045.2.1 - EC Public Key
*/
static const unsigned char kExampleBadECPubKeyDER[] = {
0x30, 0x59, 0x30, 0x13, 0x06, 0x07, 0x2a, 0x86, 0x48, 0xce, 0x3d, 0x02,
0x02, 0x06, 0x08, 0x2a, 0x86, 0x48, 0xce, 0x3d, 0x03, 0x01, 0x07, 0x03,
0x42, 0x00, 0x04, 0xba, 0xeb, 0x83, 0xfb, 0x3b, 0xb2, 0xff, 0x30, 0x53,
0xdb, 0xce, 0x32, 0xf2, 0xac, 0xae, 0x44, 0x0d, 0x3d, 0x13, 0x53, 0xb8,
0xd1, 0x68, 0x55, 0xde, 0x44, 0x46, 0x05, 0xa6, 0xc9, 0xd2, 0x04, 0xb7,
0xe3, 0xa2, 0x96, 0xc8, 0xb2, 0x5e, 0x22, 0x03, 0xd7, 0x03, 0x7a, 0x8b,
0x13, 0x5c, 0x42, 0x49, 0xc2, 0xab, 0x86, 0xd6, 0xac, 0x6b, 0x93, 0x20,
0x56, 0x6a, 0xc6, 0xc8, 0xa5, 0x0b, 0xe5
};
Support public key and param check in EVP interface EVP_PKEY_public_check() and EVP_PKEY_param_check() Doc and test cases are added Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/4647)
2017-11-01 00:45:24 +08:00
static const unsigned char pExampleECParamDER[] = {
0x06, 0x08, 0x2a, 0x86, 0x48, 0xce, 0x3d, 0x03, 0x01, 0x07
};
Ensure that ECX keys pass EVP_PKEY_param_check() RSA keys have no parameters and pass EVP_PKEY_param_check(). Previously, ECX keys had no parammeters and failed EVP_PKEY_param_check(). We should be consistent. It makes more sense to always pass, and therefore this commit implements that behaviour. Fixes #14482 Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/14485)
2021-03-10 01:07:48 +08:00
configure: introduce no-ecx to remove ECX related feature This can effectively reduce the binary size for platforms that don't need ECX feature(~100KB). Signed-off-by: Yi Li <yi1.li@intel.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/20781)
2023-04-17 16:20:31 +08:00
# ifndef OPENSSL_NO_ECX
Ensure that ECX keys pass EVP_PKEY_param_check() RSA keys have no parameters and pass EVP_PKEY_param_check(). Previously, ECX keys had no parammeters and failed EVP_PKEY_param_check(). We should be consistent. It makes more sense to always pass, and therefore this commit implements that behaviour. Fixes #14482 Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/14485)
2021-03-10 01:07:48 +08:00
static const unsigned char kExampleED25519KeyDER[] = {
0x30, 0x2e, 0x02, 0x01, 0x00, 0x30, 0x05, 0x06, 0x03, 0x2b, 0x65, 0x70,
0x04, 0x22, 0x04, 0x20, 0xba, 0x7b, 0xba, 0x20, 0x1b, 0x02, 0x75, 0x3a,
0xe8, 0x88, 0xfe, 0x00, 0xcd, 0x8b, 0xc6, 0xf4, 0x5c, 0x47, 0x09, 0x46,
0x66, 0xe4, 0x72, 0x85, 0x25, 0x26, 0x5e, 0x12, 0x33, 0x48, 0xf6, 0x50
};
static const unsigned char kExampleED25519PubKeyDER[] = {
0x30, 0x2a, 0x30, 0x05, 0x06, 0x03, 0x2b, 0x65, 0x70, 0x03, 0x21, 0x00,
0xf5, 0xc5, 0xeb, 0x52, 0x3e, 0x7d, 0x07, 0x86, 0xb2, 0x55, 0x07, 0x45,
0xef, 0x5b, 0x7c, 0x20, 0xe8, 0x66, 0x28, 0x30, 0x3c, 0x8a, 0x82, 0x40,
0x97, 0xa3, 0x08, 0xdc, 0x65, 0x80, 0x39, 0x29
};
Add a test for custom EVP_PKEY_METHODs Adds a test for using custom EVP_PKEY_METHODs without an ENGINE. As part of this we also test having a custom EVP_PKEY_METHOD that wraps a built-in EVP_PKEY_METHOD. We do this for both legacy and provided keys. Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16118)
2021-07-20 23:18:58 +08:00
# ifndef OPENSSL_NO_DEPRECATED_3_0
static const unsigned char kExampleX25519KeyDER[] = {
0x30, 0x2e, 0x02, 0x01, 0x00, 0x30, 0x05, 0x06, 0x03, 0x2b, 0x65, 0x6e,
0x04, 0x22, 0x04, 0x20, 0xa0, 0x24, 0x3a, 0x31, 0x24, 0xc3, 0x3f, 0xf6,
0x7b, 0x96, 0x0b, 0xd4, 0x8f, 0xd1, 0xee, 0x67, 0xf2, 0x9b, 0x88, 0xac,
0x50, 0xce, 0x97, 0x36, 0xdd, 0xaf, 0x25, 0xf6, 0x10, 0x34, 0x96, 0x6e
};
configure: introduce no-ecx to remove ECX related feature This can effectively reduce the binary size for platforms that don't need ECX feature(~100KB). Signed-off-by: Yi Li <yi1.li@intel.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/20781)
2023-04-17 16:20:31 +08:00
# endif
Add a test for custom EVP_PKEY_METHODs Adds a test for using custom EVP_PKEY_METHODs without an ENGINE. As part of this we also test having a custom EVP_PKEY_METHOD that wraps a built-in EVP_PKEY_METHOD. We do this for both legacy and provided keys. Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16118)
2021-07-20 23:18:58 +08:00
# endif
#endif
/* kExampleDHKeyDER is a DH private key in ASN.1, DER format. */
#ifndef OPENSSL_NO_DEPRECATED_3_0
# ifndef OPENSSL_NO_DH
static const unsigned char kExampleDHKeyDER[] = {
0x30, 0x82, 0x01, 0x21, 0x02, 0x01, 0x00, 0x30, 0x81, 0x95, 0x06, 0x09,
0x2a, 0x86, 0x48, 0x86, 0xf7, 0x0d, 0x01, 0x03, 0x01, 0x30, 0x81, 0x87,
0x02, 0x81, 0x81, 0x00, 0xf7, 0x52, 0xc2, 0x68, 0xcc, 0x66, 0xc4, 0x8d,
0x03, 0x3f, 0xfa, 0x9c, 0x52, 0xd0, 0xd8, 0x33, 0xf2, 0xe1, 0xc9, 0x9e,
0xb7, 0xe7, 0x6e, 0x90, 0x97, 0xeb, 0x92, 0x91, 0x6a, 0x9a, 0x85, 0x63,
0x92, 0x79, 0xab, 0xb6, 0x3d, 0x23, 0x58, 0x5a, 0xe8, 0x45, 0x06, 0x81,
0x97, 0x77, 0xe1, 0xcc, 0x34, 0x4e, 0xae, 0x36, 0x80, 0xf2, 0xc4, 0x7f,
0x8a, 0x52, 0xb8, 0xdb, 0x58, 0xc8, 0x4b, 0x12, 0x4c, 0xf1, 0x4c, 0x53,
0xc1, 0x89, 0x39, 0x8d, 0xb6, 0x06, 0xd8, 0xea, 0x7f, 0x2d, 0x36, 0x53,
0x96, 0x29, 0xbe, 0xb6, 0x75, 0xfc, 0xe7, 0xf3, 0x36, 0xd6, 0xf4, 0x8f,
0x16, 0xa6, 0xc7, 0xec, 0x7b, 0xce, 0x42, 0x8d, 0x48, 0x2e, 0xb7, 0x74,
0x00, 0x11, 0x52, 0x61, 0xb4, 0x19, 0x35, 0xec, 0x5c, 0xe4, 0xbe, 0x34,
0xc6, 0x59, 0x64, 0x5e, 0x42, 0x61, 0x70, 0x54, 0xf4, 0xe9, 0x6b, 0x53,
0x02, 0x01, 0x02, 0x04, 0x81, 0x83, 0x02, 0x81, 0x80, 0x64, 0xc2, 0xe3,
0x09, 0x69, 0x37, 0x3c, 0xd2, 0x4a, 0xba, 0xc3, 0x78, 0x6a, 0x9b, 0x8a,
0x2a, 0xdb, 0xe7, 0xe6, 0xc0, 0xfa, 0x3a, 0xbe, 0x39, 0x67, 0xc0, 0xa9,
0x2a, 0xf0, 0x0a, 0xc1, 0x53, 0x1c, 0xdb, 0xfa, 0x1a, 0x26, 0x98, 0xb0,
0x8c, 0xc6, 0x06, 0x4a, 0xa2, 0x48, 0xd3, 0xa4, 0x3b, 0xbd, 0x05, 0x48,
0xea, 0x59, 0xdb, 0x18, 0xa4, 0xca, 0x66, 0xd9, 0x5d, 0xb8, 0x95, 0xd1,
0xeb, 0x97, 0x3d, 0x66, 0x97, 0x5c, 0x86, 0x8f, 0x7e, 0x90, 0xd3, 0x43,
0xd1, 0xa2, 0x0d, 0xcb, 0xe7, 0xeb, 0x90, 0xea, 0x09, 0x40, 0xb1, 0x6f,
0xf7, 0x4c, 0xf2, 0x41, 0x83, 0x1d, 0xd0, 0x76, 0xef, 0xaf, 0x55, 0x6f,
0x5d, 0xa9, 0xa3, 0x55, 0x81, 0x2a, 0xd1, 0x5d, 0x9d, 0x22, 0x77, 0x97,
0x83, 0xde, 0xad, 0xb6, 0x5d, 0x19, 0xc1, 0x53, 0xec, 0xfb, 0xaf, 0x06,
0x2e, 0x87, 0x2a, 0x0b, 0x7a
};
# endif
Fix evp_extra_test.c with no-ec When OpenSSL is configured with no-ec, then the new evp_extra_test fails to pass. This change adds appropriate OPENSSL_NO_EC guards around the code. Reviewed-by: Tim Hudson <tjh@openssl.org>
2015-02-26 18:35:50 +08:00
#endif
Import evp_test.c from BoringSSL. Unfortunately we already have a file called evp_test.c, so I have called this one evp_extra_test.c Reviewed-by: Emilia Käsper <emilia@openssl.org>
2015-02-09 17:45:35 +08:00
Test EVP_CipherInit sequences and resets Various EVP_CipherInit sequences including partial inits and initializations with different "enc" flags caused problems on s390x. Similarly, cipher reinitialization and especially GCM reinitialization with different tag length led to wrong results. Add some unit tests to cover these rather exotic use cases. Signed-off-by: Juergen Christ <jchrist@linux.ibm.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/15521)
2021-06-03 01:33:50 +08:00
static const unsigned char kCFBDefaultKey[] = {
0x2B, 0x7E, 0x15, 0x16, 0x28, 0xAE, 0xD2, 0xA6, 0xAB, 0xF7, 0x15, 0x88,
0x09, 0xCF, 0x4F, 0x3C
};
static const unsigned char kGCMDefaultKey[32] = { 0 };
static const unsigned char kGCMResetKey[] = {
0xfe, 0xff, 0xe9, 0x92, 0x86, 0x65, 0x73, 0x1c, 0x6d, 0x6a, 0x8f, 0x94,
0x67, 0x30, 0x83, 0x08, 0xfe, 0xff, 0xe9, 0x92, 0x86, 0x65, 0x73, 0x1c,
0x6d, 0x6a, 0x8f, 0x94, 0x67, 0x30, 0x83, 0x08
};
static const unsigned char iCFBIV[] = {
0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08, 0x09, 0x0A, 0x0B,
0x0C, 0x0D, 0x0E, 0x0F
};
static const unsigned char iGCMDefaultIV[12] = { 0 };
static const unsigned char iGCMResetIV1[] = {
0xca, 0xfe, 0xba, 0xbe, 0xfa, 0xce, 0xdb, 0xad
};
static const unsigned char iGCMResetIV2[] = {
0xca, 0xfe, 0xba, 0xbe, 0xfa, 0xce, 0xdb, 0xad, 0xde, 0xca, 0xf8, 0x88
};
static const unsigned char cfbPlaintext[] = {
0x6B, 0xC1, 0xBE, 0xE2, 0x2E, 0x40, 0x9F, 0x96, 0xE9, 0x3D, 0x7E, 0x11,
0x73, 0x93, 0x17, 0x2A
};
Add tests for re-using cipher contexts Add test case for re-using a cipher context with the same key, iv and cipher. It detects, if the hardware-specific cipher context is reset correctly, like reported in issue #23175. This test has encrypt and decrypt iterations for cfb128 and ofb128. All iteations use the same key, iv and plaintext. Signed-off-by: Holger Dengler <dengler@linux.ibm.com> Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/23201)
2024-01-05 02:25:08 +08:00
static const unsigned char cfbPlaintext_partial[] = {
0x6B, 0xC1, 0xBE, 0xE2, 0x2E, 0x40, 0x9F, 0x96, 0xE9, 0x3D, 0x7E, 0x11,
0x73, 0x93, 0x17, 0x2A, 0x6B, 0xC1, 0xBE, 0xE2, 0x2E, 0x40, 0x9F, 0x96,
};
Test EVP_CipherInit sequences and resets Various EVP_CipherInit sequences including partial inits and initializations with different "enc" flags caused problems on s390x. Similarly, cipher reinitialization and especially GCM reinitialization with different tag length led to wrong results. Add some unit tests to cover these rather exotic use cases. Signed-off-by: Juergen Christ <jchrist@linux.ibm.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/15521)
2021-06-03 01:33:50 +08:00
static const unsigned char gcmDefaultPlaintext[16] = { 0 };
static const unsigned char gcmResetPlaintext[] = {
0xd9, 0x31, 0x32, 0x25, 0xf8, 0x84, 0x06, 0xe5, 0xa5, 0x59, 0x09, 0xc5,
0xaf, 0xf5, 0x26, 0x9a, 0x86, 0xa7, 0xa9, 0x53, 0x15, 0x34, 0xf7, 0xda,
0x2e, 0x4c, 0x30, 0x3d, 0x8a, 0x31, 0x8a, 0x72, 0x1c, 0x3c, 0x0c, 0x95,
0x95, 0x68, 0x09, 0x53, 0x2f, 0xcf, 0x0e, 0x24, 0x49, 0xa6, 0xb5, 0x25,
0xb1, 0x6a, 0xed, 0xf5, 0xaa, 0x0d, 0xe6, 0x57, 0xba, 0x63, 0x7b, 0x39
};
static const unsigned char cfbCiphertext[] = {
0x3B, 0x3F, 0xD9, 0x2E, 0xB7, 0x2D, 0xAD, 0x20, 0x33, 0x34, 0x49, 0xF8,
0xE8, 0x3C, 0xFB, 0x4A
};
Add tests for re-using cipher contexts Add test case for re-using a cipher context with the same key, iv and cipher. It detects, if the hardware-specific cipher context is reset correctly, like reported in issue #23175. This test has encrypt and decrypt iterations for cfb128 and ofb128. All iteations use the same key, iv and plaintext. Signed-off-by: Holger Dengler <dengler@linux.ibm.com> Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/23201)
2024-01-05 02:25:08 +08:00
static const unsigned char cfbCiphertext_partial[] = {
0x3B, 0x3F, 0xD9, 0x2E, 0xB7, 0x2D, 0xAD, 0x20, 0x33, 0x34, 0x49, 0xF8,
0xE8, 0x3C, 0xFB, 0x4A, 0x0D, 0x4A, 0x71, 0x82, 0x90, 0xF0, 0x9A, 0x35
};
static const unsigned char ofbCiphertext_partial[] = {
0x3B, 0x3F, 0xD9, 0x2E, 0xB7, 0x2D, 0xAD, 0x20, 0x33, 0x34, 0x49, 0xF8,
0xE8, 0x3C, 0xFB, 0x4A, 0xB2, 0x65, 0x64, 0x38, 0x26, 0xD2, 0xBC, 0x09
};
Test EVP_CipherInit sequences and resets Various EVP_CipherInit sequences including partial inits and initializations with different "enc" flags caused problems on s390x. Similarly, cipher reinitialization and especially GCM reinitialization with different tag length led to wrong results. Add some unit tests to cover these rather exotic use cases. Signed-off-by: Juergen Christ <jchrist@linux.ibm.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/15521)
2021-06-03 01:33:50 +08:00
static const unsigned char gcmDefaultCiphertext[] = {
0xce, 0xa7, 0x40, 0x3d, 0x4d, 0x60, 0x6b, 0x6e, 0x07, 0x4e, 0xc5, 0xd3,
0xba, 0xf3, 0x9d, 0x18
};
static const unsigned char gcmResetCiphertext1[] = {
0xc3, 0x76, 0x2d, 0xf1, 0xca, 0x78, 0x7d, 0x32, 0xae, 0x47, 0xc1, 0x3b,
0xf1, 0x98, 0x44, 0xcb, 0xaf, 0x1a, 0xe1, 0x4d, 0x0b, 0x97, 0x6a, 0xfa,
0xc5, 0x2f, 0xf7, 0xd7, 0x9b, 0xba, 0x9d, 0xe0, 0xfe, 0xb5, 0x82, 0xd3,
0x39, 0x34, 0xa4, 0xf0, 0x95, 0x4c, 0xc2, 0x36, 0x3b, 0xc7, 0x3f, 0x78,
0x62, 0xac, 0x43, 0x0e, 0x64, 0xab, 0xe4, 0x99, 0xf4, 0x7c, 0x9b, 0x1f
};
static const unsigned char gcmResetCiphertext2[] = {
0x52, 0x2d, 0xc1, 0xf0, 0x99, 0x56, 0x7d, 0x07, 0xf4, 0x7f, 0x37, 0xa3,
0x2a, 0x84, 0x42, 0x7d, 0x64, 0x3a, 0x8c, 0xdc, 0xbf, 0xe5, 0xc0, 0xc9,
0x75, 0x98, 0xa2, 0xbd, 0x25, 0x55, 0xd1, 0xaa, 0x8c, 0xb0, 0x8e, 0x48,
0x59, 0x0d, 0xbb, 0x3d, 0xa7, 0xb0, 0x8b, 0x10, 0x56, 0x82, 0x88, 0x38,
0xc5, 0xf6, 0x1e, 0x63, 0x93, 0xba, 0x7a, 0x0a, 0xbc, 0xc9, 0xf6, 0x62
};
static const unsigned char gcmAAD[] = {
0xfe, 0xed, 0xfa, 0xce, 0xde, 0xad, 0xbe, 0xef, 0xfe, 0xed, 0xfa, 0xce,
0xde, 0xad, 0xbe, 0xef, 0xab, 0xad, 0xda, 0xd2
};
static const unsigned char gcmDefaultTag[] = {
0xd0, 0xd1, 0xc8, 0xa7, 0x99, 0x99, 0x6b, 0xf0, 0x26, 0x5b, 0x98, 0xb5,
0xd4, 0x8a, 0xb9, 0x19
};
static const unsigned char gcmResetTag1[] = {
0x3a, 0x33, 0x7d, 0xbf, 0x46, 0xa7, 0x92, 0xc4, 0x5e, 0x45, 0x49, 0x13,
0xfe, 0x2e, 0xa8, 0xf2
};
static const unsigned char gcmResetTag2[] = {
0x76, 0xfc, 0x6e, 0xce, 0x0f, 0x4e, 0x17, 0x68, 0xcd, 0xdf, 0x88, 0x53,
0xbb, 0x2d, 0x55, 0x1b
};
Convert more tests ct_test,evp_extra_test,wpackettest,packettest Add strncmp TEST wrappers And make some style/consistency fixes to ct_test Silence travis; gcc bug? Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/3234)
2017-04-19 04:33:15 +08:00
typedef struct APK_DATA_st {
const unsigned char *kder;
size_t size;
Deprecate EC_KEY + Update ec apps to use EVP_PKEY Co-author: Richard Levitte <levitte@openssl.org> Co-author: Tomas Mraz <tmraz@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13139)
2020-10-15 11:41:59 +08:00
const char *keytype;
Convert more tests ct_test,evp_extra_test,wpackettest,packettest Add strncmp TEST wrappers And make some style/consistency fixes to ct_test Silence travis; gcc bug? Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/3234)
2017-04-19 04:33:15 +08:00
int evptype;
Support key check in EVP interface A new method is added to EVP_PKEY_METH as: int (*check) (EVP_PKEY_CTX *ctx); and to EVP_PKEY_ASN1_METHOD as: int (*pkey_check) (EVP_PKEY_CTX *ctx); This is used to check the validity of a specific key. The order of calls is: EVP_PKEY_check -> pmeth.check -> ameth.pkey_check. Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/4337)
2017-09-04 22:02:59 +08:00
int check;
Support public key and param check in EVP interface EVP_PKEY_public_check() and EVP_PKEY_param_check() Doc and test cases are added Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/4647)
2017-11-01 00:45:24 +08:00
int pub_check;
int param_check;
int type; /* 0 for private, 1 for public, 2 for params */
Convert more tests ct_test,evp_extra_test,wpackettest,packettest Add strncmp TEST wrappers And make some style/consistency fixes to ct_test Silence travis; gcc bug? Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/3234)
2017-04-19 04:33:15 +08:00
} APK_DATA;
static APK_DATA keydata[] = {
Deprecate EC_KEY + Update ec apps to use EVP_PKEY Co-author: Richard Levitte <levitte@openssl.org> Co-author: Tomas Mraz <tmraz@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13139)
2020-10-15 11:41:59 +08:00
{kExampleRSAKeyDER, sizeof(kExampleRSAKeyDER), "RSA", EVP_PKEY_RSA},
{kExampleRSAKeyPKCS8, sizeof(kExampleRSAKeyPKCS8), "RSA", EVP_PKEY_RSA},
Fix no-ec Reviewed-by: Rich Salz <rsalz@openssl.org> (Merged from https://github.com/openssl/openssl/pull/3321)
2017-04-27 00:08:22 +08:00
#ifndef OPENSSL_NO_EC
Deprecate EC_KEY + Update ec apps to use EVP_PKEY Co-author: Richard Levitte <levitte@openssl.org> Co-author: Tomas Mraz <tmraz@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13139)
2020-10-15 11:41:59 +08:00
{kExampleECKeyDER, sizeof(kExampleECKeyDER), "EC", EVP_PKEY_EC}
Fix no-ec Reviewed-by: Rich Salz <rsalz@openssl.org> (Merged from https://github.com/openssl/openssl/pull/3321)
2017-04-27 00:08:22 +08:00
#endif
Convert more tests ct_test,evp_extra_test,wpackettest,packettest Add strncmp TEST wrappers And make some style/consistency fixes to ct_test Silence travis; gcc bug? Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/3234)
2017-04-19 04:33:15 +08:00
};
Support key check in EVP interface A new method is added to EVP_PKEY_METH as: int (*check) (EVP_PKEY_CTX *ctx); and to EVP_PKEY_ASN1_METHOD as: int (*pkey_check) (EVP_PKEY_CTX *ctx); This is used to check the validity of a specific key. The order of calls is: EVP_PKEY_check -> pmeth.check -> ameth.pkey_check. Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/4337)
2017-09-04 22:02:59 +08:00
static APK_DATA keycheckdata[] = {
Ensure that ECX keys pass EVP_PKEY_param_check() RSA keys have no parameters and pass EVP_PKEY_param_check(). Previously, ECX keys had no parammeters and failed EVP_PKEY_param_check(). We should be consistent. It makes more sense to always pass, and therefore this commit implements that behaviour. Fixes #14482 Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/14485)
2021-03-10 01:07:48 +08:00
{kExampleRSAKeyDER, sizeof(kExampleRSAKeyDER), "RSA", EVP_PKEY_RSA, 1, 1, 1,
0},
Deprecate EC_KEY + Update ec apps to use EVP_PKEY Co-author: Richard Levitte <levitte@openssl.org> Co-author: Tomas Mraz <tmraz@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13139)
2020-10-15 11:41:59 +08:00
{kExampleBadRSAKeyDER, sizeof(kExampleBadRSAKeyDER), "RSA", EVP_PKEY_RSA,
test/evp_extra_test.c: Modify to reflect provider support in test_EVP_PKEY_check With our providers, RSA now supports public key check and key parameter check. Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/13334)
2020-11-06 17:37:43 +08:00
0, 1, 1, 0},
Add a test for an all 0 RSA key Testing a key with all values set to 0. Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/18651)
2022-06-24 18:02:22 +08:00
{kExampleBad2RSAKeyDER, sizeof(kExampleBad2RSAKeyDER), "RSA", EVP_PKEY_RSA,
0, 0, 1 /* Since there are no "params" in an RSA key this passes */, 0},
Support key check in EVP interface A new method is added to EVP_PKEY_METH as: int (*check) (EVP_PKEY_CTX *ctx); and to EVP_PKEY_ASN1_METHOD as: int (*pkey_check) (EVP_PKEY_CTX *ctx); This is used to check the validity of a specific key. The order of calls is: EVP_PKEY_check -> pmeth.check -> ameth.pkey_check. Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/4337)
2017-09-04 22:02:59 +08:00
#ifndef OPENSSL_NO_EC
Deprecate EC_KEY + Update ec apps to use EVP_PKEY Co-author: Richard Levitte <levitte@openssl.org> Co-author: Tomas Mraz <tmraz@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13139)
2020-10-15 11:41:59 +08:00
{kExampleECKeyDER, sizeof(kExampleECKeyDER), "EC", EVP_PKEY_EC, 1, 1, 1, 0},
Support public key and param check in EVP interface EVP_PKEY_public_check() and EVP_PKEY_param_check() Doc and test cases are added Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/4647)
2017-11-01 00:45:24 +08:00
/* group is also associated in our pub key */
Ensure that ECX keys pass EVP_PKEY_param_check() RSA keys have no parameters and pass EVP_PKEY_param_check(). Previously, ECX keys had no parammeters and failed EVP_PKEY_param_check(). We should be consistent. It makes more sense to always pass, and therefore this commit implements that behaviour. Fixes #14482 Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/14485)
2021-03-10 01:07:48 +08:00
{kExampleECPubKeyDER, sizeof(kExampleECPubKeyDER), "EC", EVP_PKEY_EC, 0, 1,
1, 1},
{pExampleECParamDER, sizeof(pExampleECParamDER), "EC", EVP_PKEY_EC, 0, 0, 1,
2},
configure: introduce no-ecx to remove ECX related feature This can effectively reduce the binary size for platforms that don't need ECX feature(~100KB). Signed-off-by: Yi Li <yi1.li@intel.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/20781)
2023-04-17 16:20:31 +08:00
# ifndef OPENSSL_NO_ECX
Ensure that ECX keys pass EVP_PKEY_param_check() RSA keys have no parameters and pass EVP_PKEY_param_check(). Previously, ECX keys had no parammeters and failed EVP_PKEY_param_check(). We should be consistent. It makes more sense to always pass, and therefore this commit implements that behaviour. Fixes #14482 Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/14485)
2021-03-10 01:07:48 +08:00
{kExampleED25519KeyDER, sizeof(kExampleED25519KeyDER), "ED25519",
EVP_PKEY_ED25519, 1, 1, 1, 0},
{kExampleED25519PubKeyDER, sizeof(kExampleED25519PubKeyDER), "ED25519",
EVP_PKEY_ED25519, 0, 1, 1, 1},
configure: introduce no-ecx to remove ECX related feature This can effectively reduce the binary size for platforms that don't need ECX feature(~100KB). Signed-off-by: Yi Li <yi1.li@intel.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/20781)
2023-04-17 16:20:31 +08:00
# endif
Support key check in EVP interface A new method is added to EVP_PKEY_METH as: int (*check) (EVP_PKEY_CTX *ctx); and to EVP_PKEY_ASN1_METHOD as: int (*pkey_check) (EVP_PKEY_CTX *ctx); This is used to check the validity of a specific key. The order of calls is: EVP_PKEY_check -> pmeth.check -> ameth.pkey_check. Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/4337)
2017-09-04 22:02:59 +08:00
#endif
};
Deprecate RSA harder This deprecates all functions that deal with the types RSA and RSA_METHOD Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/13096)
2020-10-04 22:34:31 +08:00
static EVP_PKEY *load_example_key(const char *keytype,
const unsigned char *data, size_t data_len)
Import evp_test.c from BoringSSL. Unfortunately we already have a file called evp_test.c, so I have called this one evp_extra_test.c Reviewed-by: Emilia Käsper <emilia@openssl.org>
2015-02-09 17:45:35 +08:00
{
Deprecate RSA harder This deprecates all functions that deal with the types RSA and RSA_METHOD Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/13096)
2020-10-04 22:34:31 +08:00
const unsigned char **pdata = &data;
Import evp_test.c from BoringSSL. Unfortunately we already have a file called evp_test.c, so I have called this one evp_extra_test.c Reviewed-by: Emilia Käsper <emilia@openssl.org>
2015-02-09 17:45:35 +08:00
EVP_PKEY *pkey = NULL;
Deprecate RSA harder This deprecates all functions that deal with the types RSA and RSA_METHOD Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/13096)
2020-10-04 22:34:31 +08:00
OSSL_DECODER_CTX *dctx =
Rename OSSL_ENCODER_CTX_new_by_EVP_PKEY and OSSL_DECODER_CTX_new_by_EVP_PKEY Additional renames done in encoder and decoder implementation to follow the style. Fixes #13622 Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14155)
2021-02-11 23:57:37 +08:00
OSSL_DECODER_CTX_new_for_pkey(&pkey, "DER", NULL, keytype, 0,
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
testctx, testpropq);
Import evp_test.c from BoringSSL. Unfortunately we already have a file called evp_test.c, so I have called this one evp_extra_test.c Reviewed-by: Emilia Käsper <emilia@openssl.org>
2015-02-09 17:45:35 +08:00
Deprecate RSA harder This deprecates all functions that deal with the types RSA and RSA_METHOD Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/13096)
2020-10-04 22:34:31 +08:00
/* |pkey| will be NULL on error */
(void)OSSL_DECODER_from_data(dctx, pdata, &data_len);
OSSL_DECODER_CTX_free(dctx);
return pkey;
}
Import evp_test.c from BoringSSL. Unfortunately we already have a file called evp_test.c, so I have called this one evp_extra_test.c Reviewed-by: Emilia Käsper <emilia@openssl.org>
2015-02-09 17:45:35 +08:00
Deprecate RSA harder This deprecates all functions that deal with the types RSA and RSA_METHOD Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/13096)
2020-10-04 22:34:31 +08:00
static EVP_PKEY *load_example_rsa_key(void)
{
return load_example_key("RSA", kExampleRSAKeyDER,
sizeof(kExampleRSAKeyDER));
Import evp_test.c from BoringSSL. Unfortunately we already have a file called evp_test.c, so I have called this one evp_extra_test.c Reviewed-by: Emilia Käsper <emilia@openssl.org>
2015-02-09 17:45:35 +08:00
}
Fix no-dsa Make sure we don't try and load a DSA key in the tests if DSA has been disabled. Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10375)
2019-11-07 17:24:18 +08:00
#ifndef OPENSSL_NO_DSA
Add a test for EVP_Digest[Sign|Verify]* with a BIO MD If an EVP_MD_CTX is initialised with EVP_DigestSignInit_ex() or EVP_DigestVerifyInit_ex() it should work with an MD BIO to provide the "Updates". Test that this is the case. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10116)
2019-10-08 01:21:39 +08:00
static EVP_PKEY *load_example_dsa_key(void)
{
Fix 'no-deprecated' Some of the handling of no-deprecated stuff wasn't quite complete, or even plain wrong. This restores i2d_PublicKey() to be able to handle EVP_PKEYs with legacy internal keys. This also refactors the DSA key tests in test/evp_extra_test.c to use EVP functionality entirely. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/13706)
2020-12-18 20:17:33 +08:00
return load_example_key("DSA", kExampleDSAKeyDER,
sizeof(kExampleDSAKeyDER));
Add a test for EVP_Digest[Sign|Verify]* with a BIO MD If an EVP_MD_CTX is initialised with EVP_DigestSignInit_ex() or EVP_DigestVerifyInit_ex() it should work with an MD BIO to provide the "Updates". Test that this is the case. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10116)
2019-10-08 01:21:39 +08:00
}
Fix no-dsa Make sure we don't try and load a DSA key in the tests if DSA has been disabled. Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10375)
2019-11-07 17:24:18 +08:00
#endif
Add a test for EVP_Digest[Sign|Verify]* with a BIO MD If an EVP_MD_CTX is initialised with EVP_DigestSignInit_ex() or EVP_DigestVerifyInit_ex() it should work with an MD BIO to provide the "Updates". Test that this is the case. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10116)
2019-10-08 01:21:39 +08:00
Test short buffers Test that calling EVP_DigestSign(), EVP_DigestSignFinal(), EVP_PKEY_sign(), EVP_PKEY_get_raw_private_key(), or EVP_PKEY_get_raw_public_key() with a short output buffer results in a failure. Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16789)
2021-10-07 21:15:47 +08:00
#ifndef OPENSSL_NO_EC
static EVP_PKEY *load_example_ec_key(void)
{
return load_example_key("EC", kExampleECKeyDER,
sizeof(kExampleECKeyDER));
}
#endif
Add a test for custom EVP_PKEY_METHODs Adds a test for using custom EVP_PKEY_METHODs without an ENGINE. As part of this we also test having a custom EVP_PKEY_METHOD that wraps a built-in EVP_PKEY_METHOD. We do this for both legacy and provided keys. Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16118)
2021-07-20 23:18:58 +08:00
#ifndef OPENSSL_NO_DEPRECATED_3_0
# ifndef OPENSSL_NO_DH
static EVP_PKEY *load_example_dh_key(void)
{
return load_example_key("DH", kExampleDHKeyDER,
sizeof(kExampleDHKeyDER));
}
# endif
configure: introduce no-ecx to remove ECX related feature This can effectively reduce the binary size for platforms that don't need ECX feature(~100KB). Signed-off-by: Yi Li <yi1.li@intel.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/20781)
2023-04-17 16:20:31 +08:00
# ifndef OPENSSL_NO_ECX
Add a test for custom EVP_PKEY_METHODs Adds a test for using custom EVP_PKEY_METHODs without an ENGINE. As part of this we also test having a custom EVP_PKEY_METHOD that wraps a built-in EVP_PKEY_METHOD. We do this for both legacy and provided keys. Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16118)
2021-07-20 23:18:58 +08:00
static EVP_PKEY *load_example_ed25519_key(void)
{
return load_example_key("ED25519", kExampleED25519KeyDER,
sizeof(kExampleED25519KeyDER));
}
static EVP_PKEY *load_example_x25519_key(void)
{
return load_example_key("X25519", kExampleX25519KeyDER,
sizeof(kExampleX25519KeyDER));
}
# endif
#endif /* OPENSSL_NO_DEPRECATED_3_0 */
Add a test for HMAC via EVP_DigestSign* We test with both an implicitly fetched digest and an explicitly fetched digest. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/10796)
2020-01-09 23:21:14 +08:00
static EVP_PKEY *load_example_hmac_key(void)
{
EVP_PKEY *pkey = NULL;
unsigned char key[] = {
0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08, 0x09, 0x0a, 0x0b,
0x0c, 0x0d, 0x0e, 0x0f, 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17,
0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f
};
Trivial shortcuts for EVP_PKEY_eq() Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14942)
2021-04-20 22:39:00 +08:00
pkey = EVP_PKEY_new_raw_private_key_ex(testctx, "HMAC",
NULL, key, sizeof(key));
Add a test for HMAC via EVP_DigestSign* We test with both an implicitly fetched digest and an explicitly fetched digest. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/10796)
2020-01-09 23:21:14 +08:00
if (!TEST_ptr(pkey))
return NULL;
return pkey;
}
EVP_get_default_properties - tests Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/25434)
2024-09-11 22:48:44 +08:00
static int test_EVP_set_config_properties(void)
{
char *fetched_properties = NULL;
const char test_propq[] = "test.fizzbuzz=buzzfizz";
int res = 0;
fetched_properties = EVP_get1_default_properties(OSSL_LIB_CTX_get0_global_default());
if (!TEST_ptr(fetched_properties)
|| !TEST_str_eq(fetched_properties, test_propq))
goto err;
OPENSSL_free(fetched_properties);
fetched_properties = NULL;
res = 1;
err:
OPENSSL_free(fetched_properties);
return res;
}
property: Move global default properties to the library context. Fixes a problem where global properties don't work with a NULL query. Specifying an algorithm with a NULL query ignores the default properties. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/12123)
2020-06-12 08:34:46 +08:00
static int test_EVP_set_default_properties(void)
{
Rename OPENSSL_CTX prefix to OSSL_LIB_CTX Many of the new types introduced by OpenSSL 3.0 have an OSSL_ prefix, e.g., OSSL_CALLBACK, OSSL_PARAM, OSSL_ALGORITHM, OSSL_SERIALIZER. The OPENSSL_CTX type stands out a little by using a different prefix. For consistency reasons, this type is renamed to OSSL_LIB_CTX. Reviewed-by: Paul Dale <paul.dale@oracle.com> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/12621)
2020-10-15 17:55:50 +08:00
OSSL_LIB_CTX *ctx;
property: Move global default properties to the library context. Fixes a problem where global properties don't work with a NULL query. Specifying an algorithm with a NULL query ignores the default properties. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/12123)
2020-06-12 08:34:46 +08:00
EVP_MD *md = NULL;
int res = 0;
EVP_get_default_properties - tests Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/25434)
2024-09-11 22:48:44 +08:00
char *fetched_properties = NULL;
const char test_propq[] = "provider=fizzbang";
const char test_fips_propq[] = "fips=yes,provider=fizzbang";
property: Move global default properties to the library context. Fixes a problem where global properties don't work with a NULL query. Specifying an algorithm with a NULL query ignores the default properties. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/12123)
2020-06-12 08:34:46 +08:00
Rename OPENSSL_CTX prefix to OSSL_LIB_CTX Many of the new types introduced by OpenSSL 3.0 have an OSSL_ prefix, e.g., OSSL_CALLBACK, OSSL_PARAM, OSSL_ALGORITHM, OSSL_SERIALIZER. The OPENSSL_CTX type stands out a little by using a different prefix. For consistency reasons, this type is renamed to OSSL_LIB_CTX. Reviewed-by: Paul Dale <paul.dale@oracle.com> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/12621)
2020-10-15 17:55:50 +08:00
if (!TEST_ptr(ctx = OSSL_LIB_CTX_new())
property: Move global default properties to the library context. Fixes a problem where global properties don't work with a NULL query. Specifying an algorithm with a NULL query ignores the default properties. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/12123)
2020-06-12 08:34:46 +08:00
|| !TEST_ptr(md = EVP_MD_fetch(ctx, "sha256", NULL)))
goto err;
EVP_MD_free(md);
md = NULL;
EVP_get_default_properties - tests Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/25434)
2024-09-11 22:48:44 +08:00
if (!TEST_true(EVP_set_default_properties(ctx, test_propq))
property: Move global default properties to the library context. Fixes a problem where global properties don't work with a NULL query. Specifying an algorithm with a NULL query ignores the default properties. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/12123)
2020-06-12 08:34:46 +08:00
|| !TEST_ptr_null(md = EVP_MD_fetch(ctx, "sha256", NULL))
|| !TEST_ptr(md = EVP_MD_fetch(ctx, "sha256", "-provider")))
goto err;
EVP_MD_free(md);
md = NULL;
EVP_get_default_properties - tests Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/25434)
2024-09-11 22:48:44 +08:00
fetched_properties = EVP_get1_default_properties(ctx);
if (!TEST_ptr(fetched_properties)
|| !TEST_str_eq(fetched_properties, test_propq))
goto err;
OPENSSL_free(fetched_properties);
fetched_properties = NULL;
if (!TEST_true(EVP_default_properties_enable_fips(ctx, 1)))
goto err;
fetched_properties = EVP_get1_default_properties(ctx);
if (!TEST_ptr(fetched_properties)
|| !TEST_str_eq(fetched_properties, test_fips_propq))
goto err;
OPENSSL_free(fetched_properties);
fetched_properties = NULL;
if (!TEST_true(EVP_default_properties_enable_fips(ctx, 0)))
goto err;
property: Move global default properties to the library context. Fixes a problem where global properties don't work with a NULL query. Specifying an algorithm with a NULL query ignores the default properties. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/12123)
2020-06-12 08:34:46 +08:00
if (!TEST_true(EVP_set_default_properties(ctx, NULL))
|| !TEST_ptr(md = EVP_MD_fetch(ctx, "sha256", NULL)))
goto err;
res = 1;
err:
EVP_get_default_properties - tests Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/25434)
2024-09-11 22:48:44 +08:00
OPENSSL_free(fetched_properties);
property: Move global default properties to the library context. Fixes a problem where global properties don't work with a NULL query. Specifying an algorithm with a NULL query ignores the default properties. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/12123)
2020-06-12 08:34:46 +08:00
EVP_MD_free(md);
Rename OPENSSL_CTX prefix to OSSL_LIB_CTX Many of the new types introduced by OpenSSL 3.0 have an OSSL_ prefix, e.g., OSSL_CALLBACK, OSSL_PARAM, OSSL_ALGORITHM, OSSL_SERIALIZER. The OPENSSL_CTX type stands out a little by using a different prefix. For consistency reasons, this type is renamed to OSSL_LIB_CTX. Reviewed-by: Paul Dale <paul.dale@oracle.com> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/12621)
2020-10-15 17:55:50 +08:00
OSSL_LIB_CTX_free(ctx);
property: Move global default properties to the library context. Fixes a problem where global properties don't work with a NULL query. Specifying an algorithm with a NULL query ignores the default properties. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/12123)
2020-06-12 08:34:46 +08:00
return res;
}
Test that EC keys without a public key in them work as expected We create EC keys via both the "fromdata" and legacy key routes to make sure that they can be used without a public key. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13922)
2021-01-21 23:14:15 +08:00
#if !defined(OPENSSL_NO_DH) || !defined(OPENSSL_NO_DSA) || !defined(OPENSSL_NO_EC)
test/evp_extra_test.c: Refactor test_fromdata() test_fromdata() turns out to be a bit inflexible, so we split it into two functions, make_key_fromdata() and test_selection(), and adjust test_EVP_PKEY_ffc_priv_pub() and test_EC_priv_pub() accordingly. This allows us to check the resulting keys further, not only to check that the bits we expect are there, but also that the bits that we expect not to be there to actually not be there! Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16765)
2021-12-10 20:15:10 +08:00
static EVP_PKEY *make_key_fromdata(char *keytype, OSSL_PARAM *params)
Test that EC keys without a public key in them work as expected We create EC keys via both the "fromdata" and legacy key routes to make sure that they can be used without a public key. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13922)
2021-01-21 23:14:15 +08:00
{
EVP_PKEY_CTX *pctx = NULL;
test/evp_extra_test.c: Refactor test_fromdata() test_fromdata() turns out to be a bit inflexible, so we split it into two functions, make_key_fromdata() and test_selection(), and adjust test_EVP_PKEY_ffc_priv_pub() and test_EC_priv_pub() accordingly. This allows us to check the resulting keys further, not only to check that the bits we expect are there, but also that the bits that we expect not to be there to actually not be there! Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16765)
2021-12-10 20:15:10 +08:00
EVP_PKEY *tmp_pkey = NULL, *pkey = NULL;
Test that EC keys without a public key in them work as expected We create EC keys via both the "fromdata" and legacy key routes to make sure that they can be used without a public key. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13922)
2021-01-21 23:14:15 +08:00
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
if (!TEST_ptr(pctx = EVP_PKEY_CTX_new_from_name(testctx, keytype, testpropq)))
Test that EC keys without a public key in them work as expected We create EC keys via both the "fromdata" and legacy key routes to make sure that they can be used without a public key. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13922)
2021-01-21 23:14:15 +08:00
goto err;
Simplify the EVP_PKEY_XXX_fromdata_XX methods. The existing names such as EVP_PKEY_param_fromdata_settable were a bit confusing since the 'param' referred to key params not OSSL_PARAM. To simplify the interface a 'selection' parameter will be passed instead. The changes are: (1) EVP_PKEY_fromdata_init() replaces both EVP_PKEY_key_fromdata_init() and EVP_PKEY_param_fromdata_init(). (2) EVP_PKEY_fromdata() has an additional selection parameter. (3) EVP_PKEY_fromdata_settable() replaces EVP_PKEY_key_fromdata_settable() and EVP_PKEY_param_fromdata_settable(). EVP_PKEY_fromdata_settable() also uses a selection parameter. Fixes #12989 Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14076)
2021-02-05 11:55:50 +08:00
if (!TEST_int_gt(EVP_PKEY_fromdata_init(pctx), 0)
test/evp_extra_test.c: Refactor test_fromdata() test_fromdata() turns out to be a bit inflexible, so we split it into two functions, make_key_fromdata() and test_selection(), and adjust test_EVP_PKEY_ffc_priv_pub() and test_EC_priv_pub() accordingly. This allows us to check the resulting keys further, not only to check that the bits we expect are there, but also that the bits that we expect not to be there to actually not be there! Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16765)
2021-12-10 20:15:10 +08:00
|| !TEST_int_gt(EVP_PKEY_fromdata(pctx, &tmp_pkey, EVP_PKEY_KEYPAIR,
Simplify the EVP_PKEY_XXX_fromdata_XX methods. The existing names such as EVP_PKEY_param_fromdata_settable were a bit confusing since the 'param' referred to key params not OSSL_PARAM. To simplify the interface a 'selection' parameter will be passed instead. The changes are: (1) EVP_PKEY_fromdata_init() replaces both EVP_PKEY_key_fromdata_init() and EVP_PKEY_param_fromdata_init(). (2) EVP_PKEY_fromdata() has an additional selection parameter. (3) EVP_PKEY_fromdata_settable() replaces EVP_PKEY_key_fromdata_settable() and EVP_PKEY_param_fromdata_settable(). EVP_PKEY_fromdata_settable() also uses a selection parameter. Fixes #12989 Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14076)
2021-02-05 11:55:50 +08:00
params), 0))
Test that EC keys without a public key in them work as expected We create EC keys via both the "fromdata" and legacy key routes to make sure that they can be used without a public key. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13922)
2021-01-21 23:14:15 +08:00
goto err;
test/evp_extra_test.c: Refactor test_fromdata() test_fromdata() turns out to be a bit inflexible, so we split it into two functions, make_key_fromdata() and test_selection(), and adjust test_EVP_PKEY_ffc_priv_pub() and test_EC_priv_pub() accordingly. This allows us to check the resulting keys further, not only to check that the bits we expect are there, but also that the bits that we expect not to be there to actually not be there! Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16765)
2021-12-10 20:15:10 +08:00
if (!TEST_ptr(tmp_pkey))
Test that EC keys without a public key in them work as expected We create EC keys via both the "fromdata" and legacy key routes to make sure that they can be used without a public key. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13922)
2021-01-21 23:14:15 +08:00
goto err;
test/evp_extra_test.c: Refactor test_fromdata() test_fromdata() turns out to be a bit inflexible, so we split it into two functions, make_key_fromdata() and test_selection(), and adjust test_EVP_PKEY_ffc_priv_pub() and test_EC_priv_pub() accordingly. This allows us to check the resulting keys further, not only to check that the bits we expect are there, but also that the bits that we expect not to be there to actually not be there! Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16765)
2021-12-10 20:15:10 +08:00
pkey = tmp_pkey;
tmp_pkey = NULL;
err:
EVP_PKEY_free(tmp_pkey);
EVP_PKEY_CTX_free(pctx);
return pkey;
}
static int test_selection(EVP_PKEY *pkey, int selection)
{
int testresult = 0;
int ret;
BIO *bio = BIO_new(BIO_s_mem());
Test that a key is usable after an EVP_PKEY_fromdata call Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16911)
2021-10-25 21:34:38 +08:00
ret = PEM_write_bio_PUBKEY(bio, pkey);
if ((selection & OSSL_KEYMGMT_SELECT_PUBLIC_KEY) != 0) {
if (!TEST_true(ret))
goto err;
} else {
if (!TEST_false(ret))
goto err;
}
ret = PEM_write_bio_PrivateKey_ex(bio, pkey, NULL, NULL, 0, NULL, NULL,
testctx, NULL);
if ((selection & OSSL_KEYMGMT_SELECT_PRIVATE_KEY) != 0) {
if (!TEST_true(ret))
goto err;
} else {
if (!TEST_false(ret))
goto err;
}
Test that EC keys without a public key in them work as expected We create EC keys via both the "fromdata" and legacy key routes to make sure that they can be used without a public key. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13922)
2021-01-21 23:14:15 +08:00
testresult = 1;
err:
Test that a key is usable after an EVP_PKEY_fromdata call Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16911)
2021-10-25 21:34:38 +08:00
BIO_free(bio);
Test that EC keys without a public key in them work as expected We create EC keys via both the "fromdata" and legacy key routes to make sure that they can be used without a public key. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13922)
2021-01-21 23:14:15 +08:00
return testresult;
}
#endif /* !OPENSSL_NO_DH || !OPENSSL_NO_DSA || !OPENSSL_NO_EC */
Allow EVP_PKEY private key objects to be created without a public component Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/13855)
2021-01-13 12:10:51 +08:00
/*
Fix no-dh and no-dsa Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13915)
2021-01-21 17:19:16 +08:00
* Test combinations of private, public, missing and private + public key
Allow EVP_PKEY private key objects to be created without a public component Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/13855)
2021-01-13 12:10:51 +08:00
* params to ensure they are all accepted
*/
Test that EC keys without a public key in them work as expected We create EC keys via both the "fromdata" and legacy key routes to make sure that they can be used without a public key. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13922)
2021-01-21 23:14:15 +08:00
#if !defined(OPENSSL_NO_DH) || !defined(OPENSSL_NO_DSA)
Allow EVP_PKEY private key objects to be created without a public component Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/13855)
2021-01-13 12:10:51 +08:00
static int test_EVP_PKEY_ffc_priv_pub(char *keytype)
{
OSSL_PARAM_BLD *bld = NULL;
OSSL_PARAM *params = NULL;
test/evp_extra_test.c: Refactor test_fromdata() test_fromdata() turns out to be a bit inflexible, so we split it into two functions, make_key_fromdata() and test_selection(), and adjust test_EVP_PKEY_ffc_priv_pub() and test_EC_priv_pub() accordingly. This allows us to check the resulting keys further, not only to check that the bits we expect are there, but also that the bits that we expect not to be there to actually not be there! Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16765)
2021-12-10 20:15:10 +08:00
EVP_PKEY *just_params = NULL;
EVP_PKEY *params_and_priv = NULL;
EVP_PKEY *params_and_pub = NULL;
EVP_PKEY *params_and_keypair = NULL;
Allow EVP_PKEY private key objects to be created without a public component Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/13855)
2021-01-13 12:10:51 +08:00
BIGNUM *p = NULL, *q = NULL, *g = NULL, *pub = NULL, *priv = NULL;
int ret = 0;
/*
* Setup the parameters for our pkey object. For our purposes they don't
* have to actually be *valid* parameters. We just need to set something.
*/
Test that EC keys without a public key in them work as expected We create EC keys via both the "fromdata" and legacy key routes to make sure that they can be used without a public key. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13922)
2021-01-21 23:14:15 +08:00
if (!TEST_ptr(p = BN_new())
Allow EVP_PKEY private key objects to be created without a public component Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/13855)
2021-01-13 12:10:51 +08:00
|| !TEST_ptr(q = BN_new())
|| !TEST_ptr(g = BN_new())
|| !TEST_ptr(pub = BN_new())
|| !TEST_ptr(priv = BN_new()))
goto err;
/* Test !priv and !pub */
if (!TEST_ptr(bld = OSSL_PARAM_BLD_new())
|| !TEST_true(OSSL_PARAM_BLD_push_BN(bld, OSSL_PKEY_PARAM_FFC_P, p))
|| !TEST_true(OSSL_PARAM_BLD_push_BN(bld, OSSL_PKEY_PARAM_FFC_Q, q))
|| !TEST_true(OSSL_PARAM_BLD_push_BN(bld, OSSL_PKEY_PARAM_FFC_G, g)))
goto err;
test/evp_extra_test.c: Refactor test_fromdata() test_fromdata() turns out to be a bit inflexible, so we split it into two functions, make_key_fromdata() and test_selection(), and adjust test_EVP_PKEY_ffc_priv_pub() and test_EC_priv_pub() accordingly. This allows us to check the resulting keys further, not only to check that the bits we expect are there, but also that the bits that we expect not to be there to actually not be there! Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16765)
2021-12-10 20:15:10 +08:00
if (!TEST_ptr(params = OSSL_PARAM_BLD_to_param(bld))
|| !TEST_ptr(just_params = make_key_fromdata(keytype, params)))
Allow EVP_PKEY private key objects to be created without a public component Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/13855)
2021-01-13 12:10:51 +08:00
goto err;
Replace OSSL_PARAM_BLD_free_params() with OSSL_PARAM_free(). Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14785)
2021-04-07 11:45:19 +08:00
OSSL_PARAM_free(params);
Allow EVP_PKEY private key objects to be created without a public component Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/13855)
2021-01-13 12:10:51 +08:00
OSSL_PARAM_BLD_free(bld);
test/evp_extra_test.c: Refactor test_fromdata() test_fromdata() turns out to be a bit inflexible, so we split it into two functions, make_key_fromdata() and test_selection(), and adjust test_EVP_PKEY_ffc_priv_pub() and test_EC_priv_pub() accordingly. This allows us to check the resulting keys further, not only to check that the bits we expect are there, but also that the bits that we expect not to be there to actually not be there! Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16765)
2021-12-10 20:15:10 +08:00
params = NULL;
bld = NULL;
if (!test_selection(just_params, OSSL_KEYMGMT_SELECT_ALL_PARAMETERS)
|| test_selection(just_params, OSSL_KEYMGMT_SELECT_KEYPAIR))
goto err;
Allow EVP_PKEY private key objects to be created without a public component Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/13855)
2021-01-13 12:10:51 +08:00
/* Test priv and !pub */
if (!TEST_ptr(bld = OSSL_PARAM_BLD_new())
|| !TEST_true(OSSL_PARAM_BLD_push_BN(bld, OSSL_PKEY_PARAM_FFC_P, p))
|| !TEST_true(OSSL_PARAM_BLD_push_BN(bld, OSSL_PKEY_PARAM_FFC_Q, q))
|| !TEST_true(OSSL_PARAM_BLD_push_BN(bld, OSSL_PKEY_PARAM_FFC_G, g))
|| !TEST_true(OSSL_PARAM_BLD_push_BN(bld, OSSL_PKEY_PARAM_PRIV_KEY,
priv)))
goto err;
test/evp_extra_test.c: Refactor test_fromdata() test_fromdata() turns out to be a bit inflexible, so we split it into two functions, make_key_fromdata() and test_selection(), and adjust test_EVP_PKEY_ffc_priv_pub() and test_EC_priv_pub() accordingly. This allows us to check the resulting keys further, not only to check that the bits we expect are there, but also that the bits that we expect not to be there to actually not be there! Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16765)
2021-12-10 20:15:10 +08:00
if (!TEST_ptr(params = OSSL_PARAM_BLD_to_param(bld))
|| !TEST_ptr(params_and_priv = make_key_fromdata(keytype, params)))
Allow EVP_PKEY private key objects to be created without a public component Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/13855)
2021-01-13 12:10:51 +08:00
goto err;
Replace OSSL_PARAM_BLD_free_params() with OSSL_PARAM_free(). Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14785)
2021-04-07 11:45:19 +08:00
OSSL_PARAM_free(params);
Allow EVP_PKEY private key objects to be created without a public component Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/13855)
2021-01-13 12:10:51 +08:00
OSSL_PARAM_BLD_free(bld);
test/evp_extra_test.c: Refactor test_fromdata() test_fromdata() turns out to be a bit inflexible, so we split it into two functions, make_key_fromdata() and test_selection(), and adjust test_EVP_PKEY_ffc_priv_pub() and test_EC_priv_pub() accordingly. This allows us to check the resulting keys further, not only to check that the bits we expect are there, but also that the bits that we expect not to be there to actually not be there! Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16765)
2021-12-10 20:15:10 +08:00
params = NULL;
bld = NULL;
if (!test_selection(params_and_priv, OSSL_KEYMGMT_SELECT_PRIVATE_KEY)
|| test_selection(params_and_priv, OSSL_KEYMGMT_SELECT_PUBLIC_KEY))
goto err;
Allow EVP_PKEY private key objects to be created without a public component Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/13855)
2021-01-13 12:10:51 +08:00
/* Test !priv and pub */
if (!TEST_ptr(bld = OSSL_PARAM_BLD_new())
|| !TEST_true(OSSL_PARAM_BLD_push_BN(bld, OSSL_PKEY_PARAM_FFC_P, p))
|| !TEST_true(OSSL_PARAM_BLD_push_BN(bld, OSSL_PKEY_PARAM_FFC_Q, q))
|| !TEST_true(OSSL_PARAM_BLD_push_BN(bld, OSSL_PKEY_PARAM_FFC_G, g))
|| !TEST_true(OSSL_PARAM_BLD_push_BN(bld, OSSL_PKEY_PARAM_PUB_KEY,
pub)))
goto err;
test/evp_extra_test.c: Refactor test_fromdata() test_fromdata() turns out to be a bit inflexible, so we split it into two functions, make_key_fromdata() and test_selection(), and adjust test_EVP_PKEY_ffc_priv_pub() and test_EC_priv_pub() accordingly. This allows us to check the resulting keys further, not only to check that the bits we expect are there, but also that the bits that we expect not to be there to actually not be there! Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16765)
2021-12-10 20:15:10 +08:00
if (!TEST_ptr(params = OSSL_PARAM_BLD_to_param(bld))
|| !TEST_ptr(params_and_pub = make_key_fromdata(keytype, params)))
Allow EVP_PKEY private key objects to be created without a public component Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/13855)
2021-01-13 12:10:51 +08:00
goto err;
Replace OSSL_PARAM_BLD_free_params() with OSSL_PARAM_free(). Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14785)
2021-04-07 11:45:19 +08:00
OSSL_PARAM_free(params);
Allow EVP_PKEY private key objects to be created without a public component Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/13855)
2021-01-13 12:10:51 +08:00
OSSL_PARAM_BLD_free(bld);
test/evp_extra_test.c: Refactor test_fromdata() test_fromdata() turns out to be a bit inflexible, so we split it into two functions, make_key_fromdata() and test_selection(), and adjust test_EVP_PKEY_ffc_priv_pub() and test_EC_priv_pub() accordingly. This allows us to check the resulting keys further, not only to check that the bits we expect are there, but also that the bits that we expect not to be there to actually not be there! Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16765)
2021-12-10 20:15:10 +08:00
params = NULL;
bld = NULL;
if (!test_selection(params_and_pub, OSSL_KEYMGMT_SELECT_PUBLIC_KEY)
|| test_selection(params_and_pub, OSSL_KEYMGMT_SELECT_PRIVATE_KEY))
goto err;
Allow EVP_PKEY private key objects to be created without a public component Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/13855)
2021-01-13 12:10:51 +08:00
/* Test priv and pub */
if (!TEST_ptr(bld = OSSL_PARAM_BLD_new())
|| !TEST_true(OSSL_PARAM_BLD_push_BN(bld, OSSL_PKEY_PARAM_FFC_P, p))
|| !TEST_true(OSSL_PARAM_BLD_push_BN(bld, OSSL_PKEY_PARAM_FFC_Q, q))
|| !TEST_true(OSSL_PARAM_BLD_push_BN(bld, OSSL_PKEY_PARAM_FFC_G, g))
|| !TEST_true(OSSL_PARAM_BLD_push_BN(bld, OSSL_PKEY_PARAM_PUB_KEY,
pub))
|| !TEST_true(OSSL_PARAM_BLD_push_BN(bld, OSSL_PKEY_PARAM_PRIV_KEY,
priv)))
goto err;
test/evp_extra_test.c: Refactor test_fromdata() test_fromdata() turns out to be a bit inflexible, so we split it into two functions, make_key_fromdata() and test_selection(), and adjust test_EVP_PKEY_ffc_priv_pub() and test_EC_priv_pub() accordingly. This allows us to check the resulting keys further, not only to check that the bits we expect are there, but also that the bits that we expect not to be there to actually not be there! Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16765)
2021-12-10 20:15:10 +08:00
if (!TEST_ptr(params = OSSL_PARAM_BLD_to_param(bld))
|| !TEST_ptr(params_and_keypair = make_key_fromdata(keytype, params)))
Allow EVP_PKEY private key objects to be created without a public component Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/13855)
2021-01-13 12:10:51 +08:00
goto err;
test/evp_extra_test.c: Refactor test_fromdata() test_fromdata() turns out to be a bit inflexible, so we split it into two functions, make_key_fromdata() and test_selection(), and adjust test_EVP_PKEY_ffc_priv_pub() and test_EC_priv_pub() accordingly. This allows us to check the resulting keys further, not only to check that the bits we expect are there, but also that the bits that we expect not to be there to actually not be there! Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16765)
2021-12-10 20:15:10 +08:00
if (!test_selection(params_and_keypair, EVP_PKEY_KEYPAIR))
Allow EVP_PKEY private key objects to be created without a public component Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/13855)
2021-01-13 12:10:51 +08:00
goto err;
ret = 1;
err:
Replace OSSL_PARAM_BLD_free_params() with OSSL_PARAM_free(). Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14785)
2021-04-07 11:45:19 +08:00
OSSL_PARAM_free(params);
Allow EVP_PKEY private key objects to be created without a public component Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/13855)
2021-01-13 12:10:51 +08:00
OSSL_PARAM_BLD_free(bld);
test/evp_extra_test.c: Refactor test_fromdata() test_fromdata() turns out to be a bit inflexible, so we split it into two functions, make_key_fromdata() and test_selection(), and adjust test_EVP_PKEY_ffc_priv_pub() and test_EC_priv_pub() accordingly. This allows us to check the resulting keys further, not only to check that the bits we expect are there, but also that the bits that we expect not to be there to actually not be there! Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16765)
2021-12-10 20:15:10 +08:00
EVP_PKEY_free(just_params);
EVP_PKEY_free(params_and_priv);
EVP_PKEY_free(params_and_pub);
EVP_PKEY_free(params_and_keypair);
Allow EVP_PKEY private key objects to be created without a public component Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/13855)
2021-01-13 12:10:51 +08:00
BN_free(p);
BN_free(q);
BN_free(g);
BN_free(pub);
BN_free(priv);
return ret;
}
Fix no-dh and no-dsa Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13915)
2021-01-21 17:19:16 +08:00
#endif /* !OPENSSL_NO_DH || !OPENSSL_NO_DSA */
Allow EVP_PKEY private key objects to be created without a public component Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/13855)
2021-01-13 12:10:51 +08:00
Test that EC keys without a public key in them work as expected We create EC keys via both the "fromdata" and legacy key routes to make sure that they can be used without a public key. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13922)
2021-01-21 23:14:15 +08:00
/*
* Test combinations of private, public, missing and private + public key
* params to ensure they are all accepted for EC keys
*/
#ifndef OPENSSL_NO_EC
static unsigned char ec_priv[] = {
0xe9, 0x25, 0xf7, 0x66, 0x58, 0xa4, 0xdd, 0x99, 0x61, 0xe7, 0xe8, 0x23,
0x85, 0xc2, 0xe8, 0x33, 0x27, 0xc5, 0x5c, 0xeb, 0xdb, 0x43, 0x9f, 0xd5,
0xf2, 0x5a, 0x75, 0x55, 0xd0, 0x2e, 0x6d, 0x16
};
static unsigned char ec_pub[] = {
0x04, 0xad, 0x11, 0x90, 0x77, 0x4b, 0x46, 0xee, 0x72, 0x51, 0x15, 0x97,
0x4a, 0x6a, 0xa7, 0xaf, 0x59, 0xfa, 0x4b, 0xf2, 0x41, 0xc8, 0x3a, 0x81,
0x23, 0xb6, 0x90, 0x04, 0x6c, 0x67, 0x66, 0xd0, 0xdc, 0xf2, 0x15, 0x1d,
0x41, 0x61, 0xb7, 0x95, 0x85, 0x38, 0x5a, 0x84, 0x56, 0xe8, 0xb3, 0x0e,
0xf5, 0xc6, 0x5d, 0xa4, 0x54, 0x26, 0xb0, 0xf7, 0xa5, 0x4a, 0x33, 0xf1,
0x08, 0x09, 0xb8, 0xdb, 0x03
};
static int test_EC_priv_pub(void)
{
OSSL_PARAM_BLD *bld = NULL;
OSSL_PARAM *params = NULL;
test/evp_extra_test.c: Refactor test_fromdata() test_fromdata() turns out to be a bit inflexible, so we split it into two functions, make_key_fromdata() and test_selection(), and adjust test_EVP_PKEY_ffc_priv_pub() and test_EC_priv_pub() accordingly. This allows us to check the resulting keys further, not only to check that the bits we expect are there, but also that the bits that we expect not to be there to actually not be there! Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16765)
2021-12-10 20:15:10 +08:00
EVP_PKEY *just_params = NULL;
EVP_PKEY *params_and_priv = NULL;
EVP_PKEY *params_and_pub = NULL;
EVP_PKEY *params_and_keypair = NULL;
Test that EC keys without a public key in them work as expected We create EC keys via both the "fromdata" and legacy key routes to make sure that they can be used without a public key. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13922)
2021-01-21 23:14:15 +08:00
BIGNUM *priv = NULL;
int ret = 0;
Add testcases for EVP_PKEY_get1_encoded_public_key Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/18902)
2022-07-28 20:13:06 +08:00
unsigned char *encoded = NULL;
provider: return error if buf too small when getting ec pubkey param Fixes #20889 There was an incorrect value passed to EC_POINT_point2oct() for the buffer size of the param passed-in. Added testcases. Signed-off-by: Yi Li <yi1.li@intel.com> Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/20890)
2023-05-05 11:30:05 +08:00
size_t len = 0;
unsigned char buffer[128];
Test that EC keys without a public key in them work as expected We create EC keys via both the "fromdata" and legacy key routes to make sure that they can be used without a public key. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13922)
2021-01-21 23:14:15 +08:00
/*
* Setup the parameters for our pkey object. For our purposes they don't
* have to actually be *valid* parameters. We just need to set something.
*/
if (!TEST_ptr(priv = BN_bin2bn(ec_priv, sizeof(ec_priv), NULL)))
goto err;
/* Test !priv and !pub */
if (!TEST_ptr(bld = OSSL_PARAM_BLD_new())
|| !TEST_true(OSSL_PARAM_BLD_push_utf8_string(bld,
OSSL_PKEY_PARAM_GROUP_NAME,
"P-256", 0)))
goto err;
test/evp_extra_test.c: Refactor test_fromdata() test_fromdata() turns out to be a bit inflexible, so we split it into two functions, make_key_fromdata() and test_selection(), and adjust test_EVP_PKEY_ffc_priv_pub() and test_EC_priv_pub() accordingly. This allows us to check the resulting keys further, not only to check that the bits we expect are there, but also that the bits that we expect not to be there to actually not be there! Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16765)
2021-12-10 20:15:10 +08:00
if (!TEST_ptr(params = OSSL_PARAM_BLD_to_param(bld))
|| !TEST_ptr(just_params = make_key_fromdata("EC", params)))
Test that EC keys without a public key in them work as expected We create EC keys via both the "fromdata" and legacy key routes to make sure that they can be used without a public key. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13922)
2021-01-21 23:14:15 +08:00
goto err;
Replace OSSL_PARAM_BLD_free_params() with OSSL_PARAM_free(). Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14785)
2021-04-07 11:45:19 +08:00
OSSL_PARAM_free(params);
Test that EC keys without a public key in them work as expected We create EC keys via both the "fromdata" and legacy key routes to make sure that they can be used without a public key. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13922)
2021-01-21 23:14:15 +08:00
OSSL_PARAM_BLD_free(bld);
test/evp_extra_test.c: Refactor test_fromdata() test_fromdata() turns out to be a bit inflexible, so we split it into two functions, make_key_fromdata() and test_selection(), and adjust test_EVP_PKEY_ffc_priv_pub() and test_EC_priv_pub() accordingly. This allows us to check the resulting keys further, not only to check that the bits we expect are there, but also that the bits that we expect not to be there to actually not be there! Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16765)
2021-12-10 20:15:10 +08:00
params = NULL;
bld = NULL;
if (!test_selection(just_params, OSSL_KEYMGMT_SELECT_ALL_PARAMETERS)
|| test_selection(just_params, OSSL_KEYMGMT_SELECT_KEYPAIR))
goto err;
Test that EC keys without a public key in them work as expected We create EC keys via both the "fromdata" and legacy key routes to make sure that they can be used without a public key. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13922)
2021-01-21 23:14:15 +08:00
/* Test priv and !pub */
if (!TEST_ptr(bld = OSSL_PARAM_BLD_new())
|| !TEST_true(OSSL_PARAM_BLD_push_utf8_string(bld,
OSSL_PKEY_PARAM_GROUP_NAME,
"P-256", 0))
|| !TEST_true(OSSL_PARAM_BLD_push_BN(bld, OSSL_PKEY_PARAM_PRIV_KEY,
priv)))
goto err;
test/evp_extra_test.c: Refactor test_fromdata() test_fromdata() turns out to be a bit inflexible, so we split it into two functions, make_key_fromdata() and test_selection(), and adjust test_EVP_PKEY_ffc_priv_pub() and test_EC_priv_pub() accordingly. This allows us to check the resulting keys further, not only to check that the bits we expect are there, but also that the bits that we expect not to be there to actually not be there! Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16765)
2021-12-10 20:15:10 +08:00
if (!TEST_ptr(params = OSSL_PARAM_BLD_to_param(bld))
|| !TEST_ptr(params_and_priv = make_key_fromdata("EC", params)))
Test that EC keys without a public key in them work as expected We create EC keys via both the "fromdata" and legacy key routes to make sure that they can be used without a public key. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13922)
2021-01-21 23:14:15 +08:00
goto err;
test/evp_extra_test.c: Refactor test_fromdata() test_fromdata() turns out to be a bit inflexible, so we split it into two functions, make_key_fromdata() and test_selection(), and adjust test_EVP_PKEY_ffc_priv_pub() and test_EC_priv_pub() accordingly. This allows us to check the resulting keys further, not only to check that the bits we expect are there, but also that the bits that we expect not to be there to actually not be there! Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16765)
2021-12-10 20:15:10 +08:00
OSSL_PARAM_free(params);
OSSL_PARAM_BLD_free(bld);
params = NULL;
bld = NULL;
Test that a key is usable after an EVP_PKEY_fromdata call Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16911)
2021-10-25 21:34:38 +08:00
/*
test/evp_extra_test.c: Refactor test_fromdata() test_fromdata() turns out to be a bit inflexible, so we split it into two functions, make_key_fromdata() and test_selection(), and adjust test_EVP_PKEY_ffc_priv_pub() and test_EC_priv_pub() accordingly. This allows us to check the resulting keys further, not only to check that the bits we expect are there, but also that the bits that we expect not to be there to actually not be there! Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16765)
2021-12-10 20:15:10 +08:00
* We indicate only parameters here, in spite of having built a key that
* has a private part, because the PEM_write_bio_PrivateKey_ex call is
* expected to fail because it does not support exporting a private EC
* key without a corresponding public key
Test that a key is usable after an EVP_PKEY_fromdata call Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16911)
2021-10-25 21:34:38 +08:00
*/
test/evp_extra_test.c: Refactor test_fromdata() test_fromdata() turns out to be a bit inflexible, so we split it into two functions, make_key_fromdata() and test_selection(), and adjust test_EVP_PKEY_ffc_priv_pub() and test_EC_priv_pub() accordingly. This allows us to check the resulting keys further, not only to check that the bits we expect are there, but also that the bits that we expect not to be there to actually not be there! Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16765)
2021-12-10 20:15:10 +08:00
if (!test_selection(params_and_priv, OSSL_KEYMGMT_SELECT_ALL_PARAMETERS)
|| test_selection(params_and_priv, OSSL_KEYMGMT_SELECT_PUBLIC_KEY))
Test that EC keys without a public key in them work as expected We create EC keys via both the "fromdata" and legacy key routes to make sure that they can be used without a public key. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13922)
2021-01-21 23:14:15 +08:00
goto err;
/* Test !priv and pub */
if (!TEST_ptr(bld = OSSL_PARAM_BLD_new())
|| !TEST_true(OSSL_PARAM_BLD_push_utf8_string(bld,
OSSL_PKEY_PARAM_GROUP_NAME,
"P-256", 0))
|| !TEST_true(OSSL_PARAM_BLD_push_octet_string(bld,
OSSL_PKEY_PARAM_PUB_KEY,
ec_pub, sizeof(ec_pub))))
goto err;
test/evp_extra_test.c: Refactor test_fromdata() test_fromdata() turns out to be a bit inflexible, so we split it into two functions, make_key_fromdata() and test_selection(), and adjust test_EVP_PKEY_ffc_priv_pub() and test_EC_priv_pub() accordingly. This allows us to check the resulting keys further, not only to check that the bits we expect are there, but also that the bits that we expect not to be there to actually not be there! Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16765)
2021-12-10 20:15:10 +08:00
if (!TEST_ptr(params = OSSL_PARAM_BLD_to_param(bld))
|| !TEST_ptr(params_and_pub = make_key_fromdata("EC", params)))
Test that EC keys without a public key in them work as expected We create EC keys via both the "fromdata" and legacy key routes to make sure that they can be used without a public key. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13922)
2021-01-21 23:14:15 +08:00
goto err;
Replace OSSL_PARAM_BLD_free_params() with OSSL_PARAM_free(). Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14785)
2021-04-07 11:45:19 +08:00
OSSL_PARAM_free(params);
Test that EC keys without a public key in them work as expected We create EC keys via both the "fromdata" and legacy key routes to make sure that they can be used without a public key. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13922)
2021-01-21 23:14:15 +08:00
OSSL_PARAM_BLD_free(bld);
test/evp_extra_test.c: Refactor test_fromdata() test_fromdata() turns out to be a bit inflexible, so we split it into two functions, make_key_fromdata() and test_selection(), and adjust test_EVP_PKEY_ffc_priv_pub() and test_EC_priv_pub() accordingly. This allows us to check the resulting keys further, not only to check that the bits we expect are there, but also that the bits that we expect not to be there to actually not be there! Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16765)
2021-12-10 20:15:10 +08:00
params = NULL;
bld = NULL;
if (!test_selection(params_and_pub, OSSL_KEYMGMT_SELECT_PUBLIC_KEY)
|| test_selection(params_and_pub, OSSL_KEYMGMT_SELECT_PRIVATE_KEY))
goto err;
Test that EC keys without a public key in them work as expected We create EC keys via both the "fromdata" and legacy key routes to make sure that they can be used without a public key. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13922)
2021-01-21 23:14:15 +08:00
/* Test priv and pub */
if (!TEST_ptr(bld = OSSL_PARAM_BLD_new())
|| !TEST_true(OSSL_PARAM_BLD_push_utf8_string(bld,
OSSL_PKEY_PARAM_GROUP_NAME,
"P-256", 0))
|| !TEST_true(OSSL_PARAM_BLD_push_octet_string(bld,
OSSL_PKEY_PARAM_PUB_KEY,
ec_pub, sizeof(ec_pub)))
Fix a copy&paste error in evp_extra_test test_EC_priv_pub fails to test the case where both a private and public key have been supplied. Fixes #14349 Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14351)
2021-03-01 18:48:59 +08:00
|| !TEST_true(OSSL_PARAM_BLD_push_BN(bld, OSSL_PKEY_PARAM_PRIV_KEY,
priv)))
Test that EC keys without a public key in them work as expected We create EC keys via both the "fromdata" and legacy key routes to make sure that they can be used without a public key. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13922)
2021-01-21 23:14:15 +08:00
goto err;
test/evp_extra_test.c: Refactor test_fromdata() test_fromdata() turns out to be a bit inflexible, so we split it into two functions, make_key_fromdata() and test_selection(), and adjust test_EVP_PKEY_ffc_priv_pub() and test_EC_priv_pub() accordingly. This allows us to check the resulting keys further, not only to check that the bits we expect are there, but also that the bits that we expect not to be there to actually not be there! Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16765)
2021-12-10 20:15:10 +08:00
if (!TEST_ptr(params = OSSL_PARAM_BLD_to_param(bld))
|| !TEST_ptr(params_and_keypair = make_key_fromdata("EC", params)))
Test that EC keys without a public key in them work as expected We create EC keys via both the "fromdata" and legacy key routes to make sure that they can be used without a public key. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13922)
2021-01-21 23:14:15 +08:00
goto err;
test/evp_extra_test.c: Refactor test_fromdata() test_fromdata() turns out to be a bit inflexible, so we split it into two functions, make_key_fromdata() and test_selection(), and adjust test_EVP_PKEY_ffc_priv_pub() and test_EC_priv_pub() accordingly. This allows us to check the resulting keys further, not only to check that the bits we expect are there, but also that the bits that we expect not to be there to actually not be there! Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16765)
2021-12-10 20:15:10 +08:00
if (!test_selection(params_and_keypair, EVP_PKEY_KEYPAIR))
Test that EC keys without a public key in them work as expected We create EC keys via both the "fromdata" and legacy key routes to make sure that they can be used without a public key. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13922)
2021-01-21 23:14:15 +08:00
goto err;
test/evp_extra_test.c: Add EVP_PKEY comparisons in test_EC_priv_pub() Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16765)
2021-12-10 20:18:42 +08:00
/* Try key equality */
if (!TEST_int_gt(EVP_PKEY_parameters_eq(just_params, just_params), 0)
|| !TEST_int_gt(EVP_PKEY_parameters_eq(just_params, params_and_pub),
0)
|| !TEST_int_gt(EVP_PKEY_parameters_eq(just_params, params_and_priv),
0)
|| !TEST_int_gt(EVP_PKEY_parameters_eq(just_params, params_and_keypair),
0)
|| !TEST_int_gt(EVP_PKEY_eq(params_and_pub, params_and_pub), 0)
|| !TEST_int_gt(EVP_PKEY_eq(params_and_priv, params_and_priv), 0)
|| !TEST_int_gt(EVP_PKEY_eq(params_and_keypair, params_and_pub), 0)
|| !TEST_int_gt(EVP_PKEY_eq(params_and_keypair, params_and_priv), 0))
goto err;
Add testcases for EVP_PKEY_get1_encoded_public_key Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/18902)
2022-07-28 20:13:06 +08:00
/* Positive and negative testcase for EVP_PKEY_get1_encoded_public_key */
if (!TEST_int_gt(EVP_PKEY_get1_encoded_public_key(params_and_pub, &encoded), 0))
goto err;
OPENSSL_free(encoded);
encoded = NULL;
if (!TEST_int_eq(EVP_PKEY_get1_encoded_public_key(just_params, &encoded), 0)) {
OPENSSL_free(encoded);
encoded = NULL;
goto err;
}
provider: return error if buf too small when getting ec pubkey param Fixes #20889 There was an incorrect value passed to EC_POINT_point2oct() for the buffer size of the param passed-in. Added testcases. Signed-off-by: Yi Li <yi1.li@intel.com> Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/20890)
2023-05-05 11:30:05 +08:00
/* Positive and negative testcase for EVP_PKEY_get_octet_string_param */
if (!TEST_int_eq(EVP_PKEY_get_octet_string_param(params_and_pub,
OSSL_PKEY_PARAM_ENCODED_PUBLIC_KEY,
buffer, sizeof(buffer), &len), 1)
|| !TEST_int_eq(len, 65))
goto err;
len = 0;
if (!TEST_int_eq(EVP_PKEY_get_octet_string_param(params_and_pub,
OSSL_PKEY_PARAM_ENCODED_PUBLIC_KEY,
NULL, 0, &len), 1)
|| !TEST_int_eq(len, 65))
goto err;
/* too-short buffer len*/
if (!TEST_int_eq(EVP_PKEY_get_octet_string_param(params_and_pub,
OSSL_PKEY_PARAM_ENCODED_PUBLIC_KEY,
buffer, 10, &len), 0))
goto err;
Test that EC keys without a public key in them work as expected We create EC keys via both the "fromdata" and legacy key routes to make sure that they can be used without a public key. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13922)
2021-01-21 23:14:15 +08:00
ret = 1;
err:
Replace OSSL_PARAM_BLD_free_params() with OSSL_PARAM_free(). Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14785)
2021-04-07 11:45:19 +08:00
OSSL_PARAM_free(params);
Test that EC keys without a public key in them work as expected We create EC keys via both the "fromdata" and legacy key routes to make sure that they can be used without a public key. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13922)
2021-01-21 23:14:15 +08:00
OSSL_PARAM_BLD_free(bld);
test/evp_extra_test.c: Refactor test_fromdata() test_fromdata() turns out to be a bit inflexible, so we split it into two functions, make_key_fromdata() and test_selection(), and adjust test_EVP_PKEY_ffc_priv_pub() and test_EC_priv_pub() accordingly. This allows us to check the resulting keys further, not only to check that the bits we expect are there, but also that the bits that we expect not to be there to actually not be there! Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16765)
2021-12-10 20:15:10 +08:00
EVP_PKEY_free(just_params);
EVP_PKEY_free(params_and_priv);
EVP_PKEY_free(params_and_pub);
EVP_PKEY_free(params_and_keypair);
Test that EC keys without a public key in them work as expected We create EC keys via both the "fromdata" and legacy key routes to make sure that they can be used without a public key. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13922)
2021-01-21 23:14:15 +08:00
BN_free(priv);
return ret;
}
test: evp_extra: EC, read affine coordinates Add a test to read the EC X,Y coordinates. Support legacy keys. Signed-off-by: Jorge Ramirez-Ortiz <jorge@foundries.io> Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/20535)
2023-03-20 19:51:55 +08:00
/* Also test that we can read the EC PUB affine coordinates */
static int test_evp_get_ec_pub(void)
{
OSSL_PARAM_BLD *bld = NULL;
OSSL_PARAM *params = NULL;
unsigned char *pad = NULL;
EVP_PKEY *keypair = NULL;
BIGNUM *priv = NULL;
BIGNUM *x = NULL;
BIGNUM *y = NULL;
int ret = 0;
if (!TEST_ptr(priv = BN_bin2bn(ec_priv, sizeof(ec_priv), NULL)))
goto err;
if (!TEST_ptr(bld = OSSL_PARAM_BLD_new())
|| !TEST_true(OSSL_PARAM_BLD_push_utf8_string(bld,
OSSL_PKEY_PARAM_GROUP_NAME,
"P-256", 0))
|| !TEST_true(OSSL_PARAM_BLD_push_octet_string(bld,
OSSL_PKEY_PARAM_PUB_KEY,
ec_pub, sizeof(ec_pub)))
|| !TEST_true(OSSL_PARAM_BLD_push_BN(bld, OSSL_PKEY_PARAM_PRIV_KEY,
priv)))
goto err;
if (!TEST_ptr(params = OSSL_PARAM_BLD_to_param(bld))
|| !TEST_ptr(keypair = make_key_fromdata("EC", params)))
goto err;
if (!test_selection(keypair, EVP_PKEY_KEYPAIR))
goto err;
if (!EVP_PKEY_get_bn_param(keypair, OSSL_PKEY_PARAM_EC_PUB_X, &x)
|| !EVP_PKEY_get_bn_param(keypair, OSSL_PKEY_PARAM_EC_PUB_Y, &y))
goto err;
if (!TEST_ptr(pad = OPENSSL_zalloc(sizeof(ec_pub))))
goto err;
pad[0] = ec_pub[0];
BN_bn2bin(x, &pad[1]);
BN_bn2bin(y, &pad[33]);
if (!TEST_true(memcmp(ec_pub, pad, sizeof(ec_pub)) == 0))
goto err;
ret = 1;
err:
OSSL_PARAM_free(params);
OSSL_PARAM_BLD_free(bld);
EVP_PKEY_free(keypair);
OPENSSL_free(pad);
BN_free(priv);
BN_free(x);
BN_free(y);
return ret;
}
Test that EC keys without a public key in them work as expected We create EC keys via both the "fromdata" and legacy key routes to make sure that they can be used without a public key. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13922)
2021-01-21 23:14:15 +08:00
/* Test that using a legacy EC key with only a private key in it works */
# ifndef OPENSSL_NO_DEPRECATED_3_0
static int test_EC_priv_only_legacy(void)
{
BIGNUM *priv = NULL;
int ret = 0;
EC_KEY *eckey = NULL;
Implement EVP_PKEY_dup() function Fixes #14501 Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14624)
2021-03-20 01:45:43 +08:00
EVP_PKEY *pkey = NULL, *dup_pk = NULL;
Test that EC keys without a public key in them work as expected We create EC keys via both the "fromdata" and legacy key routes to make sure that they can be used without a public key. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13922)
2021-01-21 23:14:15 +08:00
EVP_MD_CTX *ctx = NULL;
/* Create the low level EC_KEY */
if (!TEST_ptr(priv = BN_bin2bn(ec_priv, sizeof(ec_priv), NULL)))
goto err;
eckey = EC_KEY_new_by_curve_name(NID_X9_62_prime256v1);
if (!TEST_ptr(eckey))
goto err;
if (!TEST_true(EC_KEY_set_private_key(eckey, priv)))
goto err;
pkey = EVP_PKEY_new();
if (!TEST_ptr(pkey))
goto err;
if (!TEST_true(EVP_PKEY_assign_EC_KEY(pkey, eckey)))
goto err;
eckey = NULL;
Fix testcases to run on duplicated keys The existing loop pattern did not really run the expected tests on the duplicated keys. Fixes #23129 Reviewed-by: Neil Horman <nhorman@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/23292)
2024-01-13 01:47:56 +08:00
for (;;) {
Implement EVP_PKEY_dup() function Fixes #14501 Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14624)
2021-03-20 01:45:43 +08:00
ret = 0;
ctx = EVP_MD_CTX_new();
if (!TEST_ptr(ctx))
goto err;
/*
* The EVP_DigestSignInit function should create the key on the
* provider side which is sufficient for this test.
*/
if (!TEST_true(EVP_DigestSignInit_ex(ctx, NULL, NULL, testctx,
testpropq, pkey, NULL)))
goto err;
EVP_MD_CTX_free(ctx);
ctx = NULL;
Fix testcases to run on duplicated keys The existing loop pattern did not really run the expected tests on the duplicated keys. Fixes #23129 Reviewed-by: Neil Horman <nhorman@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/23292)
2024-01-13 01:47:56 +08:00
if (dup_pk != NULL)
break;
Implement EVP_PKEY_dup() function Fixes #14501 Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14624)
2021-03-20 01:45:43 +08:00
if (!TEST_ptr(dup_pk = EVP_PKEY_dup(pkey)))
goto err;
/* EVP_PKEY_eq() returns -2 with missing public keys */
ret = TEST_int_eq(EVP_PKEY_eq(pkey, dup_pk), -2);
EVP_PKEY_free(pkey);
pkey = dup_pk;
if (!ret)
goto err;
}
Fix testcases to run on duplicated keys The existing loop pattern did not really run the expected tests on the duplicated keys. Fixes #23129 Reviewed-by: Neil Horman <nhorman@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/23292)
2024-01-13 01:47:56 +08:00
ret = 1;
Test that EC keys without a public key in them work as expected We create EC keys via both the "fromdata" and legacy key routes to make sure that they can be used without a public key. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13922)
2021-01-21 23:14:15 +08:00
err:
EVP_MD_CTX_free(ctx);
EVP_PKEY_free(pkey);
EC_KEY_free(eckey);
BN_free(priv);
return ret;
}
test: evp_extra: EC, read affine coordinates Add a test to read the EC X,Y coordinates. Support legacy keys. Signed-off-by: Jorge Ramirez-Ortiz <jorge@foundries.io> Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/20535)
2023-03-20 19:51:55 +08:00
static int test_evp_get_ec_pub_legacy(void)
{
OSSL_LIB_CTX *libctx = NULL;
unsigned char *pad = NULL;
EVP_PKEY *pkey = NULL;
EC_KEY *eckey = NULL;
BIGNUM *priv = NULL;
BIGNUM *x = NULL;
BIGNUM *y = NULL;
int ret = 0;
if (!TEST_ptr(libctx = OSSL_LIB_CTX_new()))
goto err;
/* Create the legacy key */
if (!TEST_ptr(eckey = EC_KEY_new_by_curve_name_ex(libctx, NULL,
NID_X9_62_prime256v1)))
goto err;
if (!TEST_ptr(priv = BN_bin2bn(ec_priv, sizeof(ec_priv), NULL)))
goto err;
if (!TEST_true(EC_KEY_set_private_key(eckey, priv)))
goto err;
if (!TEST_ptr(x = BN_bin2bn(&ec_pub[1], 32, NULL)))
goto err;
if (!TEST_ptr(y = BN_bin2bn(&ec_pub[33], 32, NULL)))
goto err;
if (!TEST_true(EC_KEY_set_public_key_affine_coordinates(eckey, x, y)))
goto err;
if (!TEST_ptr(pkey = EVP_PKEY_new()))
goto err;
/* Transfer the legacy key */
if (!TEST_true(EVP_PKEY_assign_EC_KEY(pkey, eckey)))
goto err;
eckey = NULL;
if (!TEST_true(EVP_PKEY_get_bn_param(pkey, OSSL_PKEY_PARAM_EC_PUB_X, &x))
|| !TEST_true(EVP_PKEY_get_bn_param(pkey, OSSL_PKEY_PARAM_EC_PUB_Y, &y)))
goto err;
if (!TEST_ptr(pad = OPENSSL_zalloc(sizeof(ec_pub))))
goto err;
pad[0] = ec_pub[0];
BN_bn2bin(x, &pad[1]);
BN_bn2bin(y, &pad[33]);
if (!TEST_true(memcmp(ec_pub, pad, sizeof(ec_pub)) == 0))
goto err;
ret = 1;
err:
OSSL_LIB_CTX_free(libctx);
EVP_PKEY_free(pkey);
EC_KEY_free(eckey);
OPENSSL_free(pad);
BN_free(priv);
BN_free(x);
BN_free(y);
return ret;
}
Test that EC keys without a public key in them work as expected We create EC keys via both the "fromdata" and legacy key routes to make sure that they can be used without a public key. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13922)
2021-01-21 23:14:15 +08:00
# endif /* OPENSSL_NO_DEPRECATED_3_0 */
#endif /* OPENSSL_NO_EC */
Test short buffers Test that calling EVP_DigestSign(), EVP_DigestSignFinal(), EVP_PKEY_sign(), EVP_PKEY_get_raw_private_key(), or EVP_PKEY_get_raw_public_key() with a short output buffer results in a failure. Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16789)
2021-10-07 21:15:47 +08:00
static int test_EVP_PKEY_sign(int tst)
{
int ret = 0;
EVP_PKEY *pkey = NULL;
unsigned char *sig = NULL;
size_t sig_len = 0, shortsig_len = 1;
EVP_PKEY_CTX *ctx = NULL;
unsigned char tbs[] = {
0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08, 0x09, 0x0a, 0x0b,
0x0c, 0x0d, 0x0e, 0x0f, 0x10, 0x11, 0x12, 0x13
};
fix some code with obvious wrong coding style Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16918)
2021-10-26 15:16:18 +08:00
if (tst == 0) {
Test short buffers Test that calling EVP_DigestSign(), EVP_DigestSignFinal(), EVP_PKEY_sign(), EVP_PKEY_get_raw_private_key(), or EVP_PKEY_get_raw_public_key() with a short output buffer results in a failure. Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16789)
2021-10-07 21:15:47 +08:00
if (!TEST_ptr(pkey = load_example_rsa_key()))
test: evp_extra: fix indentation error Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/22185)
2023-09-24 07:24:59 +08:00
goto out;
Test short buffers Test that calling EVP_DigestSign(), EVP_DigestSignFinal(), EVP_PKEY_sign(), EVP_PKEY_get_raw_private_key(), or EVP_PKEY_get_raw_public_key() with a short output buffer results in a failure. Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16789)
2021-10-07 21:15:47 +08:00
} else if (tst == 1) {
#ifndef OPENSSL_NO_DSA
if (!TEST_ptr(pkey = load_example_dsa_key()))
test: evp_extra: fix indentation error Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/22185)
2023-09-24 07:24:59 +08:00
goto out;
Test short buffers Test that calling EVP_DigestSign(), EVP_DigestSignFinal(), EVP_PKEY_sign(), EVP_PKEY_get_raw_private_key(), or EVP_PKEY_get_raw_public_key() with a short output buffer results in a failure. Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16789)
2021-10-07 21:15:47 +08:00
#else
ret = 1;
goto out;
#endif
} else {
#ifndef OPENSSL_NO_EC
if (!TEST_ptr(pkey = load_example_ec_key()))
test: evp_extra: fix indentation error Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/22185)
2023-09-24 07:24:59 +08:00
goto out;
Test short buffers Test that calling EVP_DigestSign(), EVP_DigestSignFinal(), EVP_PKEY_sign(), EVP_PKEY_get_raw_private_key(), or EVP_PKEY_get_raw_public_key() with a short output buffer results in a failure. Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16789)
2021-10-07 21:15:47 +08:00
#else
ret = 1;
goto out;
#endif
}
ctx = EVP_PKEY_CTX_new_from_pkey(testctx, pkey, NULL);
if (!TEST_ptr(ctx)
|| !TEST_int_gt(EVP_PKEY_sign_init(ctx), 0)
|| !TEST_int_gt(EVP_PKEY_sign(ctx, NULL, &sig_len, tbs,
sizeof(tbs)), 0))
goto out;
sig = OPENSSL_malloc(sig_len);
if (!TEST_ptr(sig)
/* Test sending a signature buffer that is too short is rejected */
|| !TEST_int_le(EVP_PKEY_sign(ctx, sig, &shortsig_len, tbs,
sizeof(tbs)), 0)
|| !TEST_int_gt(EVP_PKEY_sign(ctx, sig, &sig_len, tbs, sizeof(tbs)),
0)
/* Test the signature round-trips */
|| !TEST_int_gt(EVP_PKEY_verify_init(ctx), 0)
|| !TEST_int_gt(EVP_PKEY_verify(ctx, sig, sig_len, tbs, sizeof(tbs)),
0))
goto out;
ret = 1;
out:
EVP_PKEY_CTX_free(ctx);
OPENSSL_free(sig);
EVP_PKEY_free(pkey);
return ret;
}
test: evp_extra: test signing with legacy app method based keys This commit adds `test_EVP_PKEY_sign_with_app_method`, a regression test for the bug fix in commit 1acc3e8cc3c6 (pull request #22163). It is analogous to `test_EVP_PKEY_sign`, only with a fake app method based key. (The EC key test case was omitted, because there is no `EC_KEY_METHOD_dup` method.) Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/22185)
2023-09-24 07:13:20 +08:00
#ifndef OPENSSL_NO_DEPRECATED_3_0
static int test_EVP_PKEY_sign_with_app_method(int tst)
{
int ret = 0;
EVP_PKEY *pkey = NULL;
RSA *rsa = NULL;
RSA_METHOD *rsa_meth = NULL;
#ifndef OPENSSL_NO_DSA
DSA *dsa = NULL;
DSA_METHOD *dsa_meth = NULL;
#endif
unsigned char *sig = NULL;
size_t sig_len = 0, shortsig_len = 1;
EVP_PKEY_CTX *ctx = NULL;
unsigned char tbs[] = {
0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08, 0x09, 0x0a, 0x0b,
0x0c, 0x0d, 0x0e, 0x0f, 0x10, 0x11, 0x12, 0x13
};
if (tst == 0) {
if (!TEST_ptr(pkey = load_example_rsa_key()))
goto out;
if (!TEST_ptr(rsa_meth = RSA_meth_dup(RSA_get_default_method())))
goto out;
if (!TEST_ptr(rsa = EVP_PKEY_get1_RSA(pkey))
|| !TEST_int_gt(RSA_set_method(rsa, rsa_meth), 0)
|| !TEST_int_gt(EVP_PKEY_assign_RSA(pkey, rsa), 0))
goto out;
rsa = NULL; /* now owned by the pkey */
} else {
#ifndef OPENSSL_NO_DSA
if (!TEST_ptr(pkey = load_example_dsa_key()))
goto out;
if (!TEST_ptr(dsa_meth = DSA_meth_dup(DSA_get_default_method())))
goto out;
if (!TEST_ptr(dsa = EVP_PKEY_get1_DSA(pkey))
|| !TEST_int_gt(DSA_set_method(dsa, dsa_meth), 0)
|| !TEST_int_gt(EVP_PKEY_assign_DSA(pkey, dsa), 0))
goto out;
dsa = NULL; /* now owned by the pkey */
#else
ret = 1;
goto out;
#endif
}
ctx = EVP_PKEY_CTX_new_from_pkey(testctx, pkey, NULL);
if (!TEST_ptr(ctx)
|| !TEST_int_gt(EVP_PKEY_sign_init(ctx), 0)
|| !TEST_int_gt(EVP_PKEY_sign(ctx, NULL, &sig_len, tbs,
sizeof(tbs)), 0))
goto out;
sig = OPENSSL_malloc(sig_len);
if (!TEST_ptr(sig)
/* Test sending a signature buffer that is too short is rejected */
|| !TEST_int_le(EVP_PKEY_sign(ctx, sig, &shortsig_len, tbs,
sizeof(tbs)), 0)
|| !TEST_int_gt(EVP_PKEY_sign(ctx, sig, &sig_len, tbs, sizeof(tbs)),
0)
/* Test the signature round-trips */
|| !TEST_int_gt(EVP_PKEY_verify_init(ctx), 0)
|| !TEST_int_gt(EVP_PKEY_verify(ctx, sig, sig_len, tbs, sizeof(tbs)),
0))
goto out;
ret = 1;
out:
EVP_PKEY_CTX_free(ctx);
OPENSSL_free(sig);
EVP_PKEY_free(pkey);
RSA_free(rsa);
RSA_meth_free(rsa_meth);
#ifndef OPENSSL_NO_DSA
DSA_free(dsa);
DSA_meth_free(dsa_meth);
#endif
return ret;
}
#endif /* !OPENSSL_NO_DEPRECATED_3_0 */
seal: make EVP_SealInit() library context aware Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/15300)
2021-05-17 10:18:53 +08:00
/*
* n = 0 => test using legacy cipher
* n = 1 => test using fetched cipher
*/
static int test_EVP_Enveloped(int n)
Key zeroization fix for EVP_SealInit + added simple test Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com> Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/7105)
2018-09-06 06:34:45 +08:00
{
int ret = 0;
EVP_CIPHER_CTX *ctx = NULL;
EVP_PKEY *keypair = NULL;
unsigned char *kek = NULL;
unsigned char iv[EVP_MAX_IV_LENGTH];
static const unsigned char msg[] = { 1, 2, 3, 4, 5, 6, 7, 8 };
int len, kek_len, ciphertext_len, plaintext_len;
unsigned char ciphertext[32], plaintext[16];
seal: make EVP_SealInit() library context aware Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/15300)
2021-05-17 10:18:53 +08:00
EVP_CIPHER *type = NULL;
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
if (nullprov != NULL)
return TEST_skip("Test does not support a non-default library context");
seal: make EVP_SealInit() library context aware Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/15300)
2021-05-17 10:18:53 +08:00
if (n == 0)
type = (EVP_CIPHER *)EVP_aes_256_cbc();
else if (!TEST_ptr(type = EVP_CIPHER_fetch(testctx, "AES-256-CBC",
testpropq)))
goto err;
Key zeroization fix for EVP_SealInit + added simple test Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com> Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/7105)
2018-09-06 06:34:45 +08:00
if (!TEST_ptr(keypair = load_example_rsa_key())
Rename all getters to use get/get0 in name For functions that exist in 1.1.1 provide a simple aliases via #define. Fixes #15236 Functions with OSSL_DECODER_, OSSL_ENCODER_, OSSL_STORE_LOADER_, EVP_KEYEXCH_, EVP_KEM_, EVP_ASYM_CIPHER_, EVP_SIGNATURE_, EVP_KEYMGMT_, EVP_RAND_, EVP_MAC_, EVP_KDF_, EVP_PKEY_, EVP_MD_, and EVP_CIPHER_ prefixes are renamed. Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/15405)
2021-05-21 22:58:08 +08:00
|| !TEST_ptr(kek = OPENSSL_zalloc(EVP_PKEY_get_size(keypair)))
Key zeroization fix for EVP_SealInit + added simple test Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com> Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/7105)
2018-09-06 06:34:45 +08:00
|| !TEST_ptr(ctx = EVP_CIPHER_CTX_new())
|| !TEST_true(EVP_SealInit(ctx, type, &kek, &kek_len, iv,
&keypair, 1))
|| !TEST_true(EVP_SealUpdate(ctx, ciphertext, &ciphertext_len,
msg, sizeof(msg)))
|| !TEST_true(EVP_SealFinal(ctx, ciphertext + ciphertext_len,
&len)))
goto err;
ciphertext_len += len;
if (!TEST_true(EVP_OpenInit(ctx, type, kek, kek_len, iv, keypair))
|| !TEST_true(EVP_OpenUpdate(ctx, plaintext, &plaintext_len,
ciphertext, ciphertext_len))
|| !TEST_true(EVP_OpenFinal(ctx, plaintext + plaintext_len, &len)))
goto err;
plaintext_len += len;
if (!TEST_mem_eq(msg, sizeof(msg), plaintext, plaintext_len))
goto err;
ret = 1;
err:
seal: make EVP_SealInit() library context aware Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/15300)
2021-05-17 10:18:53 +08:00
if (n != 0)
EVP_CIPHER_free(type);
Key zeroization fix for EVP_SealInit + added simple test Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com> Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/7105)
2018-09-06 06:34:45 +08:00
OPENSSL_free(kek);
EVP_PKEY_free(keypair);
EVP_CIPHER_CTX_free(ctx);
return ret;
}
Add a test for EVP_Digest[Sign|Verify]* with a BIO MD If an EVP_MD_CTX is initialised with EVP_DigestSignInit_ex() or EVP_DigestVerifyInit_ex() it should work with an MD BIO to provide the "Updates". Test that this is the case. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10116)
2019-10-08 01:21:39 +08:00
/*
Add a test for HMAC via EVP_DigestSign* We test with both an implicitly fetched digest and an explicitly fetched digest. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/10796)
2020-01-09 23:21:14 +08:00
* Test 0: Standard calls to EVP_DigestSignInit/Update/Final (Implicit fetch digest, RSA)
* Test 1: Standard calls to EVP_DigestSignInit/Update/Final (Implicit fetch digest, DSA)
* Test 2: Standard calls to EVP_DigestSignInit/Update/Final (Implicit fetch digest, HMAC)
* Test 3: Standard calls to EVP_DigestSignInit/Update/Final (Explicit fetch digest, RSA)
* Test 4: Standard calls to EVP_DigestSignInit/Update/Final (Explicit fetch digest, DSA)
* Test 5: Standard calls to EVP_DigestSignInit/Update/Final (Explicit fetch diegst, HMAC)
* Test 6: Use an MD BIO to do the Update calls instead (RSA)
* Test 7: Use an MD BIO to do the Update calls instead (DSA)
* Test 8: Use an MD BIO to do the Update calls instead (HMAC)
Test short buffers Test that calling EVP_DigestSign(), EVP_DigestSignFinal(), EVP_PKEY_sign(), EVP_PKEY_get_raw_private_key(), or EVP_PKEY_get_raw_public_key() with a short output buffer results in a failure. Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16789)
2021-10-07 21:15:47 +08:00
* Test 9: Use EVP_DigestSign (Implicit fetch digest, RSA, short sig)
* Test 10: Use EVP_DigestSign (Implicit fetch digest, DSA, short sig)
* Test 11: Use EVP_DigestSign (Implicit fetch digest, HMAC, short sig)
* Test 12: Use EVP_DigestSign (Implicit fetch digest, RSA)
* Test 13: Use EVP_DigestSign (Implicit fetch digest, DSA)
* Test 14: Use EVP_DigestSign (Implicit fetch digest, HMAC)
test: Add testing of reinitialization via EVP_DigestSignInit() Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16964)
2021-11-04 22:35:40 +08:00
* Test 15-29: Same as above with reinitialization
Add a test for EVP_Digest[Sign|Verify]* with a BIO MD If an EVP_MD_CTX is initialised with EVP_DigestSignInit_ex() or EVP_DigestVerifyInit_ex() it should work with an MD BIO to provide the "Updates". Test that this is the case. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10116)
2019-10-08 01:21:39 +08:00
*/
static int test_EVP_DigestSignInit(int tst)
Import evp_test.c from BoringSSL. Unfortunately we already have a file called evp_test.c, so I have called this one evp_extra_test.c Reviewed-by: Emilia Käsper <emilia@openssl.org>
2015-02-09 17:45:35 +08:00
{
int ret = 0;
EVP_PKEY *pkey = NULL;
EVP_DigestSign/VerifyFinal: Duplicate the pctx to allow multiple calls The legacy implementation duplicates the pctx before creating/verifying the signature unless EVP_MD_CTX_FLAG_FINALISE is set. We have to do the same with provided implementations. Fixes #16321 Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16422)
2021-08-25 19:50:40 +08:00
unsigned char *sig = NULL, *sig2 = NULL;
Test short buffers Test that calling EVP_DigestSign(), EVP_DigestSignFinal(), EVP_PKEY_sign(), EVP_PKEY_get_raw_private_key(), or EVP_PKEY_get_raw_public_key() with a short output buffer results in a failure. Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16789)
2021-10-07 21:15:47 +08:00
size_t sig_len = 0, sig2_len = 0, shortsig_len = 1;
Add a test for EVP_Digest[Sign|Verify]* with a BIO MD If an EVP_MD_CTX is initialised with EVP_DigestSignInit_ex() or EVP_DigestVerifyInit_ex() it should work with an MD BIO to provide the "Updates". Test that this is the case. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10116)
2019-10-08 01:21:39 +08:00
EVP_MD_CTX *md_ctx = NULL, *md_ctx_verify = NULL;
EVP_MD_CTX *a_md_ctx = NULL, *a_md_ctx_verify = NULL;
BIO *mdbio = NULL, *membio = NULL;
size_t written;
Add a test for HMAC via EVP_DigestSign* We test with both an implicitly fetched digest and an explicitly fetched digest. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/10796)
2020-01-09 23:21:14 +08:00
const EVP_MD *md;
EVP_MD *mdexp = NULL;
test: Add testing of reinitialization via EVP_DigestSignInit() Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16964)
2021-11-04 22:35:40 +08:00
int reinit = 0;
Add a test for EVP_Digest[Sign|Verify]* with a BIO MD If an EVP_MD_CTX is initialised with EVP_DigestSignInit_ex() or EVP_DigestVerifyInit_ex() it should work with an MD BIO to provide the "Updates". Test that this is the case. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10116)
2019-10-08 01:21:39 +08:00
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
if (nullprov != NULL)
return TEST_skip("Test does not support a non-default library context");
test: Add testing of reinitialization via EVP_DigestSignInit() Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16964)
2021-11-04 22:35:40 +08:00
if (tst >= 15) {
reinit = 1;
tst -= 15;
}
Test short buffers Test that calling EVP_DigestSign(), EVP_DigestSignFinal(), EVP_PKEY_sign(), EVP_PKEY_get_raw_private_key(), or EVP_PKEY_get_raw_public_key() with a short output buffer results in a failure. Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16789)
2021-10-07 21:15:47 +08:00
if (tst >= 6 && tst <= 8) {
Add a test for EVP_Digest[Sign|Verify]* with a BIO MD If an EVP_MD_CTX is initialised with EVP_DigestSignInit_ex() or EVP_DigestVerifyInit_ex() it should work with an MD BIO to provide the "Updates". Test that this is the case. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10116)
2019-10-08 01:21:39 +08:00
membio = BIO_new(BIO_s_mem());
mdbio = BIO_new(BIO_f_md());
if (!TEST_ptr(membio) || !TEST_ptr(mdbio))
goto out;
BIO_push(mdbio, membio);
if (!TEST_int_gt(BIO_get_md_ctx(mdbio, &md_ctx), 0))
goto out;
} else {
if (!TEST_ptr(a_md_ctx = md_ctx = EVP_MD_CTX_new())
|| !TEST_ptr(a_md_ctx_verify = md_ctx_verify = EVP_MD_CTX_new()))
goto out;
}
Import evp_test.c from BoringSSL. Unfortunately we already have a file called evp_test.c, so I have called this one evp_extra_test.c Reviewed-by: Emilia Käsper <emilia@openssl.org>
2015-02-09 17:45:35 +08:00
Test short buffers Test that calling EVP_DigestSign(), EVP_DigestSignFinal(), EVP_PKEY_sign(), EVP_PKEY_get_raw_private_key(), or EVP_PKEY_get_raw_public_key() with a short output buffer results in a failure. Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16789)
2021-10-07 21:15:47 +08:00
if (tst % 3 == 0) {
Add a test for EVP_Digest[Sign|Verify]* with a BIO MD If an EVP_MD_CTX is initialised with EVP_DigestSignInit_ex() or EVP_DigestVerifyInit_ex() it should work with an MD BIO to provide the "Updates". Test that this is the case. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10116)
2019-10-08 01:21:39 +08:00
if (!TEST_ptr(pkey = load_example_rsa_key()))
goto out;
Test short buffers Test that calling EVP_DigestSign(), EVP_DigestSignFinal(), EVP_PKEY_sign(), EVP_PKEY_get_raw_private_key(), or EVP_PKEY_get_raw_public_key() with a short output buffer results in a failure. Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16789)
2021-10-07 21:15:47 +08:00
} else if (tst % 3 == 1) {
Fix no-dsa Make sure we don't try and load a DSA key in the tests if DSA has been disabled. Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10375)
2019-11-07 17:24:18 +08:00
#ifndef OPENSSL_NO_DSA
Add a test for EVP_Digest[Sign|Verify]* with a BIO MD If an EVP_MD_CTX is initialised with EVP_DigestSignInit_ex() or EVP_DigestVerifyInit_ex() it should work with an MD BIO to provide the "Updates". Test that this is the case. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10116)
2019-10-08 01:21:39 +08:00
if (!TEST_ptr(pkey = load_example_dsa_key()))
goto out;
Fix no-dsa Make sure we don't try and load a DSA key in the tests if DSA has been disabled. Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10375)
2019-11-07 17:24:18 +08:00
#else
ret = 1;
goto out;
#endif
Add a test for HMAC via EVP_DigestSign* We test with both an implicitly fetched digest and an explicitly fetched digest. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/10796)
2020-01-09 23:21:14 +08:00
} else {
if (!TEST_ptr(pkey = load_example_hmac_key()))
goto out;
Add a test for EVP_Digest[Sign|Verify]* with a BIO MD If an EVP_MD_CTX is initialised with EVP_DigestSignInit_ex() or EVP_DigestVerifyInit_ex() it should work with an MD BIO to provide the "Updates". Test that this is the case. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10116)
2019-10-08 01:21:39 +08:00
}
Import evp_test.c from BoringSSL. Unfortunately we already have a file called evp_test.c, so I have called this one evp_extra_test.c Reviewed-by: Emilia Käsper <emilia@openssl.org>
2015-02-09 17:45:35 +08:00
Add a test for HMAC via EVP_DigestSign* We test with both an implicitly fetched digest and an explicitly fetched digest. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/10796)
2020-01-09 23:21:14 +08:00
if (tst >= 3 && tst <= 5)
md = mdexp = EVP_MD_fetch(NULL, "SHA256", NULL);
else
md = EVP_sha256();
if (!TEST_true(EVP_DigestSignInit(md_ctx, NULL, md, NULL, pkey)))
Import evp_test.c from BoringSSL. Unfortunately we already have a file called evp_test.c, so I have called this one evp_extra_test.c Reviewed-by: Emilia Käsper <emilia@openssl.org>
2015-02-09 17:45:35 +08:00
goto out;
Convert more tests ct_test,evp_extra_test,wpackettest,packettest Add strncmp TEST wrappers And make some style/consistency fixes to ct_test Silence travis; gcc bug? Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/3234)
2017-04-19 04:33:15 +08:00
test: Add testing of reinitialization via EVP_DigestSignInit() Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16964)
2021-11-04 22:35:40 +08:00
if (reinit && !TEST_true(EVP_DigestSignInit(md_ctx, NULL, NULL, NULL, NULL)))
goto out;
Test short buffers Test that calling EVP_DigestSign(), EVP_DigestSignFinal(), EVP_PKEY_sign(), EVP_PKEY_get_raw_private_key(), or EVP_PKEY_get_raw_public_key() with a short output buffer results in a failure. Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16789)
2021-10-07 21:15:47 +08:00
if (tst >= 6 && tst <= 8) {
Add a test for EVP_Digest[Sign|Verify]* with a BIO MD If an EVP_MD_CTX is initialised with EVP_DigestSignInit_ex() or EVP_DigestVerifyInit_ex() it should work with an MD BIO to provide the "Updates". Test that this is the case. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10116)
2019-10-08 01:21:39 +08:00
if (!BIO_write_ex(mdbio, kMsg, sizeof(kMsg), &written))
goto out;
Test short buffers Test that calling EVP_DigestSign(), EVP_DigestSignFinal(), EVP_PKEY_sign(), EVP_PKEY_get_raw_private_key(), or EVP_PKEY_get_raw_public_key() with a short output buffer results in a failure. Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16789)
2021-10-07 21:15:47 +08:00
} else if (tst < 6) {
Add a test for EVP_Digest[Sign|Verify]* with a BIO MD If an EVP_MD_CTX is initialised with EVP_DigestSignInit_ex() or EVP_DigestVerifyInit_ex() it should work with an MD BIO to provide the "Updates". Test that this is the case. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10116)
2019-10-08 01:21:39 +08:00
if (!TEST_true(EVP_DigestSignUpdate(md_ctx, kMsg, sizeof(kMsg))))
goto out;
}
Test short buffers Test that calling EVP_DigestSign(), EVP_DigestSignFinal(), EVP_PKEY_sign(), EVP_PKEY_get_raw_private_key(), or EVP_PKEY_get_raw_public_key() with a short output buffer results in a failure. Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16789)
2021-10-07 21:15:47 +08:00
if (tst >= 9) {
/* Determine the size of the signature. */
if (!TEST_true(EVP_DigestSign(md_ctx, NULL, &sig_len, kMsg,
sizeof(kMsg)))
|| !TEST_ptr(sig = OPENSSL_malloc(sig_len)))
goto out;
if (tst <= 11) {
/* Test that supply a short sig buffer fails */
if (!TEST_false(EVP_DigestSign(md_ctx, sig, &shortsig_len, kMsg,
sizeof(kMsg))))
goto out;
/*
* We end here because once EVP_DigestSign() has failed you should
* not call it again without re-initing the ctx
*/
ret = 1;
goto out;
}
if (!TEST_true(EVP_DigestSign(md_ctx, sig, &sig_len, kMsg,
sizeof(kMsg))))
goto out;
} else {
/* Determine the size of the signature. */
if (!TEST_true(EVP_DigestSignFinal(md_ctx, NULL, &sig_len))
|| !TEST_ptr(sig = OPENSSL_malloc(sig_len))
/*
* Trying to create a signature with a deliberately short
* buffer should fail.
*/
|| !TEST_false(EVP_DigestSignFinal(md_ctx, sig, &shortsig_len))
|| !TEST_true(EVP_DigestSignFinal(md_ctx, sig, &sig_len)))
goto out;
}
Import evp_test.c from BoringSSL. Unfortunately we already have a file called evp_test.c, so I have called this one evp_extra_test.c Reviewed-by: Emilia Käsper <emilia@openssl.org>
2015-02-09 17:45:35 +08:00
Add a test for HMAC via EVP_DigestSign* We test with both an implicitly fetched digest and an explicitly fetched digest. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/10796)
2020-01-09 23:21:14 +08:00
/*
* Ensure that the signature round-trips (Verification isn't supported for
* HMAC via EVP_DigestVerify*)
*/
Test short buffers Test that calling EVP_DigestSign(), EVP_DigestSignFinal(), EVP_PKEY_sign(), EVP_PKEY_get_raw_private_key(), or EVP_PKEY_get_raw_public_key() with a short output buffer results in a failure. Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16789)
2021-10-07 21:15:47 +08:00
if (tst % 3 != 2) {
if (tst >= 6 && tst <= 8) {
EVP_DigestSign/VerifyFinal: Duplicate the pctx to allow multiple calls The legacy implementation duplicates the pctx before creating/verifying the signature unless EVP_MD_CTX_FLAG_FINALISE is set. We have to do the same with provided implementations. Fixes #16321 Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16422)
2021-08-25 19:50:40 +08:00
if (!TEST_int_gt(BIO_reset(mdbio), 0)
|| !TEST_int_gt(BIO_get_md_ctx(mdbio, &md_ctx_verify), 0))
goto out;
}
Add a test for HMAC via EVP_DigestSign* We test with both an implicitly fetched digest and an explicitly fetched digest. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/10796)
2020-01-09 23:21:14 +08:00
if (!TEST_true(EVP_DigestVerifyInit(md_ctx_verify, NULL, md,
NULL, pkey)))
Add a test for EVP_Digest[Sign|Verify]* with a BIO MD If an EVP_MD_CTX is initialised with EVP_DigestSignInit_ex() or EVP_DigestVerifyInit_ex() it should work with an MD BIO to provide the "Updates". Test that this is the case. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10116)
2019-10-08 01:21:39 +08:00
goto out;
Add a test for HMAC via EVP_DigestSign* We test with both an implicitly fetched digest and an explicitly fetched digest. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/10796)
2020-01-09 23:21:14 +08:00
Test short buffers Test that calling EVP_DigestSign(), EVP_DigestSignFinal(), EVP_PKEY_sign(), EVP_PKEY_get_raw_private_key(), or EVP_PKEY_get_raw_public_key() with a short output buffer results in a failure. Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16789)
2021-10-07 21:15:47 +08:00
if (tst >= 6 && tst <= 8) {
Add a test for HMAC via EVP_DigestSign* We test with both an implicitly fetched digest and an explicitly fetched digest. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/10796)
2020-01-09 23:21:14 +08:00
if (!TEST_true(BIO_write_ex(mdbio, kMsg, sizeof(kMsg), &written)))
goto out;
} else {
if (!TEST_true(EVP_DigestVerifyUpdate(md_ctx_verify, kMsg,
sizeof(kMsg))))
goto out;
}
EVP_DigestVerifyFinal: fix test function and invocation Signed-off-by: Peiwei Hu <jlu.hpw@foxmail.com> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/17029)
2021-11-14 22:42:35 +08:00
if (!TEST_int_gt(EVP_DigestVerifyFinal(md_ctx_verify, sig, sig_len), 0))
Add a test for EVP_Digest[Sign|Verify]* with a BIO MD If an EVP_MD_CTX is initialised with EVP_DigestSignInit_ex() or EVP_DigestVerifyInit_ex() it should work with an MD BIO to provide the "Updates". Test that this is the case. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10116)
2019-10-08 01:21:39 +08:00
goto out;
EVP_DigestSign/VerifyFinal: Duplicate the pctx to allow multiple calls The legacy implementation duplicates the pctx before creating/verifying the signature unless EVP_MD_CTX_FLAG_FINALISE is set. We have to do the same with provided implementations. Fixes #16321 Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16422)
2021-08-25 19:50:40 +08:00
/* Multiple calls to EVP_DigestVerifyFinal should work */
EVP_DigestVerifyFinal: fix test function and invocation Signed-off-by: Peiwei Hu <jlu.hpw@foxmail.com> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/17029)
2021-11-14 22:42:35 +08:00
if (!TEST_int_gt(EVP_DigestVerifyFinal(md_ctx_verify, sig, sig_len), 0))
EVP_DigestSign/VerifyFinal: Duplicate the pctx to allow multiple calls The legacy implementation duplicates the pctx before creating/verifying the signature unless EVP_MD_CTX_FLAG_FINALISE is set. We have to do the same with provided implementations. Fixes #16321 Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16422)
2021-08-25 19:50:40 +08:00
goto out;
} else {
/*
* For HMAC a doubled call to DigestSignFinal should produce the same
* value as finalization should not happen.
*/
if (!TEST_true(EVP_DigestSignFinal(md_ctx, NULL, &sig2_len))
|| !TEST_ptr(sig2 = OPENSSL_malloc(sig2_len))
|| !TEST_true(EVP_DigestSignFinal(md_ctx, sig2, &sig2_len)))
goto out;
if (!TEST_mem_eq(sig, sig_len, sig2, sig2_len))
goto out;
Add a test for EVP_Digest[Sign|Verify]* with a BIO MD If an EVP_MD_CTX is initialised with EVP_DigestSignInit_ex() or EVP_DigestVerifyInit_ex() it should work with an MD BIO to provide the "Updates". Test that this is the case. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10116)
2019-10-08 01:21:39 +08:00
}
Import evp_test.c from BoringSSL. Unfortunately we already have a file called evp_test.c, so I have called this one evp_extra_test.c Reviewed-by: Emilia Käsper <emilia@openssl.org>
2015-02-09 17:45:35 +08:00
ret = 1;
out:
Add a test for EVP_Digest[Sign|Verify]* with a BIO MD If an EVP_MD_CTX is initialised with EVP_DigestSignInit_ex() or EVP_DigestVerifyInit_ex() it should work with an MD BIO to provide the "Updates". Test that this is the case. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10116)
2019-10-08 01:21:39 +08:00
BIO_free(membio);
BIO_free(mdbio);
EVP_MD_CTX_free(a_md_ctx);
EVP_MD_CTX_free(a_md_ctx_verify);
free NULL cleanup EVP_.*free; this gets: EVP_CIPHER_CTX_free EVP_PKEY_CTX_free EVP_PKEY_asn1_free EVP_PKEY_asn1_set_free EVP_PKEY_free EVP_PKEY_free_it EVP_PKEY_meth_free; and also EVP_CIPHER_CTX_cleanup Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2015-03-28 22:54:15 +08:00
EVP_PKEY_free(pkey);
free null cleanup finale Don't check for NULL before calling OPENSSL_free Reviewed-by: Richard Levitte <levitte@openssl.org>
2015-05-01 22:02:07 +08:00
OPENSSL_free(sig);
EVP_DigestSign/VerifyFinal: Duplicate the pctx to allow multiple calls The legacy implementation duplicates the pctx before creating/verifying the signature unless EVP_MD_CTX_FLAG_FINALISE is set. We have to do the same with provided implementations. Fixes #16321 Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16422)
2021-08-25 19:50:40 +08:00
OPENSSL_free(sig2);
Add a test for HMAC via EVP_DigestSign* We test with both an implicitly fetched digest and an explicitly fetched digest. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/10796)
2020-01-09 23:21:14 +08:00
EVP_MD_free(mdexp);
Import evp_test.c from BoringSSL. Unfortunately we already have a file called evp_test.c, so I have called this one evp_extra_test.c Reviewed-by: Emilia Käsper <emilia@openssl.org>
2015-02-09 17:45:35 +08:00
return ret;
}
static int test_EVP_DigestVerifyInit(void)
{
int ret = 0;
EVP_PKEY *pkey = NULL;
Convert more tests ct_test,evp_extra_test,wpackettest,packettest Add strncmp TEST wrappers And make some style/consistency fixes to ct_test Silence travis; gcc bug? Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/3234)
2017-04-19 04:33:15 +08:00
EVP_MD_CTX *md_ctx = NULL;
Import evp_test.c from BoringSSL. Unfortunately we already have a file called evp_test.c, so I have called this one evp_extra_test.c Reviewed-by: Emilia Käsper <emilia@openssl.org>
2015-02-09 17:45:35 +08:00
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
if (nullprov != NULL)
return TEST_skip("Test does not support a non-default library context");
Convert more tests ct_test,evp_extra_test,wpackettest,packettest Add strncmp TEST wrappers And make some style/consistency fixes to ct_test Silence travis; gcc bug? Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/3234)
2017-04-19 04:33:15 +08:00
if (!TEST_ptr(md_ctx = EVP_MD_CTX_new())
|| !TEST_ptr(pkey = load_example_rsa_key()))
goto out;
Import evp_test.c from BoringSSL. Unfortunately we already have a file called evp_test.c, so I have called this one evp_extra_test.c Reviewed-by: Emilia Käsper <emilia@openssl.org>
2015-02-09 17:45:35 +08:00
Convert more tests ct_test,evp_extra_test,wpackettest,packettest Add strncmp TEST wrappers And make some style/consistency fixes to ct_test Silence travis; gcc bug? Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/3234)
2017-04-19 04:33:15 +08:00
if (!TEST_true(EVP_DigestVerifyInit(md_ctx, NULL, EVP_sha256(), NULL, pkey))
|| !TEST_true(EVP_DigestVerifyUpdate(md_ctx, kMsg, sizeof(kMsg)))
EVP_DigestVerifyFinal: fix test function and invocation Signed-off-by: Peiwei Hu <jlu.hpw@foxmail.com> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/17029)
2021-11-14 22:42:35 +08:00
|| !TEST_int_gt(EVP_DigestVerifyFinal(md_ctx, kSignature,
sizeof(kSignature)), 0))
Import evp_test.c from BoringSSL. Unfortunately we already have a file called evp_test.c, so I have called this one evp_extra_test.c Reviewed-by: Emilia Käsper <emilia@openssl.org>
2015-02-09 17:45:35 +08:00
goto out;
do_sigver_init: Allow reinitialization of an existing operation. Fixes #16936 Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16964)
2021-11-04 18:06:26 +08:00
/* test with reinitialization */
if (!TEST_true(EVP_DigestVerifyInit(md_ctx, NULL, NULL, NULL, NULL))
|| !TEST_true(EVP_DigestVerifyUpdate(md_ctx, kMsg, sizeof(kMsg)))
EVP_DigestVerifyFinal: fix test function and invocation Signed-off-by: Peiwei Hu <jlu.hpw@foxmail.com> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/17029)
2021-11-14 22:42:35 +08:00
|| !TEST_int_gt(EVP_DigestVerifyFinal(md_ctx, kSignature,
sizeof(kSignature)), 0))
do_sigver_init: Allow reinitialization of an existing operation. Fixes #16936 Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16964)
2021-11-04 18:06:26 +08:00
goto out;
Import evp_test.c from BoringSSL. Unfortunately we already have a file called evp_test.c, so I have called this one evp_extra_test.c Reviewed-by: Emilia Käsper <emilia@openssl.org>
2015-02-09 17:45:35 +08:00
ret = 1;
out:
Cleanup: fix all sources that used EVP_MD_CTX_(create|init|destroy) Reviewed-by: Rich Salz <rsalz@openssl.org>
2015-12-02 07:49:35 +08:00
EVP_MD_CTX_free(md_ctx);
free NULL cleanup EVP_.*free; this gets: EVP_CIPHER_CTX_free EVP_PKEY_CTX_free EVP_PKEY_asn1_free EVP_PKEY_asn1_set_free EVP_PKEY_free EVP_PKEY_free_it EVP_PKEY_meth_free; and also EVP_CIPHER_CTX_cleanup Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2015-03-28 22:54:15 +08:00
EVP_PKEY_free(pkey);
Import evp_test.c from BoringSSL. Unfortunately we already have a file called evp_test.c, so I have called this one evp_extra_test.c Reviewed-by: Emilia Käsper <emilia@openssl.org>
2015-02-09 17:45:35 +08:00
return ret;
}
evp_extra_test: Add SIPHASH MAC digestsign test with reinitialization Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16964)
2021-11-05 21:04:25 +08:00
#ifndef OPENSSL_NO_SIPHASH
/* test SIPHASH MAC via EVP_PKEY with non-default parameters and reinit */
static int test_siphash_digestsign(void)
{
unsigned char key[16];
unsigned char buf[8], digest[8];
unsigned char expected[8] = {
0x6d, 0x3e, 0x54, 0xc2, 0x2f, 0xf1, 0xfe, 0xe2
};
EVP_PKEY *pkey = NULL;
EVP_MD_CTX *mdctx = NULL;
EVP_PKEY_CTX *ctx = NULL;
int ret = 0;
size_t len = 8;
if (nullprov != NULL)
return TEST_skip("Test does not support a non-default library context");
memset(buf, 0, 8);
memset(key, 1, 16);
if (!TEST_ptr(pkey = EVP_PKEY_new_raw_private_key(EVP_PKEY_SIPHASH, NULL,
key, 16)))
goto out;
if (!TEST_ptr(mdctx = EVP_MD_CTX_create()))
goto out;
if (!TEST_true(EVP_DigestSignInit(mdctx, &ctx, NULL, NULL, pkey)))
goto out;
if (!TEST_int_eq(EVP_PKEY_CTX_ctrl(ctx, -1, EVP_PKEY_OP_SIGNCTX,
EVP_PKEY_CTRL_SET_DIGEST_SIZE,
8, NULL), 1))
goto out;
/* reinitialize */
if (!TEST_true(EVP_DigestSignInit(mdctx, NULL, NULL, NULL, NULL)))
goto out;
if (!TEST_true(EVP_DigestSignUpdate(mdctx, buf, 8)))
goto out;
if (!TEST_true(EVP_DigestSignFinal(mdctx, digest, &len)))
goto out;
if (!TEST_mem_eq(digest, len, expected, sizeof(expected)))
goto out;
ret = 1;
out:
EVP_PKEY_free(pkey);
EVP_MD_CTX_free(mdctx);
return ret;
}
#endif
Add test for no reset after DigestFinal_ex and DigestFinalXOF Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com> Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/13402)
2020-11-13 21:16:35 +08:00
/*
* Test corner cases of EVP_DigestInit/Update/Final API call behavior.
*/
static int test_EVP_Digest(void)
{
int ret = 0;
EVP_MD_CTX *md_ctx = NULL;
unsigned char md[EVP_MAX_MD_SIZE];
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
EVP_MD *sha256 = NULL;
EVP_MD *shake256 = NULL;
Add test for no reset after DigestFinal_ex and DigestFinalXOF Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com> Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/13402)
2020-11-13 21:16:35 +08:00
if (!TEST_ptr(md_ctx = EVP_MD_CTX_new()))
goto out;
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
if (!TEST_ptr(sha256 = EVP_MD_fetch(testctx, "sha256", testpropq))
|| !TEST_ptr(shake256 = EVP_MD_fetch(testctx, "shake256", testpropq)))
goto out;
if (!TEST_true(EVP_DigestInit_ex(md_ctx, sha256, NULL))
Add test for no reset after DigestFinal_ex and DigestFinalXOF Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com> Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/13402)
2020-11-13 21:16:35 +08:00
|| !TEST_true(EVP_DigestUpdate(md_ctx, kMsg, sizeof(kMsg)))
|| !TEST_true(EVP_DigestFinal(md_ctx, md, NULL))
/* EVP_DigestFinal resets the EVP_MD_CTX. */
Add "origin" field to EVP_CIPHER, EVP_MD Add a "where did this EVP_{CIPHER,MD} come from" flag: global, via fetch, or via EVP_{CIPHER,MD}_meth_new. Update EVP_{CIPHER,MD}_free to handle all three origins. The flag is deliberately right before some function pointers, so that compile-time failures (int/pointer) will occur, as opposed to taking a bit in the existing "flags" field. The "global variable" flag is non-zero, so the default case of using OPENSSL_zalloc (for provider ciphers), will do the right thing. Ref-counting is a no-op for Make up_ref no-op for global MD and CIPHER objects Deprecate EVP_MD_CTX_md(). Added EVP_MD_CTX_get0_md() (same semantics as the deprecated function) and EVP_MD_CTX_get1_md(). Likewise, deprecate EVP_CIPHER_CTX_cipher() in favor of EVP_CIPHER_CTX_get0_cipher(), and add EVP_CIPHER_CTX_get1_CIPHER(). Refactor EVP_MD_free() and EVP_MD_meth_free() to call new common evp_md_free_int() function. Refactor EVP_CIPHER_free() and EVP_CIPHER_meth_free() to call new common evp_cipher_free_int() function. Also change some flags tests to explicit test == or != zero. E.g., if (flags & x) --> if ((flags & x) != 0) if (!(flags & x)) --> if ((flags & x) == 0) Only done for those lines where "get0_cipher" calls were made. Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14193)
2021-02-17 06:51:56 +08:00
|| !TEST_ptr_eq(EVP_MD_CTX_get0_md(md_ctx), NULL))
Add test for no reset after DigestFinal_ex and DigestFinalXOF Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com> Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/13402)
2020-11-13 21:16:35 +08:00
goto out;
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
if (!TEST_true(EVP_DigestInit_ex(md_ctx, sha256, NULL))
Add test for no reset after DigestFinal_ex and DigestFinalXOF Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com> Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/13402)
2020-11-13 21:16:35 +08:00
|| !TEST_true(EVP_DigestUpdate(md_ctx, kMsg, sizeof(kMsg)))
|| !TEST_true(EVP_DigestFinal_ex(md_ctx, md, NULL))
/* EVP_DigestFinal_ex does not reset the EVP_MD_CTX. */
Add "origin" field to EVP_CIPHER, EVP_MD Add a "where did this EVP_{CIPHER,MD} come from" flag: global, via fetch, or via EVP_{CIPHER,MD}_meth_new. Update EVP_{CIPHER,MD}_free to handle all three origins. The flag is deliberately right before some function pointers, so that compile-time failures (int/pointer) will occur, as opposed to taking a bit in the existing "flags" field. The "global variable" flag is non-zero, so the default case of using OPENSSL_zalloc (for provider ciphers), will do the right thing. Ref-counting is a no-op for Make up_ref no-op for global MD and CIPHER objects Deprecate EVP_MD_CTX_md(). Added EVP_MD_CTX_get0_md() (same semantics as the deprecated function) and EVP_MD_CTX_get1_md(). Likewise, deprecate EVP_CIPHER_CTX_cipher() in favor of EVP_CIPHER_CTX_get0_cipher(), and add EVP_CIPHER_CTX_get1_CIPHER(). Refactor EVP_MD_free() and EVP_MD_meth_free() to call new common evp_md_free_int() function. Refactor EVP_CIPHER_free() and EVP_CIPHER_meth_free() to call new common evp_cipher_free_int() function. Also change some flags tests to explicit test == or != zero. E.g., if (flags & x) --> if ((flags & x) != 0) if (!(flags & x)) --> if ((flags & x) == 0) Only done for those lines where "get0_cipher" calls were made. Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14193)
2021-02-17 06:51:56 +08:00
|| !TEST_ptr(EVP_MD_CTX_get0_md(md_ctx))
Add test for no reset after DigestFinal_ex and DigestFinalXOF Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com> Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/13402)
2020-11-13 21:16:35 +08:00
/*
* EVP_DigestInit_ex with NULL type should work on
* pre-initialized context.
*/
|| !TEST_true(EVP_DigestInit_ex(md_ctx, NULL, NULL)))
goto out;
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
if (!TEST_true(EVP_DigestInit_ex(md_ctx, shake256, NULL))
Add test for no reset after DigestFinal_ex and DigestFinalXOF Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com> Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/13402)
2020-11-13 21:16:35 +08:00
|| !TEST_true(EVP_DigestUpdate(md_ctx, kMsg, sizeof(kMsg)))
|| !TEST_true(EVP_DigestFinalXOF(md_ctx, md, sizeof(md)))
/* EVP_DigestFinalXOF does not reset the EVP_MD_CTX. */
Add "origin" field to EVP_CIPHER, EVP_MD Add a "where did this EVP_{CIPHER,MD} come from" flag: global, via fetch, or via EVP_{CIPHER,MD}_meth_new. Update EVP_{CIPHER,MD}_free to handle all three origins. The flag is deliberately right before some function pointers, so that compile-time failures (int/pointer) will occur, as opposed to taking a bit in the existing "flags" field. The "global variable" flag is non-zero, so the default case of using OPENSSL_zalloc (for provider ciphers), will do the right thing. Ref-counting is a no-op for Make up_ref no-op for global MD and CIPHER objects Deprecate EVP_MD_CTX_md(). Added EVP_MD_CTX_get0_md() (same semantics as the deprecated function) and EVP_MD_CTX_get1_md(). Likewise, deprecate EVP_CIPHER_CTX_cipher() in favor of EVP_CIPHER_CTX_get0_cipher(), and add EVP_CIPHER_CTX_get1_CIPHER(). Refactor EVP_MD_free() and EVP_MD_meth_free() to call new common evp_md_free_int() function. Refactor EVP_CIPHER_free() and EVP_CIPHER_meth_free() to call new common evp_cipher_free_int() function. Also change some flags tests to explicit test == or != zero. E.g., if (flags & x) --> if ((flags & x) != 0) if (!(flags & x)) --> if ((flags & x) == 0) Only done for those lines where "get0_cipher" calls were made. Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14193)
2021-02-17 06:51:56 +08:00
|| !TEST_ptr(EVP_MD_CTX_get0_md(md_ctx))
Add test for no reset after DigestFinal_ex and DigestFinalXOF Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com> Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/13402)
2020-11-13 21:16:35 +08:00
|| !TEST_true(EVP_DigestInit_ex(md_ctx, NULL, NULL)))
goto out;
ret = 1;
out:
EVP_MD_CTX_free(md_ctx);
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
EVP_MD_free(sha256);
EVP_MD_free(shake256);
Add test for no reset after DigestFinal_ex and DigestFinalXOF Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com> Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/13402)
2020-11-13 21:16:35 +08:00
return ret;
}
Add null digest implementation to the default provider This is necessary to keep compatibility with 1.1.1. Fixes #16660 Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/17016)
2021-11-12 23:31:35 +08:00
static int test_EVP_md_null(void)
{
int ret = 0;
EVP_MD_CTX *md_ctx = NULL;
const EVP_MD *md_null = EVP_md_null();
unsigned char md_value[EVP_MAX_MD_SIZE];
unsigned int md_len = sizeof(md_value);
if (nullprov != NULL)
return TEST_skip("Test does not support a non-default library context");
if (!TEST_ptr(md_null)
|| !TEST_ptr(md_ctx = EVP_MD_CTX_new()))
goto out;
if (!TEST_true(EVP_DigestInit_ex(md_ctx, md_null, NULL))
|| !TEST_true(EVP_DigestUpdate(md_ctx, "test", 4))
|| !TEST_true(EVP_DigestFinal_ex(md_ctx, md_value, &md_len)))
goto out;
if (!TEST_uint_eq(md_len, 0))
goto out;
ret = 1;
out:
EVP_MD_CTX_free(md_ctx);
return ret;
}
Convert more tests ct_test,evp_extra_test,wpackettest,packettest Add strncmp TEST wrappers And make some style/consistency fixes to ct_test Silence travis; gcc bug? Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/3234)
2017-04-19 04:33:15 +08:00
static int test_d2i_AutoPrivateKey(int i)
Import evp_test.c from BoringSSL. Unfortunately we already have a file called evp_test.c, so I have called this one evp_extra_test.c Reviewed-by: Emilia Käsper <emilia@openssl.org>
2015-02-09 17:45:35 +08:00
{
int ret = 0;
const unsigned char *p;
EVP_PKEY *pkey = NULL;
Convert more tests ct_test,evp_extra_test,wpackettest,packettest Add strncmp TEST wrappers And make some style/consistency fixes to ct_test Silence travis; gcc bug? Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/3234)
2017-04-19 04:33:15 +08:00
const APK_DATA *ak = &keydata[i];
const unsigned char *input = ak->kder;
size_t input_len = ak->size;
int expected_id = ak->evptype;
Import evp_test.c from BoringSSL. Unfortunately we already have a file called evp_test.c, so I have called this one evp_extra_test.c Reviewed-by: Emilia Käsper <emilia@openssl.org>
2015-02-09 17:45:35 +08:00
p = input;
Convert more tests ct_test,evp_extra_test,wpackettest,packettest Add strncmp TEST wrappers And make some style/consistency fixes to ct_test Silence travis; gcc bug? Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/3234)
2017-04-19 04:33:15 +08:00
if (!TEST_ptr(pkey = d2i_AutoPrivateKey(NULL, &p, input_len))
|| !TEST_ptr_eq(p, input + input_len)
Rename all getters to use get/get0 in name For functions that exist in 1.1.1 provide a simple aliases via #define. Fixes #15236 Functions with OSSL_DECODER_, OSSL_ENCODER_, OSSL_STORE_LOADER_, EVP_KEYEXCH_, EVP_KEM_, EVP_ASYM_CIPHER_, EVP_SIGNATURE_, EVP_KEYMGMT_, EVP_RAND_, EVP_MAC_, EVP_KDF_, EVP_PKEY_, EVP_MD_, and EVP_CIPHER_ prefixes are renamed. Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/15405)
2021-05-21 22:58:08 +08:00
|| !TEST_int_eq(EVP_PKEY_get_id(pkey), expected_id))
Import evp_test.c from BoringSSL. Unfortunately we already have a file called evp_test.c, so I have called this one evp_extra_test.c Reviewed-by: Emilia Käsper <emilia@openssl.org>
2015-02-09 17:45:35 +08:00
goto done;
ret = 1;
done:
free NULL cleanup EVP_.*free; this gets: EVP_CIPHER_CTX_free EVP_PKEY_CTX_free EVP_PKEY_asn1_free EVP_PKEY_asn1_set_free EVP_PKEY_free EVP_PKEY_free_it EVP_PKEY_meth_free; and also EVP_CIPHER_CTX_cleanup Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2015-03-28 22:54:15 +08:00
EVP_PKEY_free(pkey);
Import evp_test.c from BoringSSL. Unfortunately we already have a file called evp_test.c, so I have called this one evp_extra_test.c Reviewed-by: Emilia Käsper <emilia@openssl.org>
2015-02-09 17:45:35 +08:00
return ret;
}
Fix evp_extra_test.c with no-ec When OpenSSL is configured with no-ec, then the new evp_extra_test fails to pass. This change adds appropriate OPENSSL_NO_EC guards around the code. Reviewed-by: Tim Hudson <tjh@openssl.org>
2015-02-26 18:35:50 +08:00
#ifndef OPENSSL_NO_EC
fixed public range check in ec_GF2m_simple_oct2point Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de> (Merged from https://github.com/openssl/openssl/pull/8607)
2019-03-27 15:38:28 +08:00
static const unsigned char ec_public_sect163k1_validxy[] = {
0x30, 0x40, 0x30, 0x10, 0x06, 0x07, 0x2a, 0x86, 0x48, 0xce, 0x3d, 0x02,
0x01, 0x06, 0x05, 0x2b, 0x81, 0x04, 0x00, 0x01, 0x03, 0x2c, 0x00, 0x04,
0x02, 0x84, 0x58, 0xa6, 0xd4, 0xa0, 0x35, 0x2b, 0xae, 0xf0, 0xc0, 0x69,
0x05, 0xcf, 0x2a, 0x50, 0x33, 0xf9, 0xe3, 0x92, 0x79, 0x02, 0xd1, 0x7b,
0x9f, 0x22, 0x00, 0xf0, 0x3b, 0x0e, 0x5d, 0x2e, 0xb7, 0x23, 0x24, 0xf3,
0x6a, 0xd8, 0x17, 0x65, 0x41, 0x2f
};
static const unsigned char ec_public_sect163k1_badx[] = {
0x30, 0x40, 0x30, 0x10, 0x06, 0x07, 0x2a, 0x86, 0x48, 0xce, 0x3d, 0x02,
0x01, 0x06, 0x05, 0x2b, 0x81, 0x04, 0x00, 0x01, 0x03, 0x2c, 0x00, 0x04,
0x0a, 0x84, 0x58, 0xa6, 0xd4, 0xa0, 0x35, 0x2b, 0xae, 0xf0, 0xc0, 0x69,
0x05, 0xcf, 0x2a, 0x50, 0x33, 0xf9, 0xe3, 0x92, 0xb0, 0x02, 0xd1, 0x7b,
0x9f, 0x22, 0x00, 0xf0, 0x3b, 0x0e, 0x5d, 0x2e, 0xb7, 0x23, 0x24, 0xf3,
0x6a, 0xd8, 0x17, 0x65, 0x41, 0x2f
};
static const unsigned char ec_public_sect163k1_bady[] = {
0x30, 0x40, 0x30, 0x10, 0x06, 0x07, 0x2a, 0x86, 0x48, 0xce, 0x3d, 0x02,
0x01, 0x06, 0x05, 0x2b, 0x81, 0x04, 0x00, 0x01, 0x03, 0x2c, 0x00, 0x04,
0x02, 0x84, 0x58, 0xa6, 0xd4, 0xa0, 0x35, 0x2b, 0xae, 0xf0, 0xc0, 0x69,
0x05, 0xcf, 0x2a, 0x50, 0x33, 0xf9, 0xe3, 0x92, 0x79, 0x0a, 0xd1, 0x7b,
0x9f, 0x22, 0x00, 0xf0, 0x3b, 0x0e, 0x5d, 0x2e, 0xb7, 0x23, 0x24, 0xf3,
0x6a, 0xd8, 0x17, 0x65, 0x41, 0xe6
};
static struct ec_der_pub_keys_st {
const unsigned char *der;
size_t len;
int valid;
} ec_der_pub_keys[] = {
{ ec_public_sect163k1_validxy, sizeof(ec_public_sect163k1_validxy), 1 },
{ ec_public_sect163k1_badx, sizeof(ec_public_sect163k1_badx), 0 },
{ ec_public_sect163k1_bady, sizeof(ec_public_sect163k1_bady), 0 },
};
/*
* Tests the range of the decoded EC char2 public point.
* See ec_GF2m_simple_oct2point().
*/
static int test_invalide_ec_char2_pub_range_decode(int id)
{
int ret = 0;
Deprecate EC_KEY + Update ec apps to use EVP_PKEY Co-author: Richard Levitte <levitte@openssl.org> Co-author: Tomas Mraz <tmraz@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13139)
2020-10-15 11:41:59 +08:00
EVP_PKEY *pkey;
fixed public range check in ec_GF2m_simple_oct2point Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de> (Merged from https://github.com/openssl/openssl/pull/8607)
2019-03-27 15:38:28 +08:00
Deprecate EC_KEY + Update ec apps to use EVP_PKEY Co-author: Richard Levitte <levitte@openssl.org> Co-author: Tomas Mraz <tmraz@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13139)
2020-10-15 11:41:59 +08:00
pkey = load_example_key("EC", ec_der_pub_keys[id].der,
ec_der_pub_keys[id].len);
ret = (ec_der_pub_keys[id].valid && TEST_ptr(pkey))
|| TEST_ptr_null(pkey);
EVP_PKEY_free(pkey);
fixed public range check in ec_GF2m_simple_oct2point Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de> (Merged from https://github.com/openssl/openssl/pull/8607)
2019-03-27 15:38:28 +08:00
return ret;
}
Import evp_test.c from BoringSSL. Unfortunately we already have a file called evp_test.c, so I have called this one evp_extra_test.c Reviewed-by: Emilia Käsper <emilia@openssl.org>
2015-02-09 17:45:35 +08:00
/* Tests loading a bad key in PKCS8 format */
static int test_EVP_PKCS82PKEY(void)
{
int ret = 0;
const unsigned char *derp = kExampleBadECKeyDER;
PKCS8_PRIV_KEY_INFO *p8inf = NULL;
EVP_PKEY *pkey = NULL;
Convert more tests ct_test,evp_extra_test,wpackettest,packettest Add strncmp TEST wrappers And make some style/consistency fixes to ct_test Silence travis; gcc bug? Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/3234)
2017-04-19 04:33:15 +08:00
if (!TEST_ptr(p8inf = d2i_PKCS8_PRIV_KEY_INFO(NULL, &derp,
sizeof(kExampleBadECKeyDER))))
goto done;
Import evp_test.c from BoringSSL. Unfortunately we already have a file called evp_test.c, so I have called this one evp_extra_test.c Reviewed-by: Emilia Käsper <emilia@openssl.org>
2015-02-09 17:45:35 +08:00
Convert more tests ct_test,evp_extra_test,wpackettest,packettest Add strncmp TEST wrappers And make some style/consistency fixes to ct_test Silence travis; gcc bug? Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/3234)
2017-04-19 04:33:15 +08:00
if (!TEST_ptr_eq(derp,
kExampleBadECKeyDER + sizeof(kExampleBadECKeyDER)))
Import evp_test.c from BoringSSL. Unfortunately we already have a file called evp_test.c, so I have called this one evp_extra_test.c Reviewed-by: Emilia Käsper <emilia@openssl.org>
2015-02-09 17:45:35 +08:00
goto done;
Convert more tests ct_test,evp_extra_test,wpackettest,packettest Add strncmp TEST wrappers And make some style/consistency fixes to ct_test Silence travis; gcc bug? Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/3234)
2017-04-19 04:33:15 +08:00
if (!TEST_ptr_null(pkey = EVP_PKCS82PKEY(p8inf)))
Import evp_test.c from BoringSSL. Unfortunately we already have a file called evp_test.c, so I have called this one evp_extra_test.c Reviewed-by: Emilia Käsper <emilia@openssl.org>
2015-02-09 17:45:35 +08:00
goto done;
ret = 1;
done:
free NULL cleanup 9 Ongoing work to skip NULL check before calling free routine. This gets: ecp_nistz256_pre_comp_free nistp224_pre_comp_free nistp256_pre_comp_free nistp521_pre_comp_free PKCS7_free PKCS7_RECIP_INFO_free PKCS7_SIGNER_INFO_free sk_PKCS7_pop_free PKCS8_PRIV_KEY_INFO_free PKCS12_free PKCS12_SAFEBAG_free PKCS12_free sk_PKCS12_SAFEBAG_pop_free SSL_CONF_CTX_free SSL_CTX_free SSL_SESSION_free SSL_free ssl_cert_free ssl_sess_cert_free Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2015-04-12 04:32:54 +08:00
PKCS8_PRIV_KEY_INFO_free(p8inf);
free NULL cleanup EVP_.*free; this gets: EVP_CIPHER_CTX_free EVP_PKEY_CTX_free EVP_PKEY_asn1_free EVP_PKEY_asn1_set_free EVP_PKEY_free EVP_PKEY_free_it EVP_PKEY_meth_free; and also EVP_CIPHER_CTX_cleanup Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2015-03-28 22:54:15 +08:00
EVP_PKEY_free(pkey);
Import evp_test.c from BoringSSL. Unfortunately we already have a file called evp_test.c, so I have called this one evp_extra_test.c Reviewed-by: Emilia Käsper <emilia@openssl.org>
2015-02-09 17:45:35 +08:00
return ret;
}
Mark pop/clear error stack in der2key_decode_p8 This commit sets the error mark before calling d2i_X509_SIG and clear it if that function call is successful. The motivation for this is that if d2i_X509_SIG returns NULL then the else clause will be entered and d2i_PKCS8_PRIV_KEY_INFO will be called. If d2i_X509_SIG raised any errors those error will be on the error stack when d2i_PKCS8_PRIV_KEY_INFO gets called, and even if it returns successfully those errors will still be on the error stack. We ran into this issue when upgrading Node.js to 3.0.0-alpha15. More details can be found in the ref links below. Refs: https://github.com/nodejs/node/issues/38373 Refs: https://github.com/danbev/learning-libcrypto/blob/master/notes/wrong-tag-issue2.md Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Ben Kaduk <kaduk@mit.edu> (Merged from https://github.com/openssl/openssl/pull/15067)
2021-04-28 16:30:13 +08:00
Fix evp_extra_test.c with no-ec When OpenSSL is configured with no-ec, then the new evp_extra_test fails to pass. This change adds appropriate OPENSSL_NO_EC guards around the code. Reviewed-by: Tim Hudson <tjh@openssl.org>
2015-02-26 18:35:50 +08:00
#endif
Mark pop/clear error stack in der2key_decode_p8 This commit sets the error mark before calling d2i_X509_SIG and clear it if that function call is successful. The motivation for this is that if d2i_X509_SIG returns NULL then the else clause will be entered and d2i_PKCS8_PRIV_KEY_INFO will be called. If d2i_X509_SIG raised any errors those error will be on the error stack when d2i_PKCS8_PRIV_KEY_INFO gets called, and even if it returns successfully those errors will still be on the error stack. We ran into this issue when upgrading Node.js to 3.0.0-alpha15. More details can be found in the ref links below. Refs: https://github.com/nodejs/node/issues/38373 Refs: https://github.com/danbev/learning-libcrypto/blob/master/notes/wrong-tag-issue2.md Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Ben Kaduk <kaduk@mit.edu> (Merged from https://github.com/openssl/openssl/pull/15067)
2021-04-28 16:30:13 +08:00
static int test_EVP_PKCS82PKEY_wrong_tag(void)
{
EVP_PKEY *pkey = NULL;
EVP_PKEY *pkey2 = NULL;
BIO *membio = NULL;
char *membuf = NULL;
PKCS8_PRIV_KEY_INFO *p8inf = NULL;
int ok = 0;
if (testctx != NULL)
/* test not supported with non-default context */
return 1;
if (!TEST_ptr(membio = BIO_new(BIO_s_mem()))
|| !TEST_ptr(pkey = load_example_rsa_key())
|| !TEST_int_gt(i2d_PKCS8PrivateKey_bio(membio, pkey, NULL,
NULL, 0, NULL, NULL),
0)
|| !TEST_int_gt(BIO_get_mem_data(membio, &membuf), 0)
|| !TEST_ptr(p8inf = d2i_PKCS8_PRIV_KEY_INFO_bio(membio, NULL))
|| !TEST_ptr(pkey2 = EVP_PKCS82PKEY(p8inf))
test/evp_extra_test.c: Peek at the error instead of getting it. If there is an error report, we want to get it printed too. Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/15604)
2021-06-09 17:00:00 +08:00
|| !TEST_int_eq(ERR_peek_last_error(), 0)) {
Mark pop/clear error stack in der2key_decode_p8 This commit sets the error mark before calling d2i_X509_SIG and clear it if that function call is successful. The motivation for this is that if d2i_X509_SIG returns NULL then the else clause will be entered and d2i_PKCS8_PRIV_KEY_INFO will be called. If d2i_X509_SIG raised any errors those error will be on the error stack when d2i_PKCS8_PRIV_KEY_INFO gets called, and even if it returns successfully those errors will still be on the error stack. We ran into this issue when upgrading Node.js to 3.0.0-alpha15. More details can be found in the ref links below. Refs: https://github.com/nodejs/node/issues/38373 Refs: https://github.com/danbev/learning-libcrypto/blob/master/notes/wrong-tag-issue2.md Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Ben Kaduk <kaduk@mit.edu> (Merged from https://github.com/openssl/openssl/pull/15067)
2021-04-28 16:30:13 +08:00
goto done;
}
ok = 1;
done:
EVP_PKEY_free(pkey);
EVP_PKEY_free(pkey2);
PKCS8_PRIV_KEY_INFO_free(p8inf);
BIO_free_all(membio);
return ok;
}
Import evp_test.c from BoringSSL. Unfortunately we already have a file called evp_test.c, so I have called this one evp_extra_test.c Reviewed-by: Emilia Käsper <emilia@openssl.org>
2015-02-09 17:45:35 +08:00
TEST: Test i2d_PKCS8PrivateKey_bio() and PEM_write_bio_PKCS8PrivateKey() Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/11855)
2020-05-18 14:35:29 +08:00
/* This uses kExampleRSAKeyDER and kExampleRSAKeyPKCS8 to verify encoding */
static int test_privatekey_to_pkcs8(void)
{
EVP_PKEY *pkey = NULL;
BIO *membio = NULL;
char *membuf = NULL;
Fix CID 1465213: Integer handling issues (evp_extra_test.c) Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com> (Merged from https://github.com/openssl/openssl/pull/12379)
2020-07-06 15:35:23 +08:00
long membuf_len = 0;
TEST: Test i2d_PKCS8PrivateKey_bio() and PEM_write_bio_PKCS8PrivateKey() Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/11855)
2020-05-18 14:35:29 +08:00
int ok = 0;
if (!TEST_ptr(membio = BIO_new(BIO_s_mem()))
|| !TEST_ptr(pkey = load_example_rsa_key())
|| !TEST_int_gt(i2d_PKCS8PrivateKey_bio(membio, pkey, NULL,
NULL, 0, NULL, NULL),
0)
Fix CID 1465213: Integer handling issues (evp_extra_test.c) Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com> (Merged from https://github.com/openssl/openssl/pull/12379)
2020-07-06 15:35:23 +08:00
|| !TEST_int_gt(membuf_len = BIO_get_mem_data(membio, &membuf), 0)
|| !TEST_ptr(membuf)
|| !TEST_mem_eq(membuf, (size_t)membuf_len,
TEST: Test i2d_PKCS8PrivateKey_bio() and PEM_write_bio_PKCS8PrivateKey() Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/11855)
2020-05-18 14:35:29 +08:00
kExampleRSAKeyPKCS8, sizeof(kExampleRSAKeyPKCS8))
/*
* We try to write PEM as well, just to see that it doesn't err, but
* assume that the result is correct.
*/
|| !TEST_int_gt(PEM_write_bio_PKCS8PrivateKey(membio, pkey, NULL,
NULL, 0, NULL, NULL),
0))
goto done;
ok = 1;
done:
EVP_PKEY_free(pkey);
BIO_free_all(membio);
return ok;
}
TEST: Add a test of EC key generation with encoding spec The primary function of this test is to exercise EVP_PKEY_CTX_set_ec_param_enc(). Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com> Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/12853)
2020-09-11 14:36:54 +08:00
#ifndef OPENSSL_NO_EC
static const struct {
int encoding;
const char *encoding_name;
} ec_encodings[] = {
{ OPENSSL_EC_EXPLICIT_CURVE, OSSL_PKEY_EC_ENCODING_EXPLICIT },
{ OPENSSL_EC_NAMED_CURVE, OSSL_PKEY_EC_ENCODING_GROUP }
};
static int ec_export_get_encoding_cb(const OSSL_PARAM params[], void *arg)
{
const OSSL_PARAM *p;
const char *enc_name = NULL;
int *enc = arg;
size_t i;
*enc = -1;
if (!TEST_ptr(p = OSSL_PARAM_locate_const(params,
OSSL_PKEY_PARAM_EC_ENCODING))
|| !TEST_true(OSSL_PARAM_get_utf8_string_ptr(p, &enc_name)))
return 0;
for (i = 0; i < OSSL_NELEM(ec_encodings); i++) {
str[n]casecmp => OPENSSL_strncasecmp Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/18069)
2022-04-12 18:30:08 +08:00
if (OPENSSL_strcasecmp(enc_name, ec_encodings[i].encoding_name) == 0) {
TEST: Add a test of EC key generation with encoding spec The primary function of this test is to exercise EVP_PKEY_CTX_set_ec_param_enc(). Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com> Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/12853)
2020-09-11 14:36:54 +08:00
*enc = ec_encodings[i].encoding;
break;
}
}
return (*enc != -1);
}
static int test_EC_keygen_with_enc(int idx)
{
EVP_PKEY *params = NULL, *key = NULL;
EVP_PKEY_CTX *pctx = NULL, *kctx = NULL;
int enc;
int ret = 0;
enc = ec_encodings[idx].encoding;
/* Create key parameters */
if (!TEST_ptr(pctx = EVP_PKEY_CTX_new_from_name(testctx, "EC", NULL))
EVP_PKEY_paramgen_init: fix return check Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/17030)
2021-11-14 23:52:56 +08:00
|| !TEST_int_gt(EVP_PKEY_paramgen_init(pctx), 0)
Fix the erroneous checks of EVP_PKEY_CTX_set_group_name Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/18399)
2022-05-24 23:27:49 +08:00
|| !TEST_int_gt(EVP_PKEY_CTX_set_group_name(pctx, "P-256"), 0)
Fix the checks of EVP_PKEY_CTX_set/get_* functions Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/18399)
2022-05-24 23:57:33 +08:00
|| !TEST_int_gt(EVP_PKEY_CTX_set_ec_param_enc(pctx, enc), 0)
TEST: Add a test of EC key generation with encoding spec The primary function of this test is to exercise EVP_PKEY_CTX_set_ec_param_enc(). Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com> Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/12853)
2020-09-11 14:36:54 +08:00
|| !TEST_true(EVP_PKEY_paramgen(pctx, &params))
|| !TEST_ptr(params))
goto done;
/* Create key */
if (!TEST_ptr(kctx = EVP_PKEY_CTX_new_from_pkey(testctx, params, NULL))
EVP_PKEY_keygen_init: fix return check Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/17031)
2021-11-15 00:05:04 +08:00
|| !TEST_int_gt(EVP_PKEY_keygen_init(kctx), 0)
TEST: Add a test of EC key generation with encoding spec The primary function of this test is to exercise EVP_PKEY_CTX_set_ec_param_enc(). Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com> Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/12853)
2020-09-11 14:36:54 +08:00
|| !TEST_true(EVP_PKEY_keygen(kctx, &key))
|| !TEST_ptr(key))
goto done;
/* Check that the encoding got all the way into the key */
if (!TEST_true(evp_keymgmt_util_export(key, OSSL_KEYMGMT_SELECT_ALL,
ec_export_get_encoding_cb, &enc))
|| !TEST_int_eq(enc, ec_encodings[idx].encoding))
goto done;
ret = 1;
done:
EVP_PKEY_free(key);
EVP_PKEY_free(params);
EVP_PKEY_CTX_free(kctx);
EVP_PKEY_CTX_free(pctx);
return ret;
}
#endif
Remove FIPS condition on SM2 test. Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/19237)
2022-09-19 08:28:25 +08:00
#if !defined(OPENSSL_NO_SM2)
Move SM2 algos to SM2 specific PKEY method Use EVP_PKEY_set_alias_type to access Reviewed-by: Andy Polyakov <appro@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6443)
2018-06-19 03:51:56 +08:00
Add test case for SM2 evp verification This test case is originally submitted in #6757, by Jack Lloyd. The test case has been modified to use the a different method to set the ID when computing the Z hash of SM2 signature. Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Yang <yang.yang@baishancloud.com> (Merged from https://github.com/openssl/openssl/pull/7113)
2018-09-04 23:25:29 +08:00
static int test_EVP_SM2_verify(void)
{
const char *pubkey =
Add SM2 signature algorithm to default provider Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/12536)
2020-03-04 23:49:43 +08:00
"-----BEGIN PUBLIC KEY-----\n"
"MFkwEwYHKoZIzj0CAQYIKoEcz1UBgi0DQgAEp1KLWq1ZE2jmoAnnBJE1LBGxVr18\n"
"YvvqECWCpXfAQ9qUJ+UmthnUPf0iM3SaXKHe6PlLIDyNlWMWb9RUh/yU3g==\n"
"-----END PUBLIC KEY-----\n";
Add test case for SM2 evp verification This test case is originally submitted in #6757, by Jack Lloyd. The test case has been modified to use the a different method to set the ID when computing the Z hash of SM2 signature. Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Yang <yang.yang@baishancloud.com> (Merged from https://github.com/openssl/openssl/pull/7113)
2018-09-04 23:25:29 +08:00
const char *msg = "message digest";
const char *id = "ALICE123@YAHOO.COM";
const uint8_t signature[] = {
Add SM2 signature algorithm to default provider Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/12536)
2020-03-04 23:49:43 +08:00
0x30, 0x44, 0x02, 0x20, 0x5b, 0xdb, 0xab, 0x81, 0x4f, 0xbb,
0x8b, 0x69, 0xb1, 0x05, 0x9c, 0x99, 0x3b, 0xb2, 0x45, 0x06,
0x4a, 0x30, 0x15, 0x59, 0x84, 0xcd, 0xee, 0x30, 0x60, 0x36,
0x57, 0x87, 0xef, 0x5c, 0xd0, 0xbe, 0x02, 0x20, 0x43, 0x8d,
0x1f, 0xc7, 0x77, 0x72, 0x39, 0xbb, 0x72, 0xe1, 0xfd, 0x07,
0x58, 0xd5, 0x82, 0xc8, 0x2d, 0xba, 0x3b, 0x2c, 0x46, 0x24,
0xe3, 0x50, 0xff, 0x04, 0xc7, 0xa0, 0x71, 0x9f, 0xa4, 0x70
Add test case for SM2 evp verification This test case is originally submitted in #6757, by Jack Lloyd. The test case has been modified to use the a different method to set the ID when computing the Z hash of SM2 signature. Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Yang <yang.yang@baishancloud.com> (Merged from https://github.com/openssl/openssl/pull/7113)
2018-09-04 23:25:29 +08:00
};
int rc = 0;
BIO *bio = NULL;
EVP_PKEY *pkey = NULL;
EVP_MD_CTX *mctx = NULL;
EVP_PKEY_CTX *pctx = NULL;
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
EVP_MD *sm3 = NULL;
Add test case for SM2 evp verification This test case is originally submitted in #6757, by Jack Lloyd. The test case has been modified to use the a different method to set the ID when computing the Z hash of SM2 signature. Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Yang <yang.yang@baishancloud.com> (Merged from https://github.com/openssl/openssl/pull/7113)
2018-09-04 23:25:29 +08:00
bio = BIO_new_mem_buf(pubkey, strlen(pubkey));
if (!TEST_true(bio != NULL))
goto done;
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
pkey = PEM_read_bio_PUBKEY_ex(bio, NULL, NULL, NULL, testctx, testpropq);
Add test case for SM2 evp verification This test case is originally submitted in #6757, by Jack Lloyd. The test case has been modified to use the a different method to set the ID when computing the Z hash of SM2 signature. Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Yang <yang.yang@baishancloud.com> (Merged from https://github.com/openssl/openssl/pull/7113)
2018-09-04 23:25:29 +08:00
if (!TEST_true(pkey != NULL))
goto done;
TEST: Remove use of EVP_PKEY_set_alias_type() in test/evp_extra_test.c We already test EVP_PKEY_set_alias_type() quite thoroughly in test/ecdsatest.c, that should be enough. Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/12920)
2020-09-23 12:18:06 +08:00
if (!TEST_true(EVP_PKEY_is_a(pkey, "SM2")))
Add test case for SM2 evp verification This test case is originally submitted in #6757, by Jack Lloyd. The test case has been modified to use the a different method to set the ID when computing the Z hash of SM2 signature. Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Yang <yang.yang@baishancloud.com> (Merged from https://github.com/openssl/openssl/pull/7113)
2018-09-04 23:25:29 +08:00
goto done;
if (!TEST_ptr(mctx = EVP_MD_CTX_new()))
goto done;
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
if (!TEST_ptr(pctx = EVP_PKEY_CTX_new_from_pkey(testctx, pkey, testpropq)))
Add test case for SM2 evp verification This test case is originally submitted in #6757, by Jack Lloyd. The test case has been modified to use the a different method to set the ID when computing the Z hash of SM2 signature. Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Yang <yang.yang@baishancloud.com> (Merged from https://github.com/openssl/openssl/pull/7113)
2018-09-04 23:25:29 +08:00
goto done;
EVP_MD_CTX_set_pkey_ctx(mctx, pctx);
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
if (!TEST_ptr(sm3 = EVP_MD_fetch(testctx, "sm3", testpropq)))
goto done;
if (!TEST_true(EVP_DigestVerifyInit(mctx, NULL, sm3, NULL, pkey)))
Add test case for SM2 evp verification This test case is originally submitted in #6757, by Jack Lloyd. The test case has been modified to use the a different method to set the ID when computing the Z hash of SM2 signature. Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Yang <yang.yang@baishancloud.com> (Merged from https://github.com/openssl/openssl/pull/7113)
2018-09-04 23:25:29 +08:00
goto done;
Add SM2 signature algorithm to default provider Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/12536)
2020-03-04 23:49:43 +08:00
if (!TEST_int_gt(EVP_PKEY_CTX_set1_id(pctx, id, strlen(id)), 0))
goto done;
Add test case for SM2 evp verification This test case is originally submitted in #6757, by Jack Lloyd. The test case has been modified to use the a different method to set the ID when computing the Z hash of SM2 signature. Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Yang <yang.yang@baishancloud.com> (Merged from https://github.com/openssl/openssl/pull/7113)
2018-09-04 23:25:29 +08:00
if (!TEST_true(EVP_DigestVerifyUpdate(mctx, msg, strlen(msg))))
goto done;
EVP_DigestVerifyFinal: fix test function and invocation Signed-off-by: Peiwei Hu <jlu.hpw@foxmail.com> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/17029)
2021-11-14 22:42:35 +08:00
if (!TEST_int_gt(EVP_DigestVerifyFinal(mctx, signature, sizeof(signature)), 0))
Add test case for SM2 evp verification This test case is originally submitted in #6757, by Jack Lloyd. The test case has been modified to use the a different method to set the ID when computing the Z hash of SM2 signature. Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Yang <yang.yang@baishancloud.com> (Merged from https://github.com/openssl/openssl/pull/7113)
2018-09-04 23:25:29 +08:00
goto done;
rc = 1;
done:
BIO_free(bio);
EVP_PKEY_free(pkey);
EVP_PKEY_CTX_free(pctx);
EVP_MD_CTX_free(mctx);
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
EVP_MD_free(sm3);
Add test case for SM2 evp verification This test case is originally submitted in #6757, by Jack Lloyd. The test case has been modified to use the a different method to set the ID when computing the Z hash of SM2 signature. Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Yang <yang.yang@baishancloud.com> (Merged from https://github.com/openssl/openssl/pull/7113)
2018-09-04 23:25:29 +08:00
return rc;
}
Move SM2 algos to SM2 specific PKEY method Use EVP_PKEY_set_alias_type to access Reviewed-by: Andy Polyakov <appro@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6443)
2018-06-19 03:51:56 +08:00
static int test_EVP_SM2(void)
{
int ret = 0;
EVP_PKEY *pkey = NULL;
Extend the SM2 asym cipher test Ensure we test getting and setting ctx params Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/12913)
2020-09-18 18:57:24 +08:00
EVP_PKEY *pkeyparams = NULL;
Move SM2 algos to SM2 specific PKEY method Use EVP_PKEY_set_alias_type to access Reviewed-by: Andy Polyakov <appro@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6443)
2018-06-19 03:51:56 +08:00
EVP_PKEY_CTX *pctx = NULL;
EVP_PKEY_CTX *kctx = NULL;
Support setting SM2 ID zero-length ID is allowed, but it's not allowed to skip the ID. Fixes: #6534 Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7113)
2018-09-04 17:21:10 +08:00
EVP_PKEY_CTX *sctx = NULL;
Move SM2 algos to SM2 specific PKEY method Use EVP_PKEY_set_alias_type to access Reviewed-by: Andy Polyakov <appro@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6443)
2018-06-19 03:51:56 +08:00
size_t sig_len = 0;
unsigned char *sig = NULL;
EVP_MD_CTX *md_ctx = NULL;
EVP_MD_CTX *md_ctx_verify = NULL;
EVP_PKEY_CTX *cctx = NULL;
TEST: Modify how the retrieved digest name for SM2 digestsign is checked Because of aliases, retrieved names won't always match one specific string. A safer way to check is to fetch the digest from the retrieved name and check it's the expected one with the help of EVP_MD_is_a(). Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14498)
2021-03-10 19:53:51 +08:00
EVP_MD *check_md = NULL;
Move SM2 algos to SM2 specific PKEY method Use EVP_PKEY_set_alias_type to access Reviewed-by: Andy Polyakov <appro@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6443)
2018-06-19 03:51:56 +08:00
uint8_t ciphertext[128];
size_t ctext_len = sizeof(ciphertext);
uint8_t plaintext[8];
size_t ptext_len = sizeof(plaintext);
Support setting SM2 ID zero-length ID is allowed, but it's not allowed to skip the ID. Fixes: #6534 Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7113)
2018-09-04 17:21:10 +08:00
uint8_t sm2_id[] = {1, 2, 3, 4, 'l', 'e', 't', 't', 'e', 'r'};
Extend the SM2 asym cipher test Ensure we test getting and setting ctx params Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/12913)
2020-09-18 18:57:24 +08:00
OSSL_PARAM sparams[2] = {OSSL_PARAM_END, OSSL_PARAM_END};
OSSL_PARAM gparams[2] = {OSSL_PARAM_END, OSSL_PARAM_END};
int i;
TEST: Use OSSL_MAX_NAME_SIZE instead of arbitrary number of mdname Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14498)
2021-03-11 23:04:16 +08:00
char mdname[OSSL_MAX_NAME_SIZE];
Extend the SM2 asym cipher test Ensure we test getting and setting ctx params Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/12913)
2020-09-18 18:57:24 +08:00
TEST: Modify how the retrieved digest name for SM2 digestsign is checked Because of aliases, retrieved names won't always match one specific string. A safer way to check is to fetch the digest from the retrieved name and check it's the expected one with the help of EVP_MD_is_a(). Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14498)
2021-03-10 19:53:51 +08:00
if (!TEST_ptr(pctx = EVP_PKEY_CTX_new_from_name(testctx,
"SM2", testpropq)))
Move SM2 algos to SM2 specific PKEY method Use EVP_PKEY_set_alias_type to access Reviewed-by: Andy Polyakov <appro@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6443)
2018-06-19 03:51:56 +08:00
goto done;
if (!TEST_true(EVP_PKEY_paramgen_init(pctx) == 1))
goto done;
Fix the checks of EVP_PKEY_CTX_set/get_* functions Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/18399)
2022-05-24 23:57:33 +08:00
if (!TEST_int_gt(EVP_PKEY_CTX_set_ec_paramgen_curve_nid(pctx, NID_sm2), 0))
Move SM2 algos to SM2 specific PKEY method Use EVP_PKEY_set_alias_type to access Reviewed-by: Andy Polyakov <appro@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6443)
2018-06-19 03:51:56 +08:00
goto done;
Extend the SM2 asym cipher test Ensure we test getting and setting ctx params Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/12913)
2020-09-18 18:57:24 +08:00
if (!TEST_true(EVP_PKEY_paramgen(pctx, &pkeyparams)))
Move SM2 algos to SM2 specific PKEY method Use EVP_PKEY_set_alias_type to access Reviewed-by: Andy Polyakov <appro@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6443)
2018-06-19 03:51:56 +08:00
goto done;
TEST: Modify how the retrieved digest name for SM2 digestsign is checked Because of aliases, retrieved names won't always match one specific string. A safer way to check is to fetch the digest from the retrieved name and check it's the expected one with the help of EVP_MD_is_a(). Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14498)
2021-03-10 19:53:51 +08:00
if (!TEST_ptr(kctx = EVP_PKEY_CTX_new_from_pkey(testctx,
pkeyparams, testpropq)))
Move SM2 algos to SM2 specific PKEY method Use EVP_PKEY_set_alias_type to access Reviewed-by: Andy Polyakov <appro@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6443)
2018-06-19 03:51:56 +08:00
goto done;
EVP_PKEY_keygen_init: fix return check Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/17031)
2021-11-15 00:05:04 +08:00
if (!TEST_int_gt(EVP_PKEY_keygen_init(kctx), 0))
Move SM2 algos to SM2 specific PKEY method Use EVP_PKEY_set_alias_type to access Reviewed-by: Andy Polyakov <appro@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6443)
2018-06-19 03:51:56 +08:00
goto done;
if (!TEST_true(EVP_PKEY_keygen(kctx, &pkey)))
goto done;
if (!TEST_ptr(md_ctx = EVP_MD_CTX_new()))
goto done;
if (!TEST_ptr(md_ctx_verify = EVP_MD_CTX_new()))
goto done;
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
if (!TEST_ptr(sctx = EVP_PKEY_CTX_new_from_pkey(testctx, pkey, testpropq)))
Support setting SM2 ID zero-length ID is allowed, but it's not allowed to skip the ID. Fixes: #6534 Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7113)
2018-09-04 17:21:10 +08:00
goto done;
EVP_MD_CTX_set_pkey_ctx(md_ctx, sctx);
EVP_MD_CTX_set_pkey_ctx(md_ctx_verify, sctx);
TEST: Modify how the retrieved digest name for SM2 digestsign is checked Because of aliases, retrieved names won't always match one specific string. A safer way to check is to fetch the digest from the retrieved name and check it's the expected one with the help of EVP_MD_is_a(). Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14498)
2021-03-10 19:53:51 +08:00
if (!TEST_ptr(check_md = EVP_MD_fetch(testctx, "sm3", testpropq)))
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
goto done;
TEST: Modify how the retrieved digest name for SM2 digestsign is checked Because of aliases, retrieved names won't always match one specific string. A safer way to check is to fetch the digest from the retrieved name and check it's the expected one with the help of EVP_MD_is_a(). Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14498)
2021-03-10 19:53:51 +08:00
if (!TEST_true(EVP_DigestSignInit(md_ctx, NULL, check_md, NULL, pkey)))
Support setting SM2 ID zero-length ID is allowed, but it's not allowed to skip the ID. Fixes: #6534 Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7113)
2018-09-04 17:21:10 +08:00
goto done;
Add SM2 signature algorithm to default provider Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/12536)
2020-03-04 23:49:43 +08:00
if (!TEST_int_gt(EVP_PKEY_CTX_set1_id(sctx, sm2_id, sizeof(sm2_id)), 0))
Move SM2 algos to SM2 specific PKEY method Use EVP_PKEY_set_alias_type to access Reviewed-by: Andy Polyakov <appro@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6443)
2018-06-19 03:51:56 +08:00
goto done;
TEST: Modify how the retrieved digest name for SM2 digestsign is checked Because of aliases, retrieved names won't always match one specific string. A safer way to check is to fetch the digest from the retrieved name and check it's the expected one with the help of EVP_MD_is_a(). Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14498)
2021-03-10 19:53:51 +08:00
if (!TEST_true(EVP_DigestSignUpdate(md_ctx, kMsg, sizeof(kMsg))))
Move SM2 algos to SM2 specific PKEY method Use EVP_PKEY_set_alias_type to access Reviewed-by: Andy Polyakov <appro@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6443)
2018-06-19 03:51:56 +08:00
goto done;
/* Determine the size of the signature. */
if (!TEST_true(EVP_DigestSignFinal(md_ctx, NULL, &sig_len)))
goto done;
if (!TEST_ptr(sig = OPENSSL_malloc(sig_len)))
goto done;
if (!TEST_true(EVP_DigestSignFinal(md_ctx, sig, &sig_len)))
goto done;
/* Ensure that the signature round-trips. */
TEST: Modify how the retrieved digest name for SM2 digestsign is checked Because of aliases, retrieved names won't always match one specific string. A safer way to check is to fetch the digest from the retrieved name and check it's the expected one with the help of EVP_MD_is_a(). Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14498)
2021-03-10 19:53:51 +08:00
if (!TEST_true(EVP_DigestVerifyInit(md_ctx_verify, NULL, check_md, NULL,
pkey)))
Move SM2 algos to SM2 specific PKEY method Use EVP_PKEY_set_alias_type to access Reviewed-by: Andy Polyakov <appro@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6443)
2018-06-19 03:51:56 +08:00
goto done;
Add SM2 signature algorithm to default provider Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/12536)
2020-03-04 23:49:43 +08:00
if (!TEST_int_gt(EVP_PKEY_CTX_set1_id(sctx, sm2_id, sizeof(sm2_id)), 0))
goto done;
Move SM2 algos to SM2 specific PKEY method Use EVP_PKEY_set_alias_type to access Reviewed-by: Andy Polyakov <appro@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6443)
2018-06-19 03:51:56 +08:00
if (!TEST_true(EVP_DigestVerifyUpdate(md_ctx_verify, kMsg, sizeof(kMsg))))
goto done;
EVP_DigestVerifyFinal: fix test function and invocation Signed-off-by: Peiwei Hu <jlu.hpw@foxmail.com> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/17029)
2021-11-14 22:42:35 +08:00
if (!TEST_int_gt(EVP_DigestVerifyFinal(md_ctx_verify, sig, sig_len), 0))
Move SM2 algos to SM2 specific PKEY method Use EVP_PKEY_set_alias_type to access Reviewed-by: Andy Polyakov <appro@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6443)
2018-06-19 03:51:56 +08:00
goto done;
Add test for resetting SM2 dist ID Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/18052)
2022-04-06 16:29:54 +08:00
/*
* Try verify again with non-matching 0 length id but ensure that it can
* be set on the context and overrides the previous value.
*/
if (!TEST_true(EVP_DigestVerifyInit(md_ctx_verify, NULL, check_md, NULL,
pkey)))
goto done;
if (!TEST_int_gt(EVP_PKEY_CTX_set1_id(sctx, NULL, 0), 0))
goto done;
if (!TEST_true(EVP_DigestVerifyUpdate(md_ctx_verify, kMsg, sizeof(kMsg))))
goto done;
if (!TEST_int_eq(EVP_DigestVerifyFinal(md_ctx_verify, sig, sig_len), 0))
goto done;
Move SM2 algos to SM2 specific PKEY method Use EVP_PKEY_set_alias_type to access Reviewed-by: Andy Polyakov <appro@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6443)
2018-06-19 03:51:56 +08:00
/* now check encryption/decryption */
Extend the SM2 asym cipher test Ensure we test getting and setting ctx params Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/12913)
2020-09-18 18:57:24 +08:00
gparams[0] = OSSL_PARAM_construct_utf8_string(OSSL_ASYM_CIPHER_PARAM_DIGEST,
mdname, sizeof(mdname));
for (i = 0; i < 2; i++) {
TEST: Modify how the retrieved digest name for SM2 digestsign is checked Because of aliases, retrieved names won't always match one specific string. A safer way to check is to fetch the digest from the retrieved name and check it's the expected one with the help of EVP_MD_is_a(). Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14498)
2021-03-10 19:53:51 +08:00
const char *mdnames[] = {
#ifndef OPENSSL_NO_SM3
"SM3",
#else
NULL,
#endif
"SHA2-256" };
Extend the SM2 asym cipher test Ensure we test getting and setting ctx params Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/12913)
2020-09-18 18:57:24 +08:00
EVP_PKEY_CTX_free(cctx);
Move SM2 algos to SM2 specific PKEY method Use EVP_PKEY_set_alias_type to access Reviewed-by: Andy Polyakov <appro@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6443)
2018-06-19 03:51:56 +08:00
TEST: Modify how the retrieved digest name for SM2 digestsign is checked Because of aliases, retrieved names won't always match one specific string. A safer way to check is to fetch the digest from the retrieved name and check it's the expected one with the help of EVP_MD_is_a(). Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14498)
2021-03-10 19:53:51 +08:00
if (mdnames[i] == NULL)
continue;
sparams[0] =
OSSL_PARAM_construct_utf8_string(OSSL_ASYM_CIPHER_PARAM_DIGEST,
(char *)mdnames[i], 0);
Move SM2 algos to SM2 specific PKEY method Use EVP_PKEY_set_alias_type to access Reviewed-by: Andy Polyakov <appro@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6443)
2018-06-19 03:51:56 +08:00
TEST: Modify how the retrieved digest name for SM2 digestsign is checked Because of aliases, retrieved names won't always match one specific string. A safer way to check is to fetch the digest from the retrieved name and check it's the expected one with the help of EVP_MD_is_a(). Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14498)
2021-03-10 19:53:51 +08:00
if (!TEST_ptr(cctx = EVP_PKEY_CTX_new_from_pkey(testctx,
pkey, testpropq)))
Extend the SM2 asym cipher test Ensure we test getting and setting ctx params Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/12913)
2020-09-18 18:57:24 +08:00
goto done;
Move SM2 algos to SM2 specific PKEY method Use EVP_PKEY_set_alias_type to access Reviewed-by: Andy Polyakov <appro@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6443)
2018-06-19 03:51:56 +08:00
Extend the SM2 asym cipher test Ensure we test getting and setting ctx params Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/12913)
2020-09-18 18:57:24 +08:00
if (!TEST_true(EVP_PKEY_encrypt_init(cctx)))
goto done;
Move SM2 algos to SM2 specific PKEY method Use EVP_PKEY_set_alias_type to access Reviewed-by: Andy Polyakov <appro@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6443)
2018-06-19 03:51:56 +08:00
Extend the SM2 asym cipher test Ensure we test getting and setting ctx params Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/12913)
2020-09-18 18:57:24 +08:00
if (!TEST_true(EVP_PKEY_CTX_set_params(cctx, sparams)))
goto done;
Move SM2 algos to SM2 specific PKEY method Use EVP_PKEY_set_alias_type to access Reviewed-by: Andy Polyakov <appro@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6443)
2018-06-19 03:51:56 +08:00
Extend the SM2 asym cipher test Ensure we test getting and setting ctx params Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/12913)
2020-09-18 18:57:24 +08:00
if (!TEST_true(EVP_PKEY_encrypt(cctx, ciphertext, &ctext_len, kMsg,
sizeof(kMsg))))
goto done;
Move SM2 algos to SM2 specific PKEY method Use EVP_PKEY_set_alias_type to access Reviewed-by: Andy Polyakov <appro@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6443)
2018-06-19 03:51:56 +08:00
Fix the check of EVP_PKEY_decrypt_init Reviewed-by: Hugo Landau <hlandau@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/19819)
2022-12-02 16:33:02 +08:00
if (!TEST_int_gt(EVP_PKEY_decrypt_init(cctx), 0))
Extend the SM2 asym cipher test Ensure we test getting and setting ctx params Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/12913)
2020-09-18 18:57:24 +08:00
goto done;
if (!TEST_true(EVP_PKEY_CTX_set_params(cctx, sparams)))
goto done;
Fix EVP_PKEY_decrypt return check Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/17028)
2021-11-14 17:57:57 +08:00
if (!TEST_int_gt(EVP_PKEY_decrypt(cctx, plaintext, &ptext_len, ciphertext,
ctext_len), 0))
Extend the SM2 asym cipher test Ensure we test getting and setting ctx params Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/12913)
2020-09-18 18:57:24 +08:00
goto done;
if (!TEST_true(EVP_PKEY_CTX_get_params(cctx, gparams)))
goto done;
TEST: Modify how the retrieved digest name for SM2 digestsign is checked Because of aliases, retrieved names won't always match one specific string. A safer way to check is to fetch the digest from the retrieved name and check it's the expected one with the help of EVP_MD_is_a(). Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14498)
2021-03-10 19:53:51 +08:00
/*
* Test we're still using the digest we think we are.
* Because of aliases, the easiest is to fetch the digest and
* check the name with EVP_MD_is_a().
*/
EVP_MD_free(check_md);
if (!TEST_ptr(check_md = EVP_MD_fetch(testctx, mdname, testpropq)))
Extend the SM2 asym cipher test Ensure we test getting and setting ctx params Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/12913)
2020-09-18 18:57:24 +08:00
goto done;
TEST: Modify how the retrieved digest name for SM2 digestsign is checked Because of aliases, retrieved names won't always match one specific string. A safer way to check is to fetch the digest from the retrieved name and check it's the expected one with the help of EVP_MD_is_a(). Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14498)
2021-03-10 19:53:51 +08:00
if (!TEST_true(EVP_MD_is_a(check_md, mdnames[i]))) {
TEST_info("Fetched md %s isn't %s", mdname, mdnames[i]);
Extend the SM2 asym cipher test Ensure we test getting and setting ctx params Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/12913)
2020-09-18 18:57:24 +08:00
goto done;
TEST: Modify how the retrieved digest name for SM2 digestsign is checked Because of aliases, retrieved names won't always match one specific string. A safer way to check is to fetch the digest from the retrieved name and check it's the expected one with the help of EVP_MD_is_a(). Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14498)
2021-03-10 19:53:51 +08:00
}
Extend the SM2 asym cipher test Ensure we test getting and setting ctx params Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/12913)
2020-09-18 18:57:24 +08:00
if (!TEST_true(ptext_len == sizeof(kMsg)))
goto done;
if (!TEST_true(memcmp(plaintext, kMsg, sizeof(kMsg)) == 0))
goto done;
}
Move SM2 algos to SM2 specific PKEY method Use EVP_PKEY_set_alias_type to access Reviewed-by: Andy Polyakov <appro@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6443)
2018-06-19 03:51:56 +08:00
ret = 1;
done:
EVP_PKEY_CTX_free(pctx);
EVP_PKEY_CTX_free(kctx);
Support setting SM2 ID zero-length ID is allowed, but it's not allowed to skip the ID. Fixes: #6534 Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7113)
2018-09-04 17:21:10 +08:00
EVP_PKEY_CTX_free(sctx);
Move SM2 algos to SM2 specific PKEY method Use EVP_PKEY_set_alias_type to access Reviewed-by: Andy Polyakov <appro@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6443)
2018-06-19 03:51:56 +08:00
EVP_PKEY_CTX_free(cctx);
EVP_PKEY_free(pkey);
Extend the SM2 asym cipher test Ensure we test getting and setting ctx params Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/12913)
2020-09-18 18:57:24 +08:00
EVP_PKEY_free(pkeyparams);
Move SM2 algos to SM2 specific PKEY method Use EVP_PKEY_set_alias_type to access Reviewed-by: Andy Polyakov <appro@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6443)
2018-06-19 03:51:56 +08:00
EVP_MD_CTX_free(md_ctx);
EVP_MD_CTX_free(md_ctx_verify);
TEST: Modify how the retrieved digest name for SM2 digestsign is checked Because of aliases, retrieved names won't always match one specific string. A safer way to check is to fetch the digest from the retrieved name and check it's the expected one with the help of EVP_MD_is_a(). Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14498)
2021-03-10 19:53:51 +08:00
EVP_MD_free(check_md);
Move SM2 algos to SM2 specific PKEY method Use EVP_PKEY_set_alias_type to access Reviewed-by: Andy Polyakov <appro@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6443)
2018-06-19 03:51:56 +08:00
OPENSSL_free(sig);
return ret;
}
#endif
Multi-variant ML-KEM This introduces support for ML-KEM-512 and ML-KEM-1024 using the same underlying implementation parameterised by a few macros for the associated types and constants. KAT tests are added for ML-KEM 512 and 1024, to complement the previous tests for ML-KEM-768. MLKEM{512,768,1024} TLS "group" codepoints are updated to match the final IANA assigments and to make the additional KEMs known to the TLS layer. The pure-QC MLKEMs are not in the default list of supported groups, and need to be explicitly enabled by the application. Future work will introduce support for hybrids, and for more fine-grained policy of which keyshares a client should send by default, and when a server should request (HRR) a new mutually-supported group that was not sent. Tests for ML-KEM key exchange added to sslapitest to make sure that our TLS client MLKEM{512,768,1024} implementations interoperate with our TLS server, and that MLKEM* are not negotiated in TLS 1.2. Tests also added to excercise non-derandomised ML-KEM APIs, both directly (bypassing the provider layer), and through the generic EVP KEM API (exercising the provider). These make sure that RNG input is used correctly (KAT tests bypass the RNG by specifying seeds). The API interface to the provider takes an "const ML_KEM_VINFO" pointer, (obtained from ossl_ml_kem_get_vinfo()). This checks input and output buffer sizes before passing control to internal code that assumes correctly sized (for each variant) buffers. The original BoringSSL API was refactored to eliminate the opaque public/private key structure wrappers, since these structures are an internal detail between libcrypto and the provider, they are not part of the public (EVP) API. New "clangover" counter-measures added, refined with much appreciated input from David Benjamin (Chromium). The internal steps of "encrypt_cpa" were reordered to reduce the working-set size of the algorithm, now needs space for just two temporary "vectors" rather than three. The "decap" function now process the decrypted message in one call, rather than three separate calls to scalar_decode_1, scalar_decompress and scalar_add. Some loops were unrolled, improving performance of en/decapsulate (pre-expanded vectors and matrix) by around 5%. To handle, however unlikely, the SHA3 primitives not behaving like "pure" functions and failing, the implementation of `decap` was modifed: - To use the KDF to compute the Fujisaki-Okamoto (FO) failure secret first thing, and if that fails, bail out returning an error, a shared secret is still returned at random from the RNG, but it is OK for the caller to not use it. - If any of the subsequently used hash primitives fail, use the computed FO failure secret (OK, despite no longer constant-time) and return success (otherwise the RNG would replace the result). - We quite reasonably assume that chosen-ciphertext attacks (of the correct length) cannot cause hash functions to fail in a manner the depends on the private key content. Support for ML-KEM-512 required adding a centered binomial distribution helper function to deal with η_1 == 3 in just that variant. Some additional comments were added to highlight how the code relates to the ML-KEM specification in FIPS 203. Reviewed-by: Paul Dale <ppzgs1@gmail.com> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26172)
2024-11-30 22:20:58 +08:00
#ifndef OPENSSL_NO_ML_KEM
Improved import and export - On import, if a seed is provided, the keys are regenerated. - The seed is exported as a separate "seed" parameter, when available. The "ml-kem.retain_seed" parameter is also exported, when false. - The seed is optionally dropped after key generation. * When the "ml-kem.retain_seed" keygen parameter is set to zero. * When the "ml-kem.retain_seed" keygen parameter is not set to 1, and the "ml-kem.retain_seed" provider config property is set explictly false. - The exported private key parameter "priv" is always the FIPS 203 |dk|. - Private key decoding from PKCS#8 produces a transient "seed-only" form of the key, in which "retain_seed" is set to false when the "ml-kem.retain_seed" provider config property is set explictly false. The full key is generated during "load" and the seed is retained or not as specified. - Import honours the "ml-kem.retain_seed" parameter when specified, or otherwise honours the provider's "ml-kem.retain_seed" property. Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26512)
2025-01-18 00:28:51 +08:00
static const uint8_t ml_kem_seed[] = {
Encoders and Decoders for ML-KEM Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26341)
2025-01-09 11:35:03 +08:00
0x7c, 0x99, 0x35, 0xa0, 0xb0, 0x76, 0x94, 0xaa, 0x0c, 0x6d, 0x10, 0xe4,
0xdb, 0x6b, 0x1a, 0xdd, 0x2f, 0xd8, 0x1a, 0x25, 0xcc, 0xb1, 0x48, 0x03,
0x2d, 0xcd, 0x73, 0x99, 0x36, 0x73, 0x7f, 0x2d, 0x86, 0x26, 0xed, 0x79,
0xd4, 0x51, 0x14, 0x08, 0x00, 0xe0, 0x3b, 0x59, 0xb9, 0x56, 0xf8, 0x21,
0x0e, 0x55, 0x60, 0x67, 0x40, 0x7d, 0x13, 0xdc, 0x90, 0xfa, 0x9e, 0x8b,
0x87, 0x2b, 0xfb, 0x8f
};
Multi-variant ML-KEM This introduces support for ML-KEM-512 and ML-KEM-1024 using the same underlying implementation parameterised by a few macros for the associated types and constants. KAT tests are added for ML-KEM 512 and 1024, to complement the previous tests for ML-KEM-768. MLKEM{512,768,1024} TLS "group" codepoints are updated to match the final IANA assigments and to make the additional KEMs known to the TLS layer. The pure-QC MLKEMs are not in the default list of supported groups, and need to be explicitly enabled by the application. Future work will introduce support for hybrids, and for more fine-grained policy of which keyshares a client should send by default, and when a server should request (HRR) a new mutually-supported group that was not sent. Tests for ML-KEM key exchange added to sslapitest to make sure that our TLS client MLKEM{512,768,1024} implementations interoperate with our TLS server, and that MLKEM* are not negotiated in TLS 1.2. Tests also added to excercise non-derandomised ML-KEM APIs, both directly (bypassing the provider layer), and through the generic EVP KEM API (exercising the provider). These make sure that RNG input is used correctly (KAT tests bypass the RNG by specifying seeds). The API interface to the provider takes an "const ML_KEM_VINFO" pointer, (obtained from ossl_ml_kem_get_vinfo()). This checks input and output buffer sizes before passing control to internal code that assumes correctly sized (for each variant) buffers. The original BoringSSL API was refactored to eliminate the opaque public/private key structure wrappers, since these structures are an internal detail between libcrypto and the provider, they are not part of the public (EVP) API. New "clangover" counter-measures added, refined with much appreciated input from David Benjamin (Chromium). The internal steps of "encrypt_cpa" were reordered to reduce the working-set size of the algorithm, now needs space for just two temporary "vectors" rather than three. The "decap" function now process the decrypted message in one call, rather than three separate calls to scalar_decode_1, scalar_decompress and scalar_add. Some loops were unrolled, improving performance of en/decapsulate (pre-expanded vectors and matrix) by around 5%. To handle, however unlikely, the SHA3 primitives not behaving like "pure" functions and failing, the implementation of `decap` was modifed: - To use the KDF to compute the Fujisaki-Okamoto (FO) failure secret first thing, and if that fails, bail out returning an error, a shared secret is still returned at random from the RNG, but it is OK for the caller to not use it. - If any of the subsequently used hash primitives fail, use the computed FO failure secret (OK, despite no longer constant-time) and return success (otherwise the RNG would replace the result). - We quite reasonably assume that chosen-ciphertext attacks (of the correct length) cannot cause hash functions to fail in a manner the depends on the private key content. Support for ML-KEM-512 required adding a centered binomial distribution helper function to deal with η_1 == 3 in just that variant. Some additional comments were added to highlight how the code relates to the ML-KEM specification in FIPS 203. Reviewed-by: Paul Dale <ppzgs1@gmail.com> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26172)
2024-11-30 22:20:58 +08:00
static const uint8_t ml_kem_512_pubkey[] = {
0x40, 0x08, 0x65, 0xed, 0x10, 0xb6, 0x19, 0xaa, 0x58, 0x11, 0x13, 0x9b,
0xc0, 0x86, 0x82, 0x57, 0x82, 0xb2, 0xb7, 0x12, 0x4f, 0x75, 0x7c, 0x83,
0xae, 0x79, 0x44, 0x44, 0xbc, 0x78, 0xa4, 0x78, 0x96, 0xac, 0xf1, 0x26,
0x2c, 0x81, 0x35, 0x10, 0x77, 0x89, 0x3b, 0xfc, 0x56, 0xf9, 0x04, 0x49,
0xc2, 0xfa, 0x5f, 0x6e, 0x58, 0x6d, 0xd3, 0x7c, 0x0b, 0x9b, 0x58, 0x19,
0x92, 0x63, 0x8c, 0xb7, 0xe7, 0xbc, 0xbb, 0xb9, 0x9a, 0xfe, 0x47, 0x81,
0xd8, 0x0a, 0x50, 0xe6, 0x94, 0x63, 0xfb, 0xd9, 0x88, 0x72, 0x2c, 0x36,
0x35, 0x42, 0x3e, 0x27, 0x46, 0x6c, 0x71, 0xdc, 0xc6, 0x74, 0x52, 0x7c,
0xcd, 0x72, 0x89, 0x68, 0xcb, 0xcd, 0xc0, 0x0c, 0x5c, 0x90, 0x35, 0xbb,
0x0a, 0xf2, 0xc9, 0x92, 0x2c, 0x78, 0x81, 0xa4, 0x1d, 0xd2, 0x87, 0x52,
0x73, 0x92, 0x51, 0x31, 0x23, 0x0f, 0x6c, 0xa5, 0x9e, 0x91, 0x36, 0xb3,
0x9f, 0x95, 0x6c, 0x93, 0xb3, 0xb2, 0xd1, 0x4c, 0x64, 0x1b, 0x08, 0x9e,
0x07, 0xd0, 0xa8, 0x40, 0xc8, 0x93, 0xec, 0xd7, 0x6b, 0xbf, 0x92, 0xc8,
0x05, 0x45, 0x66, 0x68, 0xd0, 0x7c, 0x62, 0x14, 0x91, 0xc5, 0xc0, 0x54,
0x99, 0x1a, 0x65, 0x6f, 0x51, 0x16, 0x19, 0x55, 0x6e, 0xb9, 0x77, 0x82,
0xe2, 0x7a, 0x3c, 0x78, 0x51, 0x24, 0xc7, 0x0b, 0x0d, 0xab, 0xa6, 0xc6,
0x24, 0xd1, 0x8e, 0x0f, 0x97, 0x93, 0xf9, 0x6b, 0xa9, 0xe1, 0x59, 0x9b,
0x17, 0xb3, 0x0d, 0xcc, 0xc0, 0xb4, 0xf3, 0x76, 0x6a, 0x07, 0xb2, 0x3b,
0x25, 0x73, 0x09, 0xcd, 0x76, 0xab, 0xa0, 0x72, 0xc2, 0xb9, 0xc9, 0x74,
0x43, 0x94, 0xc6, 0xab, 0x9c, 0xb6, 0xc5, 0x4a, 0x97, 0xb5, 0xc5, 0x78,
0x61, 0xa5, 0x8d, 0xc0, 0xa0, 0x35, 0x19, 0x83, 0x2e, 0xe3, 0x2a, 0x07,
0x65, 0x4a, 0x07, 0x0c, 0x0c, 0x8c, 0x4e, 0x86, 0x48, 0xad, 0xdc, 0x35,
0x5f, 0x27, 0x4f, 0xc6, 0xb9, 0x2a, 0x08, 0x7b, 0x3f, 0x97, 0x51, 0x92,
0x3e, 0x44, 0x27, 0x4f, 0x85, 0x8c, 0x49, 0xca, 0xba, 0x72, 0xb6, 0x58,
0x51, 0xb3, 0xad, 0xc4, 0x89, 0x36, 0x95, 0x50, 0x97, 0xca, 0xd9, 0x55,
0x3f, 0x5a, 0x26, 0x3f, 0x18, 0x44, 0xb5, 0x2a, 0x02, 0x0f, 0xf7, 0xca,
0x89, 0xe8, 0x81, 0xa0, 0x1b, 0x95, 0xd9, 0x57, 0xa3, 0x15, 0x3c, 0x0a,
0x5e, 0x0a, 0x1c, 0xcd, 0x66, 0xb1, 0x82, 0x1a, 0x2b, 0x86, 0x32, 0x54,
0x6e, 0x24, 0xc7, 0xcb, 0xbc, 0x4c, 0xb0, 0x88, 0x08, 0xca, 0xc3, 0x7f,
0x7d, 0xa6, 0xb1, 0x6f, 0x8a, 0xce, 0xd0, 0x52, 0xcd, 0xb2, 0x56, 0x49,
0x48, 0xf1, 0xab, 0x0f, 0x76, 0x8a, 0x0d, 0x32, 0x86, 0xcc, 0xc7, 0xc3,
0x74, 0x9c, 0x63, 0xc7, 0x81, 0x53, 0x0f, 0xa1, 0xae, 0x67, 0x05, 0x42,
0x85, 0x50, 0x04, 0xa6, 0x45, 0xb5, 0x22, 0x88, 0x1e, 0xc1, 0x41, 0x2b,
0xda, 0xe3, 0x42, 0x08, 0x5a, 0x9d, 0xd5, 0xf8, 0x12, 0x6a, 0xf9, 0x6b,
0xbd, 0xb0, 0xc1, 0xaf, 0x69, 0xa1, 0x55, 0x62, 0xcb, 0x2a, 0x15, 0x5a,
0x10, 0x03, 0x09, 0xd1, 0xb6, 0x41, 0xd0, 0x8b, 0x2d, 0x4e, 0xd1, 0x7b,
0xfb, 0xf0, 0xbc, 0x04, 0x26, 0x5f, 0x9b, 0x10, 0xc1, 0x08, 0xf8, 0x50,
0x30, 0x95, 0x04, 0xd7, 0x72, 0x81, 0x1b, 0xba, 0x8e, 0x2b, 0xe1, 0x62,
0x49, 0xaa, 0x73, 0x7d, 0x87, 0x9f, 0xc7, 0xfb, 0x25, 0x5e, 0xe7, 0xa6,
0xa0, 0xa7, 0x53, 0xbd, 0x93, 0x74, 0x1c, 0x61, 0x65, 0x8e, 0xc0, 0x74,
0xf6, 0xe0, 0x02, 0xb0, 0x19, 0x34, 0x57, 0x69, 0x11, 0x3c, 0xc0, 0x13,
0xff, 0x74, 0x94, 0xba, 0x83, 0x78, 0xb1, 0x1a, 0x17, 0x22, 0x60, 0xaa,
0xa5, 0x34, 0x21, 0xbd, 0xe0, 0x3a, 0x35, 0x58, 0x9d, 0x57, 0xe3, 0x22,
0xfe, 0xfa, 0x41, 0x00, 0xa4, 0x74, 0x39, 0x26, 0xab, 0x7d, 0x62, 0x25,
0x8b, 0x87, 0xb3, 0x1c, 0xcb, 0xb5, 0xe6, 0xb8, 0x9c, 0xb1, 0x0b, 0x27,
0x1a, 0xa0, 0x5d, 0x99, 0x4b, 0xb5, 0x70, 0x8b, 0x23, 0xab, 0x32, 0x7e,
0xcb, 0x93, 0xc0, 0xf3, 0x15, 0x68, 0x69, 0xf0, 0x88, 0x3d, 0xa2, 0x06,
0x4f, 0x79, 0x5e, 0x0e, 0x2a, 0xb7, 0xd3, 0xc6, 0x4d, 0x61, 0xd2, 0x30,
0x3f, 0xc3, 0xa2, 0x9e, 0x16, 0x19, 0x92, 0x3c, 0xa8, 0x01, 0xe5, 0x9f,
0xd7, 0x52, 0xca, 0x6e, 0x76, 0x49, 0xd3, 0x03, 0xc9, 0xd2, 0x07, 0x88,
0xe1, 0x21, 0x46, 0x51, 0xb0, 0x69, 0x95, 0xeb, 0x26, 0x0c, 0x92, 0x9a,
0x13, 0x44, 0xa8, 0x49, 0xb2, 0x5c, 0xa0, 0xa0, 0x1f, 0x1e, 0xb5, 0x29,
0x13, 0x68, 0x6b, 0xba, 0x61, 0x9e, 0x23, 0x71, 0x44, 0x64, 0x03, 0x1a,
0x78, 0x43, 0x92, 0x87, 0xfc, 0xa7, 0x8f, 0x4c, 0x04, 0x76, 0x22, 0x3e,
0xea, 0x61, 0xb7, 0xf2, 0x5a, 0x7c, 0xe4, 0x2c, 0xca, 0x90, 0x1b, 0x2a,
0xea, 0x12, 0x98, 0x17, 0x89, 0x4b, 0xa3, 0x47, 0x08, 0x23, 0x85, 0x4f,
0x3e, 0x5b, 0x28, 0xd8, 0x6b, 0xa9, 0x79, 0xe5, 0x46, 0x71, 0x86, 0x2d,
0x90, 0x47, 0x0b, 0x1e, 0x78, 0x38, 0x97, 0x2a, 0x81, 0xa4, 0x81, 0x07,
0xd6, 0xac, 0x06, 0x11, 0x40, 0x6b, 0x21, 0xfb, 0xcc, 0xe1, 0xdb, 0x77,
0x02, 0xea, 0x9d, 0xd6, 0xba, 0x6e, 0x40, 0x52, 0x7b, 0x9d, 0xc6, 0x63,
0xf3, 0xc9, 0x3b, 0xad, 0x05, 0x6d, 0xc2, 0x85, 0x11, 0xf6, 0x6c, 0x3e,
0x0b, 0x92, 0x8d, 0xb8, 0x87, 0x9d, 0x22, 0xc5, 0x92, 0x68, 0x5c, 0xc7,
0x75, 0xa6, 0xcd, 0x57, 0x4a, 0xc3, 0xbc, 0xe3, 0xb2, 0x75, 0x91, 0xc8,
0x21, 0x92, 0x90, 0x76, 0x35, 0x8a, 0x22, 0x00, 0xb3, 0x77, 0x36, 0x5f,
0x7e, 0xfb, 0x9e, 0x40, 0xc3, 0xbf, 0x0f, 0xf0, 0x43, 0x29, 0x86, 0xae,
0x4b, 0xc1, 0xa2, 0x42, 0xce, 0x99, 0x21, 0xaa, 0x9e, 0x22, 0x44, 0x88,
0x19, 0x58, 0x5d, 0xea, 0x30, 0x8e, 0xb0, 0x39
};
static const uint8_t ml_kem_768_pubkey[] = {
0xa8, 0xe6, 0x51, 0xa1, 0xe6, 0x85, 0xf2, 0x24, 0x78, 0xa8, 0x95, 0x4f,
0x00, 0x7b, 0xc7, 0x71, 0x1b, 0x93, 0x07, 0x72, 0xc7, 0x8f, 0x09, 0x2e,
0x82, 0x87, 0x8e, 0x3e, 0x93, 0x7f, 0x36, 0x79, 0x67, 0x53, 0x29, 0x13,
0xa8, 0xd5, 0x3d, 0xfd, 0xf4, 0xbf, 0xb1, 0xf8, 0x84, 0x67, 0x46, 0x59,
0x67, 0x05, 0xcf, 0x34, 0x51, 0x42, 0xb9, 0x72, 0xa3, 0xf1, 0x63, 0x25,
0xc4, 0x0c, 0x29, 0x52, 0xa3, 0x7b, 0x25, 0x89, 0x7e, 0x5e, 0xf3, 0x5f,
0xba, 0xeb, 0x73, 0xa4, 0xac, 0xbe, 0xb6, 0xa0, 0xb8, 0x99, 0x42, 0xce,
0xb1, 0x95, 0x53, 0x1c, 0xfc, 0x0a, 0x07, 0x99, 0x39, 0x54, 0x48, 0x3e,
0x6c, 0xbc, 0x87, 0xc0, 0x6a, 0xa7, 0x4f, 0xf0, 0xca, 0xc5, 0x20, 0x7e,
0x53, 0x5b, 0x26, 0x0a, 0xa9, 0x8d, 0x11, 0x98, 0xc0, 0x7d, 0xa6, 0x05,
0xc4, 0xd1, 0x10, 0x20, 0xf6, 0xc9, 0xf7, 0xbb, 0x68, 0xbb, 0x34, 0x56,
0xc7, 0x3a, 0x01, 0xb7, 0x10, 0xbc, 0x99, 0xd1, 0x77, 0x39, 0xa5, 0x17,
0x16, 0xaa, 0x01, 0x66, 0x0c, 0x8b, 0x62, 0x8b, 0x2f, 0x56, 0x02, 0xba,
0x65, 0xf0, 0x7e, 0xa9, 0x93, 0x33, 0x6e, 0x89, 0x6e, 0x83, 0xf2, 0xc5,
0x73, 0x1b, 0xbf, 0x03, 0x46, 0x0c, 0x5b, 0x6c, 0x8a, 0xfe, 0xcb, 0x74,
0x8e, 0xe3, 0x91, 0xe9, 0x89, 0x34, 0xa2, 0xc5, 0x7d, 0x4d, 0x06, 0x9f,
0x50, 0xd8, 0x8b, 0x30, 0xd6, 0x96, 0x6f, 0x38, 0xc3, 0x7b, 0xc6, 0x49,
0xb8, 0x26, 0x34, 0xce, 0x77, 0x22, 0x64, 0x5c, 0xcd, 0x62, 0x50, 0x63,
0x36, 0x46, 0x46, 0xd6, 0xd6, 0x99, 0xdb, 0x57, 0xb4, 0x5e, 0xb6, 0x74,
0x65, 0xe1, 0x6d, 0xe4, 0xd4, 0x06, 0xa8, 0x18, 0xb9, 0xea, 0xe1, 0xca,
0x91, 0x6a, 0x25, 0x94, 0x48, 0x97, 0x08, 0xa4, 0x3c, 0xea, 0x88, 0xb0,
0x2a, 0x4c, 0x03, 0xd0, 0x9b, 0x44, 0x81, 0x5c, 0x97, 0x10, 0x1c, 0xaf,
0x50, 0x48, 0xbb, 0xcb, 0x24, 0x7a, 0xe2, 0x36, 0x6c, 0xdc, 0x25, 0x4b,
0xa2, 0x21, 0x29, 0xf4, 0x5b, 0x3b, 0x0e, 0xb3, 0x99, 0xca, 0x91, 0xa3,
0x03, 0x40, 0x28, 0x30, 0xec, 0x01, 0xdb, 0x7b, 0x2c, 0xa4, 0x80, 0xcf,
0x35, 0x04, 0x09, 0xb2, 0x16, 0x09, 0x4b, 0x7b, 0x0c, 0x3a, 0xe3, 0x3c,
0xe1, 0x0a, 0x91, 0x24, 0xe8, 0x96, 0x51, 0xab, 0x90, 0x1e, 0xa2, 0x53,
0xc8, 0x41, 0x5b, 0xd7, 0x82, 0x5f, 0x02, 0xbb, 0x22, 0x93, 0x69, 0xaf,
0x97, 0x20, 0x28, 0xf2, 0x28, 0x75, 0xea, 0x55, 0xaf, 0x16, 0xd3, 0xbc,
0x69, 0xf7, 0x0c, 0x2e, 0xe8, 0xb7, 0x5f, 0x28, 0xb4, 0x7d, 0xd3, 0x91,
0xf9, 0x89, 0xad, 0xe3, 0x14, 0x72, 0x9c, 0x33, 0x1f, 0xa0, 0x4c, 0x19,
0x17, 0xb2, 0x78, 0xc3, 0xeb, 0x60, 0x28, 0x68, 0x51, 0x28, 0x21, 0xad,
0xc8, 0x25, 0xc6, 0x45, 0x77, 0xce, 0x1e, 0x63, 0xb1, 0xd9, 0x64, 0x4a,
0x61, 0x29, 0x48, 0xa3, 0x48, 0x3c, 0x7f, 0x1b, 0x9a, 0x25, 0x80, 0x00,
0xe3, 0x01, 0x96, 0x94, 0x4a, 0x40, 0x36, 0x27, 0x60, 0x9c, 0x76, 0xc7,
0xea, 0x6b, 0x5d, 0xe0, 0x17, 0x64, 0xd2, 0x43, 0x79, 0x11, 0x7b, 0x9e,
0xa2, 0x98, 0x48, 0xdc, 0x55, 0x5c, 0x45, 0x4b, 0xce, 0xae, 0x1b, 0xa5,
0xcc, 0x72, 0xc7, 0x4a, 0xb9, 0x6b, 0x9c, 0x91, 0xb9, 0x10, 0xd2, 0x6b,
0x88, 0xb2, 0x56, 0x39, 0xd4, 0x77, 0x8a, 0xe2, 0x6c, 0x7c, 0x61, 0x51,
0xa1, 0x9c, 0x6c, 0xd7, 0x93, 0x84, 0x54, 0x37, 0x24, 0x65, 0xe4, 0xc5,
0xec, 0x29, 0x24, 0x5a, 0xcb, 0x3d, 0xb5, 0x37, 0x9d, 0xe3, 0xda, 0xbf,
0xa6, 0x29, 0xa7, 0xc0, 0x4a, 0x83, 0x53, 0xa8, 0x53, 0x0c, 0x95, 0xac,
0xb7, 0x32, 0xbb, 0x4b, 0xb8, 0x19, 0x32, 0xbb, 0x2c, 0xa7, 0xa8, 0x48,
0xcd, 0x36, 0x68, 0x01, 0x44, 0x4a, 0xbe, 0x23, 0xc8, 0x3b, 0x36, 0x6a,
0x87, 0xd6, 0xa3, 0xcf, 0x36, 0x09, 0x24, 0xc0, 0x02, 0xba, 0xe9, 0x0a,
0xf6, 0x5c, 0x48, 0x06, 0x0b, 0x37, 0x52, 0xf2, 0xba, 0xdf, 0x1a, 0xb2,
0x72, 0x20, 0x72, 0x55, 0x4a, 0x50, 0x59, 0x75, 0x35, 0x94, 0xe6, 0xa7,
0x02, 0x76, 0x1f, 0xc9, 0x76, 0x84, 0xc8, 0xc4, 0xa7, 0x54, 0x0a, 0x6b,
0x07, 0xfb, 0xc9, 0xde, 0x87, 0xc9, 0x74, 0xaa, 0x88, 0x09, 0xd9, 0x28,
0xc7, 0xf4, 0xcb, 0xbf, 0x80, 0x45, 0xae, 0xa5, 0xbc, 0x66, 0x78, 0x25,
0xfd, 0x05, 0xa5, 0x21, 0xf1, 0xa4, 0xbf, 0x53, 0x92, 0x10, 0xc7, 0x11,
0x3b, 0xc3, 0x7b, 0x3e, 0x58, 0xb0, 0xcb, 0xfc, 0x53, 0xc8, 0x41, 0xcb,
0xb0, 0x37, 0x1d, 0xe2, 0xe5, 0x11, 0xb9, 0x89, 0xcb, 0x7c, 0x70, 0xc0,
0x23, 0x36, 0x6d, 0x78, 0xf9, 0xc3, 0x7e, 0xf0, 0x47, 0xf8, 0x72, 0x0b,
0xe1, 0xc7, 0x59, 0xa8, 0xd9, 0x6b, 0x93, 0xf6, 0x5a, 0x94, 0x11, 0x4f,
0xfa, 0xf6, 0x0d, 0x9a, 0x81, 0x79, 0x5e, 0x99, 0x5c, 0x71, 0x15, 0x2a,
0x46, 0x91, 0xa5, 0xa6, 0x02, 0xa9, 0xe1, 0xf3, 0x59, 0x9e, 0x37, 0xc7,
0x68, 0xc7, 0xbc, 0x10, 0x89, 0x94, 0xc0, 0x66, 0x9f, 0x3a, 0xdc, 0x95,
0x7d, 0x46, 0xb4, 0xb6, 0x25, 0x69, 0x68, 0xe2, 0x90, 0xd7, 0x89, 0x2e,
0xa8, 0x54, 0x64, 0xee, 0x7a, 0x75, 0x0f, 0x39, 0xc5, 0xe3, 0x15, 0x2c,
0x2d, 0xfc, 0x56, 0xd8, 0xb0, 0xc9, 0x24, 0xba, 0x8a, 0x95, 0x9a, 0x68,
0x09, 0x65, 0x47, 0xf6, 0x64, 0x23, 0xc8, 0x38, 0x98, 0x2a, 0x57, 0x94,
0xb9, 0xe1, 0x53, 0x37, 0x71, 0x33, 0x1a, 0x9a, 0x65, 0x6c, 0x28, 0x82,
0x8b, 0xeb, 0x91, 0x26, 0xa6, 0x0e, 0x95, 0xe8, 0xc5, 0xd9, 0x06, 0x83,
0x2c, 0x77, 0x10, 0x70, 0x55, 0x76, 0xb1, 0xfb, 0x95, 0x07, 0x26, 0x9d,
0xda, 0xf8, 0xc9, 0x5c, 0xe9, 0x71, 0x9b, 0x2c, 0xa8, 0xdd, 0x11, 0x2b,
0xe1, 0x0b, 0xcc, 0x9f, 0x4a, 0x37, 0xbd, 0x1b, 0x1e, 0xee, 0xb3, 0x3e,
0xcd, 0xa7, 0x6a, 0xe9, 0xf6, 0x9a, 0x5d, 0x4b, 0x29, 0x23, 0xa8, 0x69,
0x57, 0x67, 0x1d, 0x61, 0x93, 0x35, 0xbe, 0x1c, 0x4c, 0x2c, 0x77, 0xce,
0x87, 0xc4, 0x1f, 0x98, 0xa8, 0xcc, 0x46, 0x64, 0x60, 0xfa, 0x30, 0x0a,
0xaf, 0x5b, 0x30, 0x1f, 0x0a, 0x1d, 0x09, 0xc8, 0x8e, 0x65, 0xda, 0x4d,
0x8e, 0xe6, 0x4f, 0x68, 0xc0, 0x21, 0x89, 0xbb, 0xb3, 0x58, 0x4b, 0xaf,
0xf7, 0x16, 0xc8, 0x5d, 0xb6, 0x54, 0x04, 0x8a, 0x00, 0x43, 0x33, 0x48,
0x93, 0x93, 0xa0, 0x74, 0x27, 0xcd, 0x3e, 0x21, 0x7e, 0x6a, 0x34, 0x5f,
0x6c, 0x2c, 0x2b, 0x13, 0xc2, 0x7b, 0x33, 0x72, 0x71, 0xc0, 0xb2, 0x7b,
0x2d, 0xba, 0xa0, 0x0d, 0x23, 0x76, 0x00, 0xb5, 0xb5, 0x94, 0xe8, 0xcf,
0x2d, 0xd6, 0x25, 0xea, 0x76, 0xcf, 0x0e, 0xd8, 0x99, 0x12, 0x2c, 0x97,
0x96, 0xb4, 0xb0, 0x18, 0x70, 0x04, 0x25, 0x80, 0x49, 0xa4, 0x77, 0xcd,
0x11, 0xd6, 0x8c, 0x49, 0xb9, 0xa0, 0xe7, 0xb0, 0x0b, 0xce, 0x8c, 0xac,
0x78, 0x64, 0xcb, 0xb3, 0x75, 0x14, 0x00, 0x84, 0x74, 0x4c, 0x93, 0x06,
0x26, 0x94, 0xca, 0x79, 0x5c, 0x4f, 0x40, 0xe7, 0xac, 0xc9, 0xc5, 0xa1,
0x88, 0x40, 0x72, 0xd8, 0xc3, 0x8d, 0xaf, 0xb5, 0x01, 0xee, 0x41, 0x84,
0xdd, 0x5a, 0x81, 0x9e, 0xc2, 0x4e, 0xc1, 0x65, 0x12, 0x61, 0xf9, 0x62,
0xb1, 0x7a, 0x72, 0x15, 0xaa, 0x4a, 0x74, 0x8c, 0x15, 0x83, 0x6c, 0x38,
0x91, 0x37, 0x67, 0x82, 0x04, 0x83, 0x8d, 0x71, 0x95, 0xa8, 0x5b, 0x4f,
0x98, 0xa1, 0xb5, 0x74, 0xc4, 0xcd, 0x79, 0x09, 0xcd, 0x1f, 0x83, 0x3e,
0xff, 0xd1, 0x48, 0x55, 0x43, 0x22, 0x9d, 0x37, 0x48, 0xd9, 0xb5, 0xcd,
0x6c, 0x17, 0xb9, 0xb3, 0xb8, 0x4a, 0xef, 0x8b, 0xce, 0x13, 0xe6, 0x83,
0x73, 0x36, 0x59, 0xc7, 0x95, 0x42, 0xd6, 0x15, 0x78, 0x2a, 0x71, 0xcd,
0xee, 0xe7, 0x92, 0xba, 0xb5, 0x1b, 0xdc, 0x4b, 0xbf, 0xe8, 0x30, 0x8e,
0x66, 0x31, 0x44, 0xed, 0xe8, 0x49, 0x18, 0x30, 0xad, 0x98, 0xb4, 0x63,
0x4f, 0x64, 0xab, 0xa8, 0xb9, 0xc0, 0x42, 0x27, 0x26, 0x53, 0x92, 0x0f,
0x38, 0x0c, 0x1a, 0x17, 0xca, 0x87, 0xce, 0xd7, 0xaa, 0xc4, 0x1c, 0x82,
0x88, 0x87, 0x93, 0x18, 0x1a, 0x6f, 0x76, 0xe1, 0x97, 0xb7, 0xb9, 0x0e,
0xf9, 0x09, 0x43, 0xbb, 0x38, 0x44, 0x91, 0x29, 0x11, 0xd8, 0x55, 0x1e,
0x54, 0x66, 0xc5, 0x76, 0x7a, 0xb0, 0xbc, 0x61, 0xa1, 0xa3, 0xf7, 0x36,
0x16, 0x2e, 0xc0, 0x98, 0xa9, 0x00, 0xb1, 0x2d, 0xd8, 0xfa, 0xbb, 0xfb,
0x3f, 0xe8, 0xcb, 0x1d, 0xc4, 0xe8, 0x31, 0x5f, 0x2a, 0xf0, 0xd3, 0x2f,
0x00, 0x17, 0xae, 0x13, 0x6e, 0x19, 0xf0, 0x28
};
static const uint8_t ml_kem_1024_pubkey[] = {
0x53, 0x79, 0x11, 0x95, 0x7c, 0x12, 0x51, 0x48, 0xa8, 0x7f, 0x41, 0x58,
0x9c, 0xb2, 0x22, 0xd0, 0xd1, 0x92, 0x29, 0xe2, 0xcb, 0x55, 0xe1, 0xa0,
0x44, 0x79, 0x1e, 0x7c, 0xa6, 0x11, 0x92, 0xa4, 0x64, 0x60, 0xc3, 0x18,
0x3d, 0x2b, 0xcd, 0x6d, 0xe0, 0x8a, 0x5e, 0x76, 0x51, 0x60, 0x3a, 0xcc,
0x34, 0x9c, 0xa1, 0x6c, 0xba, 0x18, 0xab, 0xb2, 0x3a, 0x3e, 0x8c, 0x33,
0x0d, 0x74, 0x21, 0x59, 0x8a, 0x62, 0x78, 0xec, 0x7e, 0xbf, 0xab, 0xca,
0x0e, 0xf4, 0x88, 0xb2, 0x29, 0x05, 0x54, 0x75, 0x34, 0x99, 0xc0, 0x45,
0x2e, 0x45, 0x38, 0x15, 0x30, 0x99, 0x55, 0xb8, 0x15, 0x0f, 0xa1, 0xa1,
0xe3, 0x93, 0x38, 0x6d, 0xc1, 0x2f, 0xdb, 0x27, 0xb3, 0x8c, 0x67, 0x45,
0xf2, 0x94, 0x40, 0x16, 0xec, 0x45, 0x7f, 0x39, 0xb1, 0x8d, 0x60, 0x4a,
0x07, 0xa1, 0xab, 0xe0, 0x7b, 0xc8, 0x44, 0x05, 0x0f, 0xfa, 0x8a, 0x06,
0xfa, 0x15, 0x4a, 0x49, 0xd8, 0x8f, 0xac, 0x77, 0x54, 0x52, 0xd6, 0xa7,
0xc0, 0xe5, 0x89, 0xbf, 0xb5, 0xc3, 0x70, 0xc2, 0xc4, 0xb6, 0x20, 0x1d,
0xda, 0x80, 0xc9, 0xab, 0x20, 0x76, 0xec, 0xc0, 0x8b, 0x44, 0x52, 0x2f,
0xda, 0x33, 0x26, 0xf0, 0x33, 0x80, 0x6d, 0xd2, 0x69, 0x3f, 0x31, 0x97,
0x39, 0xf4, 0x0c, 0x4f, 0x42, 0xb2, 0x4a, 0xca, 0x70, 0x98, 0xfb, 0x8f,
0xf5, 0xf9, 0xac, 0x20, 0x29, 0x2d, 0x02, 0xb5, 0x6a, 0xc7, 0x46, 0x80,
0x1a, 0xcc, 0xcc, 0x84, 0x86, 0x3d, 0xee, 0x32, 0x87, 0x84, 0x97, 0xb6,
0x94, 0x38, 0xbf, 0x99, 0x17, 0x76, 0x28, 0x66, 0x50, 0x48, 0x2c, 0x8d,
0x9d, 0x95, 0x87, 0xbc, 0x6a, 0x55, 0xb8, 0x5c, 0x4d, 0x7f, 0xa7, 0x4d,
0x02, 0x65, 0x6b, 0x42, 0x1c, 0x9e, 0x23, 0xe0, 0x3a, 0x48, 0xd4, 0xb7,
0x44, 0x25, 0xc2, 0x6e, 0x4a, 0x20, 0xdd, 0x95, 0x62, 0xa4, 0xda, 0x07,
0x93, 0xf3, 0xa3, 0x52, 0xcc, 0xc0, 0xf1, 0x82, 0x17, 0xd8, 0x68, 0xc7,
0xf5, 0x00, 0x2a, 0xbe, 0x76, 0x8b, 0x1f, 0xc7, 0x3f, 0x05, 0x74, 0x4e,
0x7c, 0xc2, 0x8f, 0x10, 0x34, 0x40, 0x62, 0xc1, 0x0e, 0x08, 0xec, 0xcc,
0xed, 0x3c, 0x1f, 0x7d, 0x39, 0x2c, 0x01, 0xd9, 0x79, 0xdd, 0x71, 0x8d,
0x83, 0x98, 0x37, 0x46, 0x65, 0xa1, 0x6a, 0x98, 0x70, 0x58, 0x5c, 0x39,
0xd5, 0x58, 0x9a, 0x50, 0xe1, 0x33, 0x38, 0x9c, 0x9b, 0x9a, 0x27, 0x6c,
0x02, 0x42, 0x60, 0xd9, 0xfc, 0x77, 0x11, 0xc8, 0x1b, 0x63, 0x37, 0xb5,
0x7d, 0xa3, 0xc3, 0x76, 0xd0, 0xcd, 0x74, 0xe1, 0x4c, 0x73, 0x72, 0x7b,
0x27, 0x66, 0x56, 0xb9, 0xd8, 0xa4, 0xeb, 0x71, 0x89, 0x6f, 0xf5, 0x89,
0xd4, 0xb8, 0x93, 0xe7, 0x11, 0x0f, 0x3b, 0xb9, 0x48, 0xec, 0xe2, 0x91,
0xdd, 0x86, 0xc0, 0xb7, 0x46, 0x8a, 0x67, 0x8c, 0x74, 0x69, 0x80, 0xc1,
0x2a, 0xa6, 0xb9, 0x5e, 0x2b, 0x0c, 0xbe, 0x43, 0x31, 0xbb, 0x24, 0xa3,
0x3a, 0x27, 0x01, 0x53, 0xaa, 0x47, 0x2c, 0x47, 0x31, 0x23, 0x82, 0xca,
0x36, 0x5c, 0x5f, 0x35, 0x25, 0x9d, 0x02, 0x57, 0x46, 0xfc, 0x65, 0x95,
0xfe, 0x63, 0x6c, 0x76, 0x75, 0x10, 0xa6, 0x9c, 0x1e, 0x8a, 0x17, 0x6b,
0x79, 0x49, 0x95, 0x8f, 0x26, 0x97, 0x39, 0x94, 0x97, 0xa2, 0xfc, 0x73,
0x64, 0xa1, 0x2c, 0x81, 0x98, 0x29, 0x52, 0x39, 0xc8, 0x26, 0xcb, 0x50,
0x82, 0x08, 0x60, 0x77, 0x28, 0x2e, 0xd6, 0x28, 0x65, 0x1f, 0xc0, 0x4c,
0x63, 0x9b, 0x43, 0x85, 0x22, 0xa9, 0xde, 0x30, 0x9b, 0x14, 0xb0, 0x86,
0xd6, 0xe9, 0x23, 0xc5, 0x51, 0x62, 0x3b, 0xd7, 0x2a, 0x73, 0x3c, 0xb0,
0xda, 0xbc, 0x54, 0xa9, 0x41, 0x6a, 0x99, 0xe7, 0x2c, 0x9f, 0xda, 0x1c,
0xb3, 0xfb, 0x9b, 0xa0, 0x6b, 0x8a, 0xdb, 0x24, 0x22, 0xd6, 0x8c, 0xad,
0xc5, 0x53, 0xc9, 0x82, 0x02, 0xa1, 0x76, 0x56, 0x47, 0x8a, 0xc0, 0x44,
0xef, 0x34, 0x56, 0x37, 0x8a, 0xbc, 0xe9, 0x99, 0x1e, 0x01, 0x41, 0xba,
0x79, 0x09, 0x4f, 0xa8, 0xf7, 0x7a, 0x30, 0x08, 0x05, 0xd2, 0xd3, 0x2f,
0xfc, 0x62, 0xbf, 0x0c, 0xa4, 0x55, 0x4c, 0x33, 0x0c, 0x2b, 0xb7, 0x04,
0x2d, 0xb3, 0x51, 0x02, 0xf6, 0x8b, 0x1a, 0x00, 0x62, 0x58, 0x38, 0x65,
0x38, 0x1c, 0x74, 0xdd, 0x91, 0x3a, 0xf7, 0x0b, 0x26, 0xcf, 0x09, 0x23,
0xd0, 0xc4, 0xcb, 0x97, 0x16, 0x92, 0x22, 0x25, 0x52, 0xa8, 0xf4, 0xb7,
0x88, 0xb4, 0xaf, 0xd1, 0x34, 0x1a, 0x9d, 0xf4, 0x15, 0xcf, 0x20, 0x39,
0x00, 0xf5, 0xcc, 0xf7, 0xf6, 0x59, 0x88, 0x94, 0x9a, 0x75, 0x58, 0x0d,
0x04, 0x96, 0x39, 0x85, 0x31, 0x00, 0x85, 0x4b, 0x21, 0xf4, 0x01, 0x80,
0x03, 0x50, 0x2b, 0xb1, 0xba, 0x95, 0xf5, 0x56, 0xa5, 0xd6, 0x7c, 0x7e,
0xb5, 0x24, 0x10, 0xeb, 0xa2, 0x88, 0xa6, 0xd0, 0x63, 0x5c, 0xa8, 0xa4,
0xf6, 0xd6, 0x96, 0xd0, 0xa0, 0x20, 0xc8, 0x26, 0x93, 0x8d, 0x34, 0x94,
0x3c, 0x38, 0x08, 0xc7, 0x9c, 0xc0, 0x07, 0x76, 0x85, 0x33, 0x21, 0x6b,
0xc1, 0xb2, 0x9d, 0xa6, 0xc8, 0x12, 0xef, 0xf3, 0x34, 0x0b, 0xaa, 0x8d,
0x2e, 0x65, 0x34, 0x4f, 0x09, 0xbd, 0x47, 0x89, 0x4f, 0x5a, 0x3a, 0x41,
0x18, 0x71, 0x5b, 0x3c, 0x50, 0x20, 0x67, 0x93, 0x27, 0xf9, 0x18, 0x9f,
0x7e, 0x10, 0x85, 0x6b, 0x23, 0x8b, 0xb9, 0xb0, 0xab, 0x4c, 0xa8, 0x5a,
0xbf, 0x4b, 0x21, 0xf5, 0xc7, 0x6b, 0xcc, 0xd7, 0x18, 0x50, 0xb2, 0x2e,
0x04, 0x59, 0x28, 0x27, 0x6a, 0x0f, 0x2e, 0x95, 0x1d, 0xb0, 0x70, 0x7c,
0x6a, 0x11, 0x6d, 0xc1, 0x91, 0x13, 0xfa, 0x76, 0x2d, 0xc5, 0xf2, 0x0b,
0xd5, 0xd2, 0xab, 0x5b, 0xe7, 0x17, 0x44, 0xdc, 0x9c, 0xbd, 0xb5, 0x1e,
0xa7, 0x57, 0x96, 0x3a, 0xac, 0x56, 0xa9, 0x0a, 0x0d, 0x80, 0x23, 0xbe,
0xd1, 0xf5, 0xca, 0xe8, 0xa6, 0x4d, 0xa0, 0x47, 0x27, 0x9b, 0x35, 0x3a,
0x09, 0x6a, 0x83, 0x5b, 0x0b, 0x2b, 0x02, 0x3b, 0x6a, 0xa0, 0x48, 0x98,
0x92, 0x33, 0x07, 0x9a, 0xeb, 0x46, 0x7e, 0x52, 0x2f, 0xa2, 0x7a, 0x58,
0x22, 0x92, 0x1e, 0x5c, 0x55, 0x1b, 0x4f, 0x53, 0x75, 0x36, 0xe4, 0x6f,
0x3a, 0x6a, 0x97, 0xe7, 0x2c, 0x3b, 0x06, 0x31, 0x04, 0xe0, 0x9a, 0x04,
0x05, 0x98, 0x94, 0x0d, 0x87, 0x2f, 0x6d, 0x87, 0x1f, 0x5e, 0xf9, 0xb4,
0x35, 0x50, 0x73, 0xb5, 0x47, 0x69, 0xe4, 0x54, 0x54, 0xe6, 0xa0, 0x81,
0x95, 0x99, 0x40, 0x86, 0x21, 0xab, 0x44, 0x13, 0xb3, 0x55, 0x07, 0xb0,
0xdf, 0x57, 0x8c, 0xe2, 0xd5, 0x11, 0xd5, 0x20, 0x58, 0xd5, 0x74, 0x9d,
0xf3, 0x8b, 0x29, 0xd6, 0xcc, 0x58, 0x87, 0x0c, 0xaf, 0x92, 0xf6, 0x9a,
0x75, 0x16, 0x14, 0x06, 0xe7, 0x1c, 0x5f, 0xf9, 0x24, 0x51, 0xa7, 0x75,
0x22, 0xb8, 0xb2, 0x96, 0x7a, 0x2d, 0x58, 0xa4, 0x9a, 0x81, 0x66, 0x1a,
0xa6, 0x5a, 0xc0, 0x9b, 0x08, 0xc9, 0xfe, 0x45, 0xab, 0xc3, 0x85, 0x1f,
0x99, 0xc7, 0x30, 0xc4, 0x50, 0x03, 0xac, 0xa2, 0xbf, 0x0f, 0x84, 0x24,
0xa1, 0x9b, 0x74, 0x08, 0xa5, 0x37, 0xd5, 0x41, 0xc1, 0x6f, 0x56, 0x82,
0xbf, 0xe3, 0xa7, 0xfa, 0xea, 0x56, 0x4f, 0x12, 0x98, 0x61, 0x1a, 0x7f,
0x5f, 0x60, 0x92, 0x2b, 0xa1, 0x9d, 0xe7, 0x3b, 0x19, 0x17, 0xf1, 0x85,
0x32, 0x73, 0x55, 0x51, 0x99, 0xa6, 0x49, 0x31, 0x8b, 0x50, 0x77, 0x33,
0x45, 0xc9, 0x97, 0x46, 0x08, 0x56, 0x97, 0x2a, 0xcb, 0x43, 0xfc, 0x81,
0xab, 0x63, 0x21, 0xb1, 0xc3, 0x3c, 0x2b, 0xb5, 0x09, 0x8b, 0xd4, 0x89,
0xd6, 0x96, 0xa0, 0xf7, 0x06, 0x79, 0xc1, 0x21, 0x38, 0x73, 0xd0, 0x8b,
0xda, 0xd4, 0x28, 0x44, 0x92, 0x72, 0x16, 0x04, 0x72, 0x05, 0x63, 0x32,
0x12, 0x31, 0x0e, 0xe9, 0xa0, 0x6c, 0xb1, 0x00, 0x16, 0xc8, 0x05, 0x50,
0x3c, 0x34, 0x1a, 0x36, 0xd8, 0x7e, 0x56, 0x07, 0x2e, 0xab, 0xe2, 0x37,
0x31, 0xe3, 0x4a, 0xf7, 0xe2, 0x32, 0x8f, 0x85, 0xcd, 0xb3, 0x70, 0xcc,
0xaf, 0x00, 0x51, 0x5b, 0x64, 0xc9, 0xc5, 0x4b, 0xc8, 0x37, 0x57, 0x84,
0x47, 0xaa, 0xcf, 0xae, 0xd5, 0x96, 0x9a, 0xa3, 0x51, 0xe7, 0xda, 0x4e,
0xfa, 0x7b, 0x11, 0x5c, 0x4c, 0x51, 0xf4, 0xa6, 0x99, 0x77, 0x98, 0x50,
0x29, 0x5c, 0xa7, 0x2d, 0x78, 0x1a, 0xd4, 0x1b, 0xc6, 0x80, 0x53, 0x2b,
0x89, 0xe7, 0x10, 0xe2, 0x18, 0x9e, 0xb3, 0xc5, 0x08, 0x17, 0xba, 0x25,
0x5c, 0x74, 0x74, 0xc9, 0x5c, 0xa9, 0x11, 0x0c, 0xc4, 0x3b, 0x8b, 0xa8,
0xe6, 0x82, 0xc7, 0xfb, 0x7b, 0x0f, 0xdc, 0x26, 0x5c, 0x04, 0x83, 0xa6,
0x5c, 0xa4, 0x51, 0x4e, 0xe4, 0xb8, 0x32, 0xaa, 0xc5, 0x80, 0x0c, 0x3b,
0x08, 0xe7, 0x4f, 0x56, 0x39, 0x51, 0xc1, 0xfb, 0xb2, 0x10, 0x35, 0x3e,
0xfa, 0x1a, 0xa8, 0x66, 0x85, 0x6b, 0xc1, 0xe0, 0x34, 0x73, 0x3b, 0x04,
0x85, 0xda, 0xb1, 0xd0, 0x20, 0xc6, 0xbf, 0x76, 0x5f, 0xf6, 0x0b, 0x3b,
0x80, 0x19, 0x84, 0xa9, 0x0c, 0x2f, 0xe9, 0x70, 0xbf, 0x1d, 0xe9, 0x70,
0x04, 0xa6, 0xcf, 0x44, 0xb4, 0x98, 0x4a, 0xb5, 0x82, 0x58, 0xb4, 0xaf,
0x71, 0x22, 0x1c, 0xd1, 0x75, 0x30, 0xa7, 0x00, 0xc3, 0x29, 0x59, 0xc9,
0x43, 0x63, 0x44, 0xb5, 0x31, 0x6f, 0x09, 0xcc, 0xca, 0x70, 0x29, 0xa2,
0x30, 0xd6, 0x39, 0xdc, 0xb0, 0x22, 0xd8, 0xba, 0x79, 0xba, 0x91, 0xcd,
0x6a, 0xb1, 0x2a, 0xe1, 0x57, 0x9c, 0x50, 0xc7, 0xbb, 0x10, 0xe3, 0x03,
0x01, 0xa6, 0x5c, 0xae, 0x31, 0x01, 0xd4, 0x0c, 0x7b, 0xa9, 0x27, 0xbb,
0x55, 0x31, 0x48, 0xd1, 0x64, 0x70, 0x24, 0xd4, 0xa0, 0x6c, 0x81, 0x66,
0xd0, 0xb0, 0xb8, 0x12, 0x69, 0xb7, 0xd5, 0xf4, 0xb3, 0x4f, 0xb0, 0x22,
0xf6, 0x91, 0x52, 0xf5, 0x14, 0x00, 0x4a, 0x7c, 0x68, 0x53, 0x68, 0x55,
0x23, 0x43, 0xbb, 0x60, 0x36, 0x0f, 0xbb, 0x99, 0x45, 0xed, 0xf4, 0x46,
0xd3, 0x45, 0xbd, 0xca, 0xa7, 0x45, 0x5c, 0x74, 0xba, 0x0a, 0x55, 0x1e,
0x18, 0x46, 0x20, 0xfe, 0xf9, 0x76, 0x88, 0x77, 0x3d, 0x50, 0xb6, 0x43,
0x3c, 0xa7, 0xa7, 0xac, 0x5c, 0xb6, 0xb7, 0xf6, 0x71, 0xa1, 0x53, 0x76,
0xe5, 0xa6, 0x74, 0x7a, 0x62, 0x3f, 0xa7, 0xbc, 0x66, 0x30, 0x37, 0x3f,
0x5b, 0x1b, 0x51, 0x26, 0x90, 0xa6, 0x61, 0x37, 0x78, 0x70, 0xa6, 0x0a,
0x7a, 0x18, 0x96, 0x83, 0xf9, 0xb0, 0xcf, 0x04, 0x66, 0xe1, 0xf7, 0x50,
0x76, 0x26, 0x31, 0xc4, 0xab, 0x09, 0xf5, 0x05, 0xc4, 0x2d, 0xd2, 0x86,
0x33, 0x56, 0x94, 0x72, 0x73, 0x54, 0x42, 0x85, 0x1e, 0x32, 0x16, 0x16,
0xd4, 0x00, 0x98, 0x10, 0x77, 0x7b, 0x6b, 0xd4, 0x6f, 0xa7, 0x22, 0x44,
0x61, 0xa5, 0xcc, 0x27, 0x40, 0x5d, 0xfb, 0xac, 0x0d, 0x39, 0xb0, 0x02,
0xca, 0xb3, 0x34, 0x33, 0xf2, 0xa8, 0x6e, 0xb8, 0xce, 0x91, 0xc1, 0x34,
0xa6, 0x38, 0x6f, 0x86, 0x0a, 0x19, 0x94, 0xeb, 0x4b, 0x68, 0x75, 0xa4,
0x6d, 0x19, 0x55, 0x81, 0xd1, 0x73, 0x85, 0x4b, 0x53, 0xd2, 0x29, 0x3d,
0xf3, 0xe9, 0xa8, 0x22, 0x75, 0x6c, 0xd8, 0xf2, 0x12, 0xb3, 0x25, 0xca,
0x29, 0xb4, 0xf9, 0xf8, 0xcf, 0xba, 0xdf, 0x2e, 0x41, 0x86, 0x9a, 0xbf,
0xba, 0xd1, 0x07, 0x38, 0xad, 0x04, 0xcc, 0x75, 0x2b, 0xc2, 0x0c, 0x39,
0x47, 0x46, 0x85, 0x0e, 0x0c, 0x48, 0x47, 0xdb
};
#endif
Add a test for the raw private/public key getters Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6394)
2018-06-01 22:06:52 +08:00
static struct keys_st {
int type;
Multi-variant ML-KEM This introduces support for ML-KEM-512 and ML-KEM-1024 using the same underlying implementation parameterised by a few macros for the associated types and constants. KAT tests are added for ML-KEM 512 and 1024, to complement the previous tests for ML-KEM-768. MLKEM{512,768,1024} TLS "group" codepoints are updated to match the final IANA assigments and to make the additional KEMs known to the TLS layer. The pure-QC MLKEMs are not in the default list of supported groups, and need to be explicitly enabled by the application. Future work will introduce support for hybrids, and for more fine-grained policy of which keyshares a client should send by default, and when a server should request (HRR) a new mutually-supported group that was not sent. Tests for ML-KEM key exchange added to sslapitest to make sure that our TLS client MLKEM{512,768,1024} implementations interoperate with our TLS server, and that MLKEM* are not negotiated in TLS 1.2. Tests also added to excercise non-derandomised ML-KEM APIs, both directly (bypassing the provider layer), and through the generic EVP KEM API (exercising the provider). These make sure that RNG input is used correctly (KAT tests bypass the RNG by specifying seeds). The API interface to the provider takes an "const ML_KEM_VINFO" pointer, (obtained from ossl_ml_kem_get_vinfo()). This checks input and output buffer sizes before passing control to internal code that assumes correctly sized (for each variant) buffers. The original BoringSSL API was refactored to eliminate the opaque public/private key structure wrappers, since these structures are an internal detail between libcrypto and the provider, they are not part of the public (EVP) API. New "clangover" counter-measures added, refined with much appreciated input from David Benjamin (Chromium). The internal steps of "encrypt_cpa" were reordered to reduce the working-set size of the algorithm, now needs space for just two temporary "vectors" rather than three. The "decap" function now process the decrypted message in one call, rather than three separate calls to scalar_decode_1, scalar_decompress and scalar_add. Some loops were unrolled, improving performance of en/decapsulate (pre-expanded vectors and matrix) by around 5%. To handle, however unlikely, the SHA3 primitives not behaving like "pure" functions and failing, the implementation of `decap` was modifed: - To use the KDF to compute the Fujisaki-Okamoto (FO) failure secret first thing, and if that fails, bail out returning an error, a shared secret is still returned at random from the RNG, but it is OK for the caller to not use it. - If any of the subsequently used hash primitives fail, use the computed FO failure secret (OK, despite no longer constant-time) and return success (otherwise the RNG would replace the result). - We quite reasonably assume that chosen-ciphertext attacks (of the correct length) cannot cause hash functions to fail in a manner the depends on the private key content. Support for ML-KEM-512 required adding a centered binomial distribution helper function to deal with η_1 == 3 in just that variant. Some additional comments were added to highlight how the code relates to the ML-KEM specification in FIPS 203. Reviewed-by: Paul Dale <ppzgs1@gmail.com> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26172)
2024-11-30 22:20:58 +08:00
const uint8_t *priv;
const uint8_t *pub;
/* If "name" is non-NULL, create via non-legacy _ex() API */
char *name;
int privlen;
int publen;
Add a test for the raw private/public key getters Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6394)
2018-06-01 22:06:52 +08:00
} keys[] = {
{
Multi-variant ML-KEM This introduces support for ML-KEM-512 and ML-KEM-1024 using the same underlying implementation parameterised by a few macros for the associated types and constants. KAT tests are added for ML-KEM 512 and 1024, to complement the previous tests for ML-KEM-768. MLKEM{512,768,1024} TLS "group" codepoints are updated to match the final IANA assigments and to make the additional KEMs known to the TLS layer. The pure-QC MLKEMs are not in the default list of supported groups, and need to be explicitly enabled by the application. Future work will introduce support for hybrids, and for more fine-grained policy of which keyshares a client should send by default, and when a server should request (HRR) a new mutually-supported group that was not sent. Tests for ML-KEM key exchange added to sslapitest to make sure that our TLS client MLKEM{512,768,1024} implementations interoperate with our TLS server, and that MLKEM* are not negotiated in TLS 1.2. Tests also added to excercise non-derandomised ML-KEM APIs, both directly (bypassing the provider layer), and through the generic EVP KEM API (exercising the provider). These make sure that RNG input is used correctly (KAT tests bypass the RNG by specifying seeds). The API interface to the provider takes an "const ML_KEM_VINFO" pointer, (obtained from ossl_ml_kem_get_vinfo()). This checks input and output buffer sizes before passing control to internal code that assumes correctly sized (for each variant) buffers. The original BoringSSL API was refactored to eliminate the opaque public/private key structure wrappers, since these structures are an internal detail between libcrypto and the provider, they are not part of the public (EVP) API. New "clangover" counter-measures added, refined with much appreciated input from David Benjamin (Chromium). The internal steps of "encrypt_cpa" were reordered to reduce the working-set size of the algorithm, now needs space for just two temporary "vectors" rather than three. The "decap" function now process the decrypted message in one call, rather than three separate calls to scalar_decode_1, scalar_decompress and scalar_add. Some loops were unrolled, improving performance of en/decapsulate (pre-expanded vectors and matrix) by around 5%. To handle, however unlikely, the SHA3 primitives not behaving like "pure" functions and failing, the implementation of `decap` was modifed: - To use the KDF to compute the Fujisaki-Okamoto (FO) failure secret first thing, and if that fails, bail out returning an error, a shared secret is still returned at random from the RNG, but it is OK for the caller to not use it. - If any of the subsequently used hash primitives fail, use the computed FO failure secret (OK, despite no longer constant-time) and return success (otherwise the RNG would replace the result). - We quite reasonably assume that chosen-ciphertext attacks (of the correct length) cannot cause hash functions to fail in a manner the depends on the private key content. Support for ML-KEM-512 required adding a centered binomial distribution helper function to deal with η_1 == 3 in just that variant. Some additional comments were added to highlight how the code relates to the ML-KEM specification in FIPS 203. Reviewed-by: Paul Dale <ppzgs1@gmail.com> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26172)
2024-11-30 22:20:58 +08:00
EVP_PKEY_HMAC,
(const uint8_t *)"0123456789",
NULL,
NULL, 10, 0
EVP_PKEY_new_raw_private_key: Allow zero length keys Allocate at least one byte to distinguish a zero length key from an unset key. Fixes #15632 Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/15643)
2021-06-07 17:54:04 +08:00
},
{
Multi-variant ML-KEM This introduces support for ML-KEM-512 and ML-KEM-1024 using the same underlying implementation parameterised by a few macros for the associated types and constants. KAT tests are added for ML-KEM 512 and 1024, to complement the previous tests for ML-KEM-768. MLKEM{512,768,1024} TLS "group" codepoints are updated to match the final IANA assigments and to make the additional KEMs known to the TLS layer. The pure-QC MLKEMs are not in the default list of supported groups, and need to be explicitly enabled by the application. Future work will introduce support for hybrids, and for more fine-grained policy of which keyshares a client should send by default, and when a server should request (HRR) a new mutually-supported group that was not sent. Tests for ML-KEM key exchange added to sslapitest to make sure that our TLS client MLKEM{512,768,1024} implementations interoperate with our TLS server, and that MLKEM* are not negotiated in TLS 1.2. Tests also added to excercise non-derandomised ML-KEM APIs, both directly (bypassing the provider layer), and through the generic EVP KEM API (exercising the provider). These make sure that RNG input is used correctly (KAT tests bypass the RNG by specifying seeds). The API interface to the provider takes an "const ML_KEM_VINFO" pointer, (obtained from ossl_ml_kem_get_vinfo()). This checks input and output buffer sizes before passing control to internal code that assumes correctly sized (for each variant) buffers. The original BoringSSL API was refactored to eliminate the opaque public/private key structure wrappers, since these structures are an internal detail between libcrypto and the provider, they are not part of the public (EVP) API. New "clangover" counter-measures added, refined with much appreciated input from David Benjamin (Chromium). The internal steps of "encrypt_cpa" were reordered to reduce the working-set size of the algorithm, now needs space for just two temporary "vectors" rather than three. The "decap" function now process the decrypted message in one call, rather than three separate calls to scalar_decode_1, scalar_decompress and scalar_add. Some loops were unrolled, improving performance of en/decapsulate (pre-expanded vectors and matrix) by around 5%. To handle, however unlikely, the SHA3 primitives not behaving like "pure" functions and failing, the implementation of `decap` was modifed: - To use the KDF to compute the Fujisaki-Okamoto (FO) failure secret first thing, and if that fails, bail out returning an error, a shared secret is still returned at random from the RNG, but it is OK for the caller to not use it. - If any of the subsequently used hash primitives fail, use the computed FO failure secret (OK, despite no longer constant-time) and return success (otherwise the RNG would replace the result). - We quite reasonably assume that chosen-ciphertext attacks (of the correct length) cannot cause hash functions to fail in a manner the depends on the private key content. Support for ML-KEM-512 required adding a centered binomial distribution helper function to deal with η_1 == 3 in just that variant. Some additional comments were added to highlight how the code relates to the ML-KEM specification in FIPS 203. Reviewed-by: Paul Dale <ppzgs1@gmail.com> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26172)
2024-11-30 22:20:58 +08:00
EVP_PKEY_HMAC, (const uint8_t *)"", NULL,
NULL, 0, 0
Trivial shortcuts for EVP_PKEY_eq() Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14942)
2021-04-20 22:39:00 +08:00
#ifndef OPENSSL_NO_POLY1305
Add a test for the raw private/public key getters Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6394)
2018-06-01 22:06:52 +08:00
}, {
Multi-variant ML-KEM This introduces support for ML-KEM-512 and ML-KEM-1024 using the same underlying implementation parameterised by a few macros for the associated types and constants. KAT tests are added for ML-KEM 512 and 1024, to complement the previous tests for ML-KEM-768. MLKEM{512,768,1024} TLS "group" codepoints are updated to match the final IANA assigments and to make the additional KEMs known to the TLS layer. The pure-QC MLKEMs are not in the default list of supported groups, and need to be explicitly enabled by the application. Future work will introduce support for hybrids, and for more fine-grained policy of which keyshares a client should send by default, and when a server should request (HRR) a new mutually-supported group that was not sent. Tests for ML-KEM key exchange added to sslapitest to make sure that our TLS client MLKEM{512,768,1024} implementations interoperate with our TLS server, and that MLKEM* are not negotiated in TLS 1.2. Tests also added to excercise non-derandomised ML-KEM APIs, both directly (bypassing the provider layer), and through the generic EVP KEM API (exercising the provider). These make sure that RNG input is used correctly (KAT tests bypass the RNG by specifying seeds). The API interface to the provider takes an "const ML_KEM_VINFO" pointer, (obtained from ossl_ml_kem_get_vinfo()). This checks input and output buffer sizes before passing control to internal code that assumes correctly sized (for each variant) buffers. The original BoringSSL API was refactored to eliminate the opaque public/private key structure wrappers, since these structures are an internal detail between libcrypto and the provider, they are not part of the public (EVP) API. New "clangover" counter-measures added, refined with much appreciated input from David Benjamin (Chromium). The internal steps of "encrypt_cpa" were reordered to reduce the working-set size of the algorithm, now needs space for just two temporary "vectors" rather than three. The "decap" function now process the decrypted message in one call, rather than three separate calls to scalar_decode_1, scalar_decompress and scalar_add. Some loops were unrolled, improving performance of en/decapsulate (pre-expanded vectors and matrix) by around 5%. To handle, however unlikely, the SHA3 primitives not behaving like "pure" functions and failing, the implementation of `decap` was modifed: - To use the KDF to compute the Fujisaki-Okamoto (FO) failure secret first thing, and if that fails, bail out returning an error, a shared secret is still returned at random from the RNG, but it is OK for the caller to not use it. - If any of the subsequently used hash primitives fail, use the computed FO failure secret (OK, despite no longer constant-time) and return success (otherwise the RNG would replace the result). - We quite reasonably assume that chosen-ciphertext attacks (of the correct length) cannot cause hash functions to fail in a manner the depends on the private key content. Support for ML-KEM-512 required adding a centered binomial distribution helper function to deal with η_1 == 3 in just that variant. Some additional comments were added to highlight how the code relates to the ML-KEM specification in FIPS 203. Reviewed-by: Paul Dale <ppzgs1@gmail.com> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26172)
2024-11-30 22:20:58 +08:00
EVP_PKEY_POLY1305,
(const uint8_t *)"01234567890123456789012345678901",
NULL,
NULL, 32, 0
Trivial shortcuts for EVP_PKEY_eq() Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14942)
2021-04-20 22:39:00 +08:00
#endif
#ifndef OPENSSL_NO_SIPHASH
Add a test for the raw private/public key getters Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6394)
2018-06-01 22:06:52 +08:00
}, {
Multi-variant ML-KEM This introduces support for ML-KEM-512 and ML-KEM-1024 using the same underlying implementation parameterised by a few macros for the associated types and constants. KAT tests are added for ML-KEM 512 and 1024, to complement the previous tests for ML-KEM-768. MLKEM{512,768,1024} TLS "group" codepoints are updated to match the final IANA assigments and to make the additional KEMs known to the TLS layer. The pure-QC MLKEMs are not in the default list of supported groups, and need to be explicitly enabled by the application. Future work will introduce support for hybrids, and for more fine-grained policy of which keyshares a client should send by default, and when a server should request (HRR) a new mutually-supported group that was not sent. Tests for ML-KEM key exchange added to sslapitest to make sure that our TLS client MLKEM{512,768,1024} implementations interoperate with our TLS server, and that MLKEM* are not negotiated in TLS 1.2. Tests also added to excercise non-derandomised ML-KEM APIs, both directly (bypassing the provider layer), and through the generic EVP KEM API (exercising the provider). These make sure that RNG input is used correctly (KAT tests bypass the RNG by specifying seeds). The API interface to the provider takes an "const ML_KEM_VINFO" pointer, (obtained from ossl_ml_kem_get_vinfo()). This checks input and output buffer sizes before passing control to internal code that assumes correctly sized (for each variant) buffers. The original BoringSSL API was refactored to eliminate the opaque public/private key structure wrappers, since these structures are an internal detail between libcrypto and the provider, they are not part of the public (EVP) API. New "clangover" counter-measures added, refined with much appreciated input from David Benjamin (Chromium). The internal steps of "encrypt_cpa" were reordered to reduce the working-set size of the algorithm, now needs space for just two temporary "vectors" rather than three. The "decap" function now process the decrypted message in one call, rather than three separate calls to scalar_decode_1, scalar_decompress and scalar_add. Some loops were unrolled, improving performance of en/decapsulate (pre-expanded vectors and matrix) by around 5%. To handle, however unlikely, the SHA3 primitives not behaving like "pure" functions and failing, the implementation of `decap` was modifed: - To use the KDF to compute the Fujisaki-Okamoto (FO) failure secret first thing, and if that fails, bail out returning an error, a shared secret is still returned at random from the RNG, but it is OK for the caller to not use it. - If any of the subsequently used hash primitives fail, use the computed FO failure secret (OK, despite no longer constant-time) and return success (otherwise the RNG would replace the result). - We quite reasonably assume that chosen-ciphertext attacks (of the correct length) cannot cause hash functions to fail in a manner the depends on the private key content. Support for ML-KEM-512 required adding a centered binomial distribution helper function to deal with η_1 == 3 in just that variant. Some additional comments were added to highlight how the code relates to the ML-KEM specification in FIPS 203. Reviewed-by: Paul Dale <ppzgs1@gmail.com> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26172)
2024-11-30 22:20:58 +08:00
EVP_PKEY_SIPHASH,
(const uint8_t *)"0123456789012345",
NULL,
NULL, 16, 0
Trivial shortcuts for EVP_PKEY_eq() Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14942)
2021-04-20 22:39:00 +08:00
#endif
Fix no-ec Reviewed-by: Rich Salz <rsalz@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6437)
2018-06-08 18:20:34 +08:00
},
configure: introduce no-ecx to remove ECX related feature This can effectively reduce the binary size for platforms that don't need ECX feature(~100KB). Signed-off-by: Yi Li <yi1.li@intel.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/20781)
2023-04-17 16:20:31 +08:00
#ifndef OPENSSL_NO_ECX
Fix no-ec Reviewed-by: Rich Salz <rsalz@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6437)
2018-06-08 18:20:34 +08:00
{
Multi-variant ML-KEM This introduces support for ML-KEM-512 and ML-KEM-1024 using the same underlying implementation parameterised by a few macros for the associated types and constants. KAT tests are added for ML-KEM 512 and 1024, to complement the previous tests for ML-KEM-768. MLKEM{512,768,1024} TLS "group" codepoints are updated to match the final IANA assigments and to make the additional KEMs known to the TLS layer. The pure-QC MLKEMs are not in the default list of supported groups, and need to be explicitly enabled by the application. Future work will introduce support for hybrids, and for more fine-grained policy of which keyshares a client should send by default, and when a server should request (HRR) a new mutually-supported group that was not sent. Tests for ML-KEM key exchange added to sslapitest to make sure that our TLS client MLKEM{512,768,1024} implementations interoperate with our TLS server, and that MLKEM* are not negotiated in TLS 1.2. Tests also added to excercise non-derandomised ML-KEM APIs, both directly (bypassing the provider layer), and through the generic EVP KEM API (exercising the provider). These make sure that RNG input is used correctly (KAT tests bypass the RNG by specifying seeds). The API interface to the provider takes an "const ML_KEM_VINFO" pointer, (obtained from ossl_ml_kem_get_vinfo()). This checks input and output buffer sizes before passing control to internal code that assumes correctly sized (for each variant) buffers. The original BoringSSL API was refactored to eliminate the opaque public/private key structure wrappers, since these structures are an internal detail between libcrypto and the provider, they are not part of the public (EVP) API. New "clangover" counter-measures added, refined with much appreciated input from David Benjamin (Chromium). The internal steps of "encrypt_cpa" were reordered to reduce the working-set size of the algorithm, now needs space for just two temporary "vectors" rather than three. The "decap" function now process the decrypted message in one call, rather than three separate calls to scalar_decode_1, scalar_decompress and scalar_add. Some loops were unrolled, improving performance of en/decapsulate (pre-expanded vectors and matrix) by around 5%. To handle, however unlikely, the SHA3 primitives not behaving like "pure" functions and failing, the implementation of `decap` was modifed: - To use the KDF to compute the Fujisaki-Okamoto (FO) failure secret first thing, and if that fails, bail out returning an error, a shared secret is still returned at random from the RNG, but it is OK for the caller to not use it. - If any of the subsequently used hash primitives fail, use the computed FO failure secret (OK, despite no longer constant-time) and return success (otherwise the RNG would replace the result). - We quite reasonably assume that chosen-ciphertext attacks (of the correct length) cannot cause hash functions to fail in a manner the depends on the private key content. Support for ML-KEM-512 required adding a centered binomial distribution helper function to deal with η_1 == 3 in just that variant. Some additional comments were added to highlight how the code relates to the ML-KEM specification in FIPS 203. Reviewed-by: Paul Dale <ppzgs1@gmail.com> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26172)
2024-11-30 22:20:58 +08:00
EVP_PKEY_X25519,
(const uint8_t *)"01234567890123456789012345678901",
(const unsigned char *)"abcdefghijklmnopqrstuvwxyzabcdef",
NULL, 32, 32
Add a test for the raw private/public key getters Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6394)
2018-06-01 22:06:52 +08:00
}, {
Multi-variant ML-KEM This introduces support for ML-KEM-512 and ML-KEM-1024 using the same underlying implementation parameterised by a few macros for the associated types and constants. KAT tests are added for ML-KEM 512 and 1024, to complement the previous tests for ML-KEM-768. MLKEM{512,768,1024} TLS "group" codepoints are updated to match the final IANA assigments and to make the additional KEMs known to the TLS layer. The pure-QC MLKEMs are not in the default list of supported groups, and need to be explicitly enabled by the application. Future work will introduce support for hybrids, and for more fine-grained policy of which keyshares a client should send by default, and when a server should request (HRR) a new mutually-supported group that was not sent. Tests for ML-KEM key exchange added to sslapitest to make sure that our TLS client MLKEM{512,768,1024} implementations interoperate with our TLS server, and that MLKEM* are not negotiated in TLS 1.2. Tests also added to excercise non-derandomised ML-KEM APIs, both directly (bypassing the provider layer), and through the generic EVP KEM API (exercising the provider). These make sure that RNG input is used correctly (KAT tests bypass the RNG by specifying seeds). The API interface to the provider takes an "const ML_KEM_VINFO" pointer, (obtained from ossl_ml_kem_get_vinfo()). This checks input and output buffer sizes before passing control to internal code that assumes correctly sized (for each variant) buffers. The original BoringSSL API was refactored to eliminate the opaque public/private key structure wrappers, since these structures are an internal detail between libcrypto and the provider, they are not part of the public (EVP) API. New "clangover" counter-measures added, refined with much appreciated input from David Benjamin (Chromium). The internal steps of "encrypt_cpa" were reordered to reduce the working-set size of the algorithm, now needs space for just two temporary "vectors" rather than three. The "decap" function now process the decrypted message in one call, rather than three separate calls to scalar_decode_1, scalar_decompress and scalar_add. Some loops were unrolled, improving performance of en/decapsulate (pre-expanded vectors and matrix) by around 5%. To handle, however unlikely, the SHA3 primitives not behaving like "pure" functions and failing, the implementation of `decap` was modifed: - To use the KDF to compute the Fujisaki-Okamoto (FO) failure secret first thing, and if that fails, bail out returning an error, a shared secret is still returned at random from the RNG, but it is OK for the caller to not use it. - If any of the subsequently used hash primitives fail, use the computed FO failure secret (OK, despite no longer constant-time) and return success (otherwise the RNG would replace the result). - We quite reasonably assume that chosen-ciphertext attacks (of the correct length) cannot cause hash functions to fail in a manner the depends on the private key content. Support for ML-KEM-512 required adding a centered binomial distribution helper function to deal with η_1 == 3 in just that variant. Some additional comments were added to highlight how the code relates to the ML-KEM specification in FIPS 203. Reviewed-by: Paul Dale <ppzgs1@gmail.com> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26172)
2024-11-30 22:20:58 +08:00
EVP_PKEY_ED25519,
(const uint8_t *)"01234567890123456789012345678901",
(const uint8_t *)"abcdefghijklmnopqrstuvwxyzabcdef",
NULL, 32, 32
Add a test for the raw private/public key getters Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6394)
2018-06-01 22:06:52 +08:00
}, {
EVP_PKEY_X448,
Multi-variant ML-KEM This introduces support for ML-KEM-512 and ML-KEM-1024 using the same underlying implementation parameterised by a few macros for the associated types and constants. KAT tests are added for ML-KEM 512 and 1024, to complement the previous tests for ML-KEM-768. MLKEM{512,768,1024} TLS "group" codepoints are updated to match the final IANA assigments and to make the additional KEMs known to the TLS layer. The pure-QC MLKEMs are not in the default list of supported groups, and need to be explicitly enabled by the application. Future work will introduce support for hybrids, and for more fine-grained policy of which keyshares a client should send by default, and when a server should request (HRR) a new mutually-supported group that was not sent. Tests for ML-KEM key exchange added to sslapitest to make sure that our TLS client MLKEM{512,768,1024} implementations interoperate with our TLS server, and that MLKEM* are not negotiated in TLS 1.2. Tests also added to excercise non-derandomised ML-KEM APIs, both directly (bypassing the provider layer), and through the generic EVP KEM API (exercising the provider). These make sure that RNG input is used correctly (KAT tests bypass the RNG by specifying seeds). The API interface to the provider takes an "const ML_KEM_VINFO" pointer, (obtained from ossl_ml_kem_get_vinfo()). This checks input and output buffer sizes before passing control to internal code that assumes correctly sized (for each variant) buffers. The original BoringSSL API was refactored to eliminate the opaque public/private key structure wrappers, since these structures are an internal detail between libcrypto and the provider, they are not part of the public (EVP) API. New "clangover" counter-measures added, refined with much appreciated input from David Benjamin (Chromium). The internal steps of "encrypt_cpa" were reordered to reduce the working-set size of the algorithm, now needs space for just two temporary "vectors" rather than three. The "decap" function now process the decrypted message in one call, rather than three separate calls to scalar_decode_1, scalar_decompress and scalar_add. Some loops were unrolled, improving performance of en/decapsulate (pre-expanded vectors and matrix) by around 5%. To handle, however unlikely, the SHA3 primitives not behaving like "pure" functions and failing, the implementation of `decap` was modifed: - To use the KDF to compute the Fujisaki-Okamoto (FO) failure secret first thing, and if that fails, bail out returning an error, a shared secret is still returned at random from the RNG, but it is OK for the caller to not use it. - If any of the subsequently used hash primitives fail, use the computed FO failure secret (OK, despite no longer constant-time) and return success (otherwise the RNG would replace the result). - We quite reasonably assume that chosen-ciphertext attacks (of the correct length) cannot cause hash functions to fail in a manner the depends on the private key content. Support for ML-KEM-512 required adding a centered binomial distribution helper function to deal with η_1 == 3 in just that variant. Some additional comments were added to highlight how the code relates to the ML-KEM specification in FIPS 203. Reviewed-by: Paul Dale <ppzgs1@gmail.com> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26172)
2024-11-30 22:20:58 +08:00
(const uint8_t *)"01234567890123456789012345678901234567890123456789012345",
(const uint8_t *)"abcdefghijklmnopqrstuvwxyzabcdefghijklmnopqrstuvwxyzabcd",
NULL, 56, 56
Add a test for the raw private/public key getters Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6394)
2018-06-01 22:06:52 +08:00
}, {
EVP_PKEY_ED448,
Multi-variant ML-KEM This introduces support for ML-KEM-512 and ML-KEM-1024 using the same underlying implementation parameterised by a few macros for the associated types and constants. KAT tests are added for ML-KEM 512 and 1024, to complement the previous tests for ML-KEM-768. MLKEM{512,768,1024} TLS "group" codepoints are updated to match the final IANA assigments and to make the additional KEMs known to the TLS layer. The pure-QC MLKEMs are not in the default list of supported groups, and need to be explicitly enabled by the application. Future work will introduce support for hybrids, and for more fine-grained policy of which keyshares a client should send by default, and when a server should request (HRR) a new mutually-supported group that was not sent. Tests for ML-KEM key exchange added to sslapitest to make sure that our TLS client MLKEM{512,768,1024} implementations interoperate with our TLS server, and that MLKEM* are not negotiated in TLS 1.2. Tests also added to excercise non-derandomised ML-KEM APIs, both directly (bypassing the provider layer), and through the generic EVP KEM API (exercising the provider). These make sure that RNG input is used correctly (KAT tests bypass the RNG by specifying seeds). The API interface to the provider takes an "const ML_KEM_VINFO" pointer, (obtained from ossl_ml_kem_get_vinfo()). This checks input and output buffer sizes before passing control to internal code that assumes correctly sized (for each variant) buffers. The original BoringSSL API was refactored to eliminate the opaque public/private key structure wrappers, since these structures are an internal detail between libcrypto and the provider, they are not part of the public (EVP) API. New "clangover" counter-measures added, refined with much appreciated input from David Benjamin (Chromium). The internal steps of "encrypt_cpa" were reordered to reduce the working-set size of the algorithm, now needs space for just two temporary "vectors" rather than three. The "decap" function now process the decrypted message in one call, rather than three separate calls to scalar_decode_1, scalar_decompress and scalar_add. Some loops were unrolled, improving performance of en/decapsulate (pre-expanded vectors and matrix) by around 5%. To handle, however unlikely, the SHA3 primitives not behaving like "pure" functions and failing, the implementation of `decap` was modifed: - To use the KDF to compute the Fujisaki-Okamoto (FO) failure secret first thing, and if that fails, bail out returning an error, a shared secret is still returned at random from the RNG, but it is OK for the caller to not use it. - If any of the subsequently used hash primitives fail, use the computed FO failure secret (OK, despite no longer constant-time) and return success (otherwise the RNG would replace the result). - We quite reasonably assume that chosen-ciphertext attacks (of the correct length) cannot cause hash functions to fail in a manner the depends on the private key content. Support for ML-KEM-512 required adding a centered binomial distribution helper function to deal with η_1 == 3 in just that variant. Some additional comments were added to highlight how the code relates to the ML-KEM specification in FIPS 203. Reviewed-by: Paul Dale <ppzgs1@gmail.com> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26172)
2024-11-30 22:20:58 +08:00
(const uint8_t *)"012345678901234567890123456789012345678901234567890123456",
(const uint8_t *)"abcdefghijklmnopqrstuvwxyzabcdefghijklmnopqrstuvwxyzabcde",
NULL, 57, 57
},
#endif
#ifndef OPENSSL_NO_ML_KEM
{
Improved import and export - On import, if a seed is provided, the keys are regenerated. - The seed is exported as a separate "seed" parameter, when available. The "ml-kem.retain_seed" parameter is also exported, when false. - The seed is optionally dropped after key generation. * When the "ml-kem.retain_seed" keygen parameter is set to zero. * When the "ml-kem.retain_seed" keygen parameter is not set to 1, and the "ml-kem.retain_seed" provider config property is set explictly false. - The exported private key parameter "priv" is always the FIPS 203 |dk|. - Private key decoding from PKCS#8 produces a transient "seed-only" form of the key, in which "retain_seed" is set to false when the "ml-kem.retain_seed" provider config property is set explictly false. The full key is generated during "load" and the seed is retained or not as specified. - Import honours the "ml-kem.retain_seed" parameter when specified, or otherwise honours the provider's "ml-kem.retain_seed" property. Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26512)
2025-01-18 00:28:51 +08:00
NID_undef, ml_kem_seed, ml_kem_512_pubkey,
Encoders and Decoders for ML-KEM Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26341)
2025-01-09 11:35:03 +08:00
"ML-KEM-512", ML_KEM_SEED_BYTES, sizeof(ml_kem_512_pubkey)
Multi-variant ML-KEM This introduces support for ML-KEM-512 and ML-KEM-1024 using the same underlying implementation parameterised by a few macros for the associated types and constants. KAT tests are added for ML-KEM 512 and 1024, to complement the previous tests for ML-KEM-768. MLKEM{512,768,1024} TLS "group" codepoints are updated to match the final IANA assigments and to make the additional KEMs known to the TLS layer. The pure-QC MLKEMs are not in the default list of supported groups, and need to be explicitly enabled by the application. Future work will introduce support for hybrids, and for more fine-grained policy of which keyshares a client should send by default, and when a server should request (HRR) a new mutually-supported group that was not sent. Tests for ML-KEM key exchange added to sslapitest to make sure that our TLS client MLKEM{512,768,1024} implementations interoperate with our TLS server, and that MLKEM* are not negotiated in TLS 1.2. Tests also added to excercise non-derandomised ML-KEM APIs, both directly (bypassing the provider layer), and through the generic EVP KEM API (exercising the provider). These make sure that RNG input is used correctly (KAT tests bypass the RNG by specifying seeds). The API interface to the provider takes an "const ML_KEM_VINFO" pointer, (obtained from ossl_ml_kem_get_vinfo()). This checks input and output buffer sizes before passing control to internal code that assumes correctly sized (for each variant) buffers. The original BoringSSL API was refactored to eliminate the opaque public/private key structure wrappers, since these structures are an internal detail between libcrypto and the provider, they are not part of the public (EVP) API. New "clangover" counter-measures added, refined with much appreciated input from David Benjamin (Chromium). The internal steps of "encrypt_cpa" were reordered to reduce the working-set size of the algorithm, now needs space for just two temporary "vectors" rather than three. The "decap" function now process the decrypted message in one call, rather than three separate calls to scalar_decode_1, scalar_decompress and scalar_add. Some loops were unrolled, improving performance of en/decapsulate (pre-expanded vectors and matrix) by around 5%. To handle, however unlikely, the SHA3 primitives not behaving like "pure" functions and failing, the implementation of `decap` was modifed: - To use the KDF to compute the Fujisaki-Okamoto (FO) failure secret first thing, and if that fails, bail out returning an error, a shared secret is still returned at random from the RNG, but it is OK for the caller to not use it. - If any of the subsequently used hash primitives fail, use the computed FO failure secret (OK, despite no longer constant-time) and return success (otherwise the RNG would replace the result). - We quite reasonably assume that chosen-ciphertext attacks (of the correct length) cannot cause hash functions to fail in a manner the depends on the private key content. Support for ML-KEM-512 required adding a centered binomial distribution helper function to deal with η_1 == 3 in just that variant. Some additional comments were added to highlight how the code relates to the ML-KEM specification in FIPS 203. Reviewed-by: Paul Dale <ppzgs1@gmail.com> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26172)
2024-11-30 22:20:58 +08:00
}, {
Improved import and export - On import, if a seed is provided, the keys are regenerated. - The seed is exported as a separate "seed" parameter, when available. The "ml-kem.retain_seed" parameter is also exported, when false. - The seed is optionally dropped after key generation. * When the "ml-kem.retain_seed" keygen parameter is set to zero. * When the "ml-kem.retain_seed" keygen parameter is not set to 1, and the "ml-kem.retain_seed" provider config property is set explictly false. - The exported private key parameter "priv" is always the FIPS 203 |dk|. - Private key decoding from PKCS#8 produces a transient "seed-only" form of the key, in which "retain_seed" is set to false when the "ml-kem.retain_seed" provider config property is set explictly false. The full key is generated during "load" and the seed is retained or not as specified. - Import honours the "ml-kem.retain_seed" parameter when specified, or otherwise honours the provider's "ml-kem.retain_seed" property. Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26512)
2025-01-18 00:28:51 +08:00
NID_undef, ml_kem_seed, ml_kem_768_pubkey,
Encoders and Decoders for ML-KEM Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26341)
2025-01-09 11:35:03 +08:00
"ML-KEM-768", ML_KEM_SEED_BYTES, sizeof(ml_kem_768_pubkey)
Multi-variant ML-KEM This introduces support for ML-KEM-512 and ML-KEM-1024 using the same underlying implementation parameterised by a few macros for the associated types and constants. KAT tests are added for ML-KEM 512 and 1024, to complement the previous tests for ML-KEM-768. MLKEM{512,768,1024} TLS "group" codepoints are updated to match the final IANA assigments and to make the additional KEMs known to the TLS layer. The pure-QC MLKEMs are not in the default list of supported groups, and need to be explicitly enabled by the application. Future work will introduce support for hybrids, and for more fine-grained policy of which keyshares a client should send by default, and when a server should request (HRR) a new mutually-supported group that was not sent. Tests for ML-KEM key exchange added to sslapitest to make sure that our TLS client MLKEM{512,768,1024} implementations interoperate with our TLS server, and that MLKEM* are not negotiated in TLS 1.2. Tests also added to excercise non-derandomised ML-KEM APIs, both directly (bypassing the provider layer), and through the generic EVP KEM API (exercising the provider). These make sure that RNG input is used correctly (KAT tests bypass the RNG by specifying seeds). The API interface to the provider takes an "const ML_KEM_VINFO" pointer, (obtained from ossl_ml_kem_get_vinfo()). This checks input and output buffer sizes before passing control to internal code that assumes correctly sized (for each variant) buffers. The original BoringSSL API was refactored to eliminate the opaque public/private key structure wrappers, since these structures are an internal detail between libcrypto and the provider, they are not part of the public (EVP) API. New "clangover" counter-measures added, refined with much appreciated input from David Benjamin (Chromium). The internal steps of "encrypt_cpa" were reordered to reduce the working-set size of the algorithm, now needs space for just two temporary "vectors" rather than three. The "decap" function now process the decrypted message in one call, rather than three separate calls to scalar_decode_1, scalar_decompress and scalar_add. Some loops were unrolled, improving performance of en/decapsulate (pre-expanded vectors and matrix) by around 5%. To handle, however unlikely, the SHA3 primitives not behaving like "pure" functions and failing, the implementation of `decap` was modifed: - To use the KDF to compute the Fujisaki-Okamoto (FO) failure secret first thing, and if that fails, bail out returning an error, a shared secret is still returned at random from the RNG, but it is OK for the caller to not use it. - If any of the subsequently used hash primitives fail, use the computed FO failure secret (OK, despite no longer constant-time) and return success (otherwise the RNG would replace the result). - We quite reasonably assume that chosen-ciphertext attacks (of the correct length) cannot cause hash functions to fail in a manner the depends on the private key content. Support for ML-KEM-512 required adding a centered binomial distribution helper function to deal with η_1 == 3 in just that variant. Some additional comments were added to highlight how the code relates to the ML-KEM specification in FIPS 203. Reviewed-by: Paul Dale <ppzgs1@gmail.com> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26172)
2024-11-30 22:20:58 +08:00
}, {
Improved import and export - On import, if a seed is provided, the keys are regenerated. - The seed is exported as a separate "seed" parameter, when available. The "ml-kem.retain_seed" parameter is also exported, when false. - The seed is optionally dropped after key generation. * When the "ml-kem.retain_seed" keygen parameter is set to zero. * When the "ml-kem.retain_seed" keygen parameter is not set to 1, and the "ml-kem.retain_seed" provider config property is set explictly false. - The exported private key parameter "priv" is always the FIPS 203 |dk|. - Private key decoding from PKCS#8 produces a transient "seed-only" form of the key, in which "retain_seed" is set to false when the "ml-kem.retain_seed" provider config property is set explictly false. The full key is generated during "load" and the seed is retained or not as specified. - Import honours the "ml-kem.retain_seed" parameter when specified, or otherwise honours the provider's "ml-kem.retain_seed" property. Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26512)
2025-01-18 00:28:51 +08:00
NID_undef, ml_kem_seed, ml_kem_1024_pubkey,
Encoders and Decoders for ML-KEM Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26341)
2025-01-09 11:35:03 +08:00
"ML-KEM-1024", ML_KEM_SEED_BYTES, sizeof(ml_kem_1024_pubkey)
Multi-variant ML-KEM This introduces support for ML-KEM-512 and ML-KEM-1024 using the same underlying implementation parameterised by a few macros for the associated types and constants. KAT tests are added for ML-KEM 512 and 1024, to complement the previous tests for ML-KEM-768. MLKEM{512,768,1024} TLS "group" codepoints are updated to match the final IANA assigments and to make the additional KEMs known to the TLS layer. The pure-QC MLKEMs are not in the default list of supported groups, and need to be explicitly enabled by the application. Future work will introduce support for hybrids, and for more fine-grained policy of which keyshares a client should send by default, and when a server should request (HRR) a new mutually-supported group that was not sent. Tests for ML-KEM key exchange added to sslapitest to make sure that our TLS client MLKEM{512,768,1024} implementations interoperate with our TLS server, and that MLKEM* are not negotiated in TLS 1.2. Tests also added to excercise non-derandomised ML-KEM APIs, both directly (bypassing the provider layer), and through the generic EVP KEM API (exercising the provider). These make sure that RNG input is used correctly (KAT tests bypass the RNG by specifying seeds). The API interface to the provider takes an "const ML_KEM_VINFO" pointer, (obtained from ossl_ml_kem_get_vinfo()). This checks input and output buffer sizes before passing control to internal code that assumes correctly sized (for each variant) buffers. The original BoringSSL API was refactored to eliminate the opaque public/private key structure wrappers, since these structures are an internal detail between libcrypto and the provider, they are not part of the public (EVP) API. New "clangover" counter-measures added, refined with much appreciated input from David Benjamin (Chromium). The internal steps of "encrypt_cpa" were reordered to reduce the working-set size of the algorithm, now needs space for just two temporary "vectors" rather than three. The "decap" function now process the decrypted message in one call, rather than three separate calls to scalar_decode_1, scalar_decompress and scalar_add. Some loops were unrolled, improving performance of en/decapsulate (pre-expanded vectors and matrix) by around 5%. To handle, however unlikely, the SHA3 primitives not behaving like "pure" functions and failing, the implementation of `decap` was modifed: - To use the KDF to compute the Fujisaki-Okamoto (FO) failure secret first thing, and if that fails, bail out returning an error, a shared secret is still returned at random from the RNG, but it is OK for the caller to not use it. - If any of the subsequently used hash primitives fail, use the computed FO failure secret (OK, despite no longer constant-time) and return success (otherwise the RNG would replace the result). - We quite reasonably assume that chosen-ciphertext attacks (of the correct length) cannot cause hash functions to fail in a manner the depends on the private key content. Support for ML-KEM-512 required adding a centered binomial distribution helper function to deal with η_1 == 3 in just that variant. Some additional comments were added to highlight how the code relates to the ML-KEM specification in FIPS 203. Reviewed-by: Paul Dale <ppzgs1@gmail.com> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26172)
2024-11-30 22:20:58 +08:00
},
Fix no-ec Reviewed-by: Rich Salz <rsalz@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6437)
2018-06-08 18:20:34 +08:00
#endif
Add a test for the raw private/public key getters Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6394)
2018-06-01 22:06:52 +08:00
};
Improved import and export - On import, if a seed is provided, the keys are regenerated. - The seed is exported as a separate "seed" parameter, when available. The "ml-kem.retain_seed" parameter is also exported, when false. - The seed is optionally dropped after key generation. * When the "ml-kem.retain_seed" keygen parameter is set to zero. * When the "ml-kem.retain_seed" keygen parameter is not set to 1, and the "ml-kem.retain_seed" provider config property is set explictly false. - The exported private key parameter "priv" is always the FIPS 203 |dk|. - Private key decoding from PKCS#8 produces a transient "seed-only" form of the key, in which "retain_seed" is set to false when the "ml-kem.retain_seed" provider config property is set explictly false. The full key is generated during "load" and the seed is retained or not as specified. - Import honours the "ml-kem.retain_seed" parameter when specified, or otherwise honours the provider's "ml-kem.retain_seed" property. Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26512)
2025-01-18 00:28:51 +08:00
#ifndef OPENSSL_NO_ML_KEM
static int
ml_kem_seed_to_priv(const char *alg, const unsigned char *seed, int seedlen,
unsigned char **ret, size_t *retlen)
{
OSSL_PARAM parr[2] = { OSSL_PARAM_END, OSSL_PARAM_END };
EVP_PKEY_CTX *ctx = NULL;
EVP_PKEY *pkey = NULL;
const OSSL_PARAM *p;
OSSL_PARAM *params = NULL;
int selection = OSSL_KEYMGMT_SELECT_PRIVATE_KEY;
int ok = 0;
/* Import the seed to generate a key */
ctx = EVP_PKEY_CTX_new_from_name(testctx, alg, NULL);
if (!TEST_ptr(ctx)
|| !TEST_int_gt(EVP_PKEY_fromdata_init(ctx), 0))
goto done;
parr[0] = OSSL_PARAM_construct_octet_string(
OSSL_PKEY_PARAM_ML_KEM_SEED, (unsigned char *)seed, seedlen);
if (!TEST_int_gt(EVP_PKEY_fromdata(ctx, &pkey, selection, parr), 0))
goto done;
/* Export the key to get the encoded form */
if (!TEST_true(EVP_PKEY_todata(pkey, OSSL_KEYMGMT_SELECT_PRIVATE_KEY, &params)))
goto done;
p = OSSL_PARAM_locate_const(params, OSSL_PKEY_PARAM_PRIV_KEY);
if (!TEST_ptr(p)
|| !TEST_true(OSSL_PARAM_get_octet_string(p, (void **)ret, 0, retlen)))
goto done;
ok = 1;
done:
EVP_PKEY_free(pkey);
OSSL_PARAM_free(params);
EVP_PKEY_CTX_free(ctx);
return ok;
}
#endif
Add some tests for the newly added raw private/public key functions We already had soem tests for the older raw private/public key functions so we expand those to call the new versions as well and pass in a libctx. Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/11635)
2020-04-28 00:17:05 +08:00
static int test_set_get_raw_keys_int(int tst, int pub, int uselibctx)
Add a test for the raw private/public key getters Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6394)
2018-06-01 22:06:52 +08:00
{
int ret = 0;
Multi-variant ML-KEM This introduces support for ML-KEM-512 and ML-KEM-1024 using the same underlying implementation parameterised by a few macros for the associated types and constants. KAT tests are added for ML-KEM 512 and 1024, to complement the previous tests for ML-KEM-768. MLKEM{512,768,1024} TLS "group" codepoints are updated to match the final IANA assigments and to make the additional KEMs known to the TLS layer. The pure-QC MLKEMs are not in the default list of supported groups, and need to be explicitly enabled by the application. Future work will introduce support for hybrids, and for more fine-grained policy of which keyshares a client should send by default, and when a server should request (HRR) a new mutually-supported group that was not sent. Tests for ML-KEM key exchange added to sslapitest to make sure that our TLS client MLKEM{512,768,1024} implementations interoperate with our TLS server, and that MLKEM* are not negotiated in TLS 1.2. Tests also added to excercise non-derandomised ML-KEM APIs, both directly (bypassing the provider layer), and through the generic EVP KEM API (exercising the provider). These make sure that RNG input is used correctly (KAT tests bypass the RNG by specifying seeds). The API interface to the provider takes an "const ML_KEM_VINFO" pointer, (obtained from ossl_ml_kem_get_vinfo()). This checks input and output buffer sizes before passing control to internal code that assumes correctly sized (for each variant) buffers. The original BoringSSL API was refactored to eliminate the opaque public/private key structure wrappers, since these structures are an internal detail between libcrypto and the provider, they are not part of the public (EVP) API. New "clangover" counter-measures added, refined with much appreciated input from David Benjamin (Chromium). The internal steps of "encrypt_cpa" were reordered to reduce the working-set size of the algorithm, now needs space for just two temporary "vectors" rather than three. The "decap" function now process the decrypted message in one call, rather than three separate calls to scalar_decode_1, scalar_decompress and scalar_add. Some loops were unrolled, improving performance of en/decapsulate (pre-expanded vectors and matrix) by around 5%. To handle, however unlikely, the SHA3 primitives not behaving like "pure" functions and failing, the implementation of `decap` was modifed: - To use the KDF to compute the Fujisaki-Okamoto (FO) failure secret first thing, and if that fails, bail out returning an error, a shared secret is still returned at random from the RNG, but it is OK for the caller to not use it. - If any of the subsequently used hash primitives fail, use the computed FO failure secret (OK, despite no longer constant-time) and return success (otherwise the RNG would replace the result). - We quite reasonably assume that chosen-ciphertext attacks (of the correct length) cannot cause hash functions to fail in a manner the depends on the private key content. Support for ML-KEM-512 required adding a centered binomial distribution helper function to deal with η_1 == 3 in just that variant. Some additional comments were added to highlight how the code relates to the ML-KEM specification in FIPS 203. Reviewed-by: Paul Dale <ppzgs1@gmail.com> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26172)
2024-11-30 22:20:58 +08:00
uint8_t *buf = NULL;
const uint8_t *in;
uint8_t shortbuf[1];
size_t inlen, len = 0, shortlen = sizeof(shortbuf);
Improved import and export - On import, if a seed is provided, the keys are regenerated. - The seed is exported as a separate "seed" parameter, when available. The "ml-kem.retain_seed" parameter is also exported, when false. - The seed is optionally dropped after key generation. * When the "ml-kem.retain_seed" keygen parameter is set to zero. * When the "ml-kem.retain_seed" keygen parameter is not set to 1, and the "ml-kem.retain_seed" provider config property is set explictly false. - The exported private key parameter "priv" is always the FIPS 203 |dk|. - Private key decoding from PKCS#8 produces a transient "seed-only" form of the key, in which "retain_seed" is set to false when the "ml-kem.retain_seed" provider config property is set explictly false. The full key is generated during "load" and the seed is retained or not as specified. - Import honours the "ml-kem.retain_seed" parameter when specified, or otherwise honours the provider's "ml-kem.retain_seed" property. Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26512)
2025-01-18 00:28:51 +08:00
EVP_PKEY *pkey = NULL;
unsigned char *privalloc = NULL;
Multi-variant ML-KEM This introduces support for ML-KEM-512 and ML-KEM-1024 using the same underlying implementation parameterised by a few macros for the associated types and constants. KAT tests are added for ML-KEM 512 and 1024, to complement the previous tests for ML-KEM-768. MLKEM{512,768,1024} TLS "group" codepoints are updated to match the final IANA assigments and to make the additional KEMs known to the TLS layer. The pure-QC MLKEMs are not in the default list of supported groups, and need to be explicitly enabled by the application. Future work will introduce support for hybrids, and for more fine-grained policy of which keyshares a client should send by default, and when a server should request (HRR) a new mutually-supported group that was not sent. Tests for ML-KEM key exchange added to sslapitest to make sure that our TLS client MLKEM{512,768,1024} implementations interoperate with our TLS server, and that MLKEM* are not negotiated in TLS 1.2. Tests also added to excercise non-derandomised ML-KEM APIs, both directly (bypassing the provider layer), and through the generic EVP KEM API (exercising the provider). These make sure that RNG input is used correctly (KAT tests bypass the RNG by specifying seeds). The API interface to the provider takes an "const ML_KEM_VINFO" pointer, (obtained from ossl_ml_kem_get_vinfo()). This checks input and output buffer sizes before passing control to internal code that assumes correctly sized (for each variant) buffers. The original BoringSSL API was refactored to eliminate the opaque public/private key structure wrappers, since these structures are an internal detail between libcrypto and the provider, they are not part of the public (EVP) API. New "clangover" counter-measures added, refined with much appreciated input from David Benjamin (Chromium). The internal steps of "encrypt_cpa" were reordered to reduce the working-set size of the algorithm, now needs space for just two temporary "vectors" rather than three. The "decap" function now process the decrypted message in one call, rather than three separate calls to scalar_decode_1, scalar_decompress and scalar_add. Some loops were unrolled, improving performance of en/decapsulate (pre-expanded vectors and matrix) by around 5%. To handle, however unlikely, the SHA3 primitives not behaving like "pure" functions and failing, the implementation of `decap` was modifed: - To use the KDF to compute the Fujisaki-Okamoto (FO) failure secret first thing, and if that fails, bail out returning an error, a shared secret is still returned at random from the RNG, but it is OK for the caller to not use it. - If any of the subsequently used hash primitives fail, use the computed FO failure secret (OK, despite no longer constant-time) and return success (otherwise the RNG would replace the result). - We quite reasonably assume that chosen-ciphertext attacks (of the correct length) cannot cause hash functions to fail in a manner the depends on the private key content. Support for ML-KEM-512 required adding a centered binomial distribution helper function to deal with η_1 == 3 in just that variant. Some additional comments were added to highlight how the code relates to the ML-KEM specification in FIPS 203. Reviewed-by: Paul Dale <ppzgs1@gmail.com> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26172)
2024-11-30 22:20:58 +08:00
const char *name;
Add a test for the raw private/public key getters Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6394)
2018-06-01 22:06:52 +08:00
/* Check if this algorithm supports public keys */
Trivial shortcuts for EVP_PKEY_eq() Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14942)
2021-04-20 22:39:00 +08:00
if (pub && keys[tst].pub == NULL)
Add a test for the raw private/public key getters Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6394)
2018-06-01 22:06:52 +08:00
return 1;
Multi-variant ML-KEM This introduces support for ML-KEM-512 and ML-KEM-1024 using the same underlying implementation parameterised by a few macros for the associated types and constants. KAT tests are added for ML-KEM 512 and 1024, to complement the previous tests for ML-KEM-768. MLKEM{512,768,1024} TLS "group" codepoints are updated to match the final IANA assigments and to make the additional KEMs known to the TLS layer. The pure-QC MLKEMs are not in the default list of supported groups, and need to be explicitly enabled by the application. Future work will introduce support for hybrids, and for more fine-grained policy of which keyshares a client should send by default, and when a server should request (HRR) a new mutually-supported group that was not sent. Tests for ML-KEM key exchange added to sslapitest to make sure that our TLS client MLKEM{512,768,1024} implementations interoperate with our TLS server, and that MLKEM* are not negotiated in TLS 1.2. Tests also added to excercise non-derandomised ML-KEM APIs, both directly (bypassing the provider layer), and through the generic EVP KEM API (exercising the provider). These make sure that RNG input is used correctly (KAT tests bypass the RNG by specifying seeds). The API interface to the provider takes an "const ML_KEM_VINFO" pointer, (obtained from ossl_ml_kem_get_vinfo()). This checks input and output buffer sizes before passing control to internal code that assumes correctly sized (for each variant) buffers. The original BoringSSL API was refactored to eliminate the opaque public/private key structure wrappers, since these structures are an internal detail between libcrypto and the provider, they are not part of the public (EVP) API. New "clangover" counter-measures added, refined with much appreciated input from David Benjamin (Chromium). The internal steps of "encrypt_cpa" were reordered to reduce the working-set size of the algorithm, now needs space for just two temporary "vectors" rather than three. The "decap" function now process the decrypted message in one call, rather than three separate calls to scalar_decode_1, scalar_decompress and scalar_add. Some loops were unrolled, improving performance of en/decapsulate (pre-expanded vectors and matrix) by around 5%. To handle, however unlikely, the SHA3 primitives not behaving like "pure" functions and failing, the implementation of `decap` was modifed: - To use the KDF to compute the Fujisaki-Okamoto (FO) failure secret first thing, and if that fails, bail out returning an error, a shared secret is still returned at random from the RNG, but it is OK for the caller to not use it. - If any of the subsequently used hash primitives fail, use the computed FO failure secret (OK, despite no longer constant-time) and return success (otherwise the RNG would replace the result). - We quite reasonably assume that chosen-ciphertext attacks (of the correct length) cannot cause hash functions to fail in a manner the depends on the private key content. Support for ML-KEM-512 required adding a centered binomial distribution helper function to deal with η_1 == 3 in just that variant. Some additional comments were added to highlight how the code relates to the ML-KEM specification in FIPS 203. Reviewed-by: Paul Dale <ppzgs1@gmail.com> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26172)
2024-11-30 22:20:58 +08:00
name = keys[tst].name ? keys[tst].name : OBJ_nid2sn(keys[tst].type);
Add a test for the raw private/public key getters Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6394)
2018-06-01 22:06:52 +08:00
if (pub) {
Trivial shortcuts for EVP_PKEY_eq() Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14942)
2021-04-20 22:39:00 +08:00
#ifndef OPENSSL_NO_EC
Multi-variant ML-KEM This introduces support for ML-KEM-512 and ML-KEM-1024 using the same underlying implementation parameterised by a few macros for the associated types and constants. KAT tests are added for ML-KEM 512 and 1024, to complement the previous tests for ML-KEM-768. MLKEM{512,768,1024} TLS "group" codepoints are updated to match the final IANA assigments and to make the additional KEMs known to the TLS layer. The pure-QC MLKEMs are not in the default list of supported groups, and need to be explicitly enabled by the application. Future work will introduce support for hybrids, and for more fine-grained policy of which keyshares a client should send by default, and when a server should request (HRR) a new mutually-supported group that was not sent. Tests for ML-KEM key exchange added to sslapitest to make sure that our TLS client MLKEM{512,768,1024} implementations interoperate with our TLS server, and that MLKEM* are not negotiated in TLS 1.2. Tests also added to excercise non-derandomised ML-KEM APIs, both directly (bypassing the provider layer), and through the generic EVP KEM API (exercising the provider). These make sure that RNG input is used correctly (KAT tests bypass the RNG by specifying seeds). The API interface to the provider takes an "const ML_KEM_VINFO" pointer, (obtained from ossl_ml_kem_get_vinfo()). This checks input and output buffer sizes before passing control to internal code that assumes correctly sized (for each variant) buffers. The original BoringSSL API was refactored to eliminate the opaque public/private key structure wrappers, since these structures are an internal detail between libcrypto and the provider, they are not part of the public (EVP) API. New "clangover" counter-measures added, refined with much appreciated input from David Benjamin (Chromium). The internal steps of "encrypt_cpa" were reordered to reduce the working-set size of the algorithm, now needs space for just two temporary "vectors" rather than three. The "decap" function now process the decrypted message in one call, rather than three separate calls to scalar_decode_1, scalar_decompress and scalar_add. Some loops were unrolled, improving performance of en/decapsulate (pre-expanded vectors and matrix) by around 5%. To handle, however unlikely, the SHA3 primitives not behaving like "pure" functions and failing, the implementation of `decap` was modifed: - To use the KDF to compute the Fujisaki-Okamoto (FO) failure secret first thing, and if that fails, bail out returning an error, a shared secret is still returned at random from the RNG, but it is OK for the caller to not use it. - If any of the subsequently used hash primitives fail, use the computed FO failure secret (OK, despite no longer constant-time) and return success (otherwise the RNG would replace the result). - We quite reasonably assume that chosen-ciphertext attacks (of the correct length) cannot cause hash functions to fail in a manner the depends on the private key content. Support for ML-KEM-512 required adding a centered binomial distribution helper function to deal with η_1 == 3 in just that variant. Some additional comments were added to highlight how the code relates to the ML-KEM specification in FIPS 203. Reviewed-by: Paul Dale <ppzgs1@gmail.com> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26172)
2024-11-30 22:20:58 +08:00
inlen = keys[tst].publen;
in = keys[tst].pub;
if (uselibctx || keys[tst].name != NULL) {
Run the withlibctx.pl script Automatically rename all instances of _with_libctx() to _ex() as per our coding style. Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/12970)
2020-09-24 17:42:23 +08:00
pkey = EVP_PKEY_new_raw_public_key_ex(
Add some tests for the newly added raw private/public key functions We already had soem tests for the older raw private/public key functions so we expand those to call the new versions as well and pass in a libctx. Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/11635)
2020-04-28 00:17:05 +08:00
testctx,
Multi-variant ML-KEM This introduces support for ML-KEM-512 and ML-KEM-1024 using the same underlying implementation parameterised by a few macros for the associated types and constants. KAT tests are added for ML-KEM 512 and 1024, to complement the previous tests for ML-KEM-768. MLKEM{512,768,1024} TLS "group" codepoints are updated to match the final IANA assigments and to make the additional KEMs known to the TLS layer. The pure-QC MLKEMs are not in the default list of supported groups, and need to be explicitly enabled by the application. Future work will introduce support for hybrids, and for more fine-grained policy of which keyshares a client should send by default, and when a server should request (HRR) a new mutually-supported group that was not sent. Tests for ML-KEM key exchange added to sslapitest to make sure that our TLS client MLKEM{512,768,1024} implementations interoperate with our TLS server, and that MLKEM* are not negotiated in TLS 1.2. Tests also added to excercise non-derandomised ML-KEM APIs, both directly (bypassing the provider layer), and through the generic EVP KEM API (exercising the provider). These make sure that RNG input is used correctly (KAT tests bypass the RNG by specifying seeds). The API interface to the provider takes an "const ML_KEM_VINFO" pointer, (obtained from ossl_ml_kem_get_vinfo()). This checks input and output buffer sizes before passing control to internal code that assumes correctly sized (for each variant) buffers. The original BoringSSL API was refactored to eliminate the opaque public/private key structure wrappers, since these structures are an internal detail between libcrypto and the provider, they are not part of the public (EVP) API. New "clangover" counter-measures added, refined with much appreciated input from David Benjamin (Chromium). The internal steps of "encrypt_cpa" were reordered to reduce the working-set size of the algorithm, now needs space for just two temporary "vectors" rather than three. The "decap" function now process the decrypted message in one call, rather than three separate calls to scalar_decode_1, scalar_decompress and scalar_add. Some loops were unrolled, improving performance of en/decapsulate (pre-expanded vectors and matrix) by around 5%. To handle, however unlikely, the SHA3 primitives not behaving like "pure" functions and failing, the implementation of `decap` was modifed: - To use the KDF to compute the Fujisaki-Okamoto (FO) failure secret first thing, and if that fails, bail out returning an error, a shared secret is still returned at random from the RNG, but it is OK for the caller to not use it. - If any of the subsequently used hash primitives fail, use the computed FO failure secret (OK, despite no longer constant-time) and return success (otherwise the RNG would replace the result). - We quite reasonably assume that chosen-ciphertext attacks (of the correct length) cannot cause hash functions to fail in a manner the depends on the private key content. Support for ML-KEM-512 required adding a centered binomial distribution helper function to deal with η_1 == 3 in just that variant. Some additional comments were added to highlight how the code relates to the ML-KEM specification in FIPS 203. Reviewed-by: Paul Dale <ppzgs1@gmail.com> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26172)
2024-11-30 22:20:58 +08:00
name,
Add some tests for the newly added raw private/public key functions We already had soem tests for the older raw private/public key functions so we expand those to call the new versions as well and pass in a libctx. Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/11635)
2020-04-28 00:17:05 +08:00
NULL,
in,
inlen);
} else {
pkey = EVP_PKEY_new_raw_public_key(keys[tst].type,
NULL,
in,
inlen);
}
Trivial shortcuts for EVP_PKEY_eq() Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14942)
2021-04-20 22:39:00 +08:00
#else
return 1;
#endif
Add a test for the raw private/public key getters Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6394)
2018-06-01 22:06:52 +08:00
} else {
Multi-variant ML-KEM This introduces support for ML-KEM-512 and ML-KEM-1024 using the same underlying implementation parameterised by a few macros for the associated types and constants. KAT tests are added for ML-KEM 512 and 1024, to complement the previous tests for ML-KEM-768. MLKEM{512,768,1024} TLS "group" codepoints are updated to match the final IANA assigments and to make the additional KEMs known to the TLS layer. The pure-QC MLKEMs are not in the default list of supported groups, and need to be explicitly enabled by the application. Future work will introduce support for hybrids, and for more fine-grained policy of which keyshares a client should send by default, and when a server should request (HRR) a new mutually-supported group that was not sent. Tests for ML-KEM key exchange added to sslapitest to make sure that our TLS client MLKEM{512,768,1024} implementations interoperate with our TLS server, and that MLKEM* are not negotiated in TLS 1.2. Tests also added to excercise non-derandomised ML-KEM APIs, both directly (bypassing the provider layer), and through the generic EVP KEM API (exercising the provider). These make sure that RNG input is used correctly (KAT tests bypass the RNG by specifying seeds). The API interface to the provider takes an "const ML_KEM_VINFO" pointer, (obtained from ossl_ml_kem_get_vinfo()). This checks input and output buffer sizes before passing control to internal code that assumes correctly sized (for each variant) buffers. The original BoringSSL API was refactored to eliminate the opaque public/private key structure wrappers, since these structures are an internal detail between libcrypto and the provider, they are not part of the public (EVP) API. New "clangover" counter-measures added, refined with much appreciated input from David Benjamin (Chromium). The internal steps of "encrypt_cpa" were reordered to reduce the working-set size of the algorithm, now needs space for just two temporary "vectors" rather than three. The "decap" function now process the decrypted message in one call, rather than three separate calls to scalar_decode_1, scalar_decompress and scalar_add. Some loops were unrolled, improving performance of en/decapsulate (pre-expanded vectors and matrix) by around 5%. To handle, however unlikely, the SHA3 primitives not behaving like "pure" functions and failing, the implementation of `decap` was modifed: - To use the KDF to compute the Fujisaki-Okamoto (FO) failure secret first thing, and if that fails, bail out returning an error, a shared secret is still returned at random from the RNG, but it is OK for the caller to not use it. - If any of the subsequently used hash primitives fail, use the computed FO failure secret (OK, despite no longer constant-time) and return success (otherwise the RNG would replace the result). - We quite reasonably assume that chosen-ciphertext attacks (of the correct length) cannot cause hash functions to fail in a manner the depends on the private key content. Support for ML-KEM-512 required adding a centered binomial distribution helper function to deal with η_1 == 3 in just that variant. Some additional comments were added to highlight how the code relates to the ML-KEM specification in FIPS 203. Reviewed-by: Paul Dale <ppzgs1@gmail.com> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26172)
2024-11-30 22:20:58 +08:00
inlen = keys[tst].privlen;
in = keys[tst].priv;
Improved import and export - On import, if a seed is provided, the keys are regenerated. - The seed is exported as a separate "seed" parameter, when available. The "ml-kem.retain_seed" parameter is also exported, when false. - The seed is optionally dropped after key generation. * When the "ml-kem.retain_seed" keygen parameter is set to zero. * When the "ml-kem.retain_seed" keygen parameter is not set to 1, and the "ml-kem.retain_seed" provider config property is set explictly false. - The exported private key parameter "priv" is always the FIPS 203 |dk|. - Private key decoding from PKCS#8 produces a transient "seed-only" form of the key, in which "retain_seed" is set to false when the "ml-kem.retain_seed" provider config property is set explictly false. The full key is generated during "load" and the seed is retained or not as specified. - Import honours the "ml-kem.retain_seed" parameter when specified, or otherwise honours the provider's "ml-kem.retain_seed" property. Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26512)
2025-01-18 00:28:51 +08:00
#ifndef OPENSSL_NO_ML_KEM
if (in == ml_kem_seed) {
if (!TEST_true(ml_kem_seed_to_priv(name, in, inlen,
&privalloc, &inlen)))
goto done;
in = privalloc;
}
#endif
Multi-variant ML-KEM This introduces support for ML-KEM-512 and ML-KEM-1024 using the same underlying implementation parameterised by a few macros for the associated types and constants. KAT tests are added for ML-KEM 512 and 1024, to complement the previous tests for ML-KEM-768. MLKEM{512,768,1024} TLS "group" codepoints are updated to match the final IANA assigments and to make the additional KEMs known to the TLS layer. The pure-QC MLKEMs are not in the default list of supported groups, and need to be explicitly enabled by the application. Future work will introduce support for hybrids, and for more fine-grained policy of which keyshares a client should send by default, and when a server should request (HRR) a new mutually-supported group that was not sent. Tests for ML-KEM key exchange added to sslapitest to make sure that our TLS client MLKEM{512,768,1024} implementations interoperate with our TLS server, and that MLKEM* are not negotiated in TLS 1.2. Tests also added to excercise non-derandomised ML-KEM APIs, both directly (bypassing the provider layer), and through the generic EVP KEM API (exercising the provider). These make sure that RNG input is used correctly (KAT tests bypass the RNG by specifying seeds). The API interface to the provider takes an "const ML_KEM_VINFO" pointer, (obtained from ossl_ml_kem_get_vinfo()). This checks input and output buffer sizes before passing control to internal code that assumes correctly sized (for each variant) buffers. The original BoringSSL API was refactored to eliminate the opaque public/private key structure wrappers, since these structures are an internal detail between libcrypto and the provider, they are not part of the public (EVP) API. New "clangover" counter-measures added, refined with much appreciated input from David Benjamin (Chromium). The internal steps of "encrypt_cpa" were reordered to reduce the working-set size of the algorithm, now needs space for just two temporary "vectors" rather than three. The "decap" function now process the decrypted message in one call, rather than three separate calls to scalar_decode_1, scalar_decompress and scalar_add. Some loops were unrolled, improving performance of en/decapsulate (pre-expanded vectors and matrix) by around 5%. To handle, however unlikely, the SHA3 primitives not behaving like "pure" functions and failing, the implementation of `decap` was modifed: - To use the KDF to compute the Fujisaki-Okamoto (FO) failure secret first thing, and if that fails, bail out returning an error, a shared secret is still returned at random from the RNG, but it is OK for the caller to not use it. - If any of the subsequently used hash primitives fail, use the computed FO failure secret (OK, despite no longer constant-time) and return success (otherwise the RNG would replace the result). - We quite reasonably assume that chosen-ciphertext attacks (of the correct length) cannot cause hash functions to fail in a manner the depends on the private key content. Support for ML-KEM-512 required adding a centered binomial distribution helper function to deal with η_1 == 3 in just that variant. Some additional comments were added to highlight how the code relates to the ML-KEM specification in FIPS 203. Reviewed-by: Paul Dale <ppzgs1@gmail.com> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26172)
2024-11-30 22:20:58 +08:00
if (uselibctx || keys[tst].name != NULL) {
Run the withlibctx.pl script Automatically rename all instances of _with_libctx() to _ex() as per our coding style. Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/12970)
2020-09-24 17:42:23 +08:00
pkey = EVP_PKEY_new_raw_private_key_ex(
Multi-variant ML-KEM This introduces support for ML-KEM-512 and ML-KEM-1024 using the same underlying implementation parameterised by a few macros for the associated types and constants. KAT tests are added for ML-KEM 512 and 1024, to complement the previous tests for ML-KEM-768. MLKEM{512,768,1024} TLS "group" codepoints are updated to match the final IANA assigments and to make the additional KEMs known to the TLS layer. The pure-QC MLKEMs are not in the default list of supported groups, and need to be explicitly enabled by the application. Future work will introduce support for hybrids, and for more fine-grained policy of which keyshares a client should send by default, and when a server should request (HRR) a new mutually-supported group that was not sent. Tests for ML-KEM key exchange added to sslapitest to make sure that our TLS client MLKEM{512,768,1024} implementations interoperate with our TLS server, and that MLKEM* are not negotiated in TLS 1.2. Tests also added to excercise non-derandomised ML-KEM APIs, both directly (bypassing the provider layer), and through the generic EVP KEM API (exercising the provider). These make sure that RNG input is used correctly (KAT tests bypass the RNG by specifying seeds). The API interface to the provider takes an "const ML_KEM_VINFO" pointer, (obtained from ossl_ml_kem_get_vinfo()). This checks input and output buffer sizes before passing control to internal code that assumes correctly sized (for each variant) buffers. The original BoringSSL API was refactored to eliminate the opaque public/private key structure wrappers, since these structures are an internal detail between libcrypto and the provider, they are not part of the public (EVP) API. New "clangover" counter-measures added, refined with much appreciated input from David Benjamin (Chromium). The internal steps of "encrypt_cpa" were reordered to reduce the working-set size of the algorithm, now needs space for just two temporary "vectors" rather than three. The "decap" function now process the decrypted message in one call, rather than three separate calls to scalar_decode_1, scalar_decompress and scalar_add. Some loops were unrolled, improving performance of en/decapsulate (pre-expanded vectors and matrix) by around 5%. To handle, however unlikely, the SHA3 primitives not behaving like "pure" functions and failing, the implementation of `decap` was modifed: - To use the KDF to compute the Fujisaki-Okamoto (FO) failure secret first thing, and if that fails, bail out returning an error, a shared secret is still returned at random from the RNG, but it is OK for the caller to not use it. - If any of the subsequently used hash primitives fail, use the computed FO failure secret (OK, despite no longer constant-time) and return success (otherwise the RNG would replace the result). - We quite reasonably assume that chosen-ciphertext attacks (of the correct length) cannot cause hash functions to fail in a manner the depends on the private key content. Support for ML-KEM-512 required adding a centered binomial distribution helper function to deal with η_1 == 3 in just that variant. Some additional comments were added to highlight how the code relates to the ML-KEM specification in FIPS 203. Reviewed-by: Paul Dale <ppzgs1@gmail.com> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26172)
2024-11-30 22:20:58 +08:00
testctx,
name,
Add some tests for the newly added raw private/public key functions We already had soem tests for the older raw private/public key functions so we expand those to call the new versions as well and pass in a libctx. Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/11635)
2020-04-28 00:17:05 +08:00
NULL,
in,
inlen);
} else {
pkey = EVP_PKEY_new_raw_private_key(keys[tst].type,
NULL,
in,
inlen);
}
Add a test for the raw private/public key getters Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6394)
2018-06-01 22:06:52 +08:00
}
if (!TEST_ptr(pkey)
Trivial shortcuts for EVP_PKEY_eq() Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14942)
2021-04-20 22:39:00 +08:00
|| !TEST_int_eq(EVP_PKEY_eq(pkey, pkey), 1)
Add a test for the raw private/public key getters Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6394)
2018-06-01 22:06:52 +08:00
|| (!pub && !TEST_true(EVP_PKEY_get_raw_private_key(pkey, NULL, &len)))
|| (pub && !TEST_true(EVP_PKEY_get_raw_public_key(pkey, NULL, &len)))
Test short buffers Test that calling EVP_DigestSign(), EVP_DigestSignFinal(), EVP_PKEY_sign(), EVP_PKEY_get_raw_private_key(), or EVP_PKEY_get_raw_public_key() with a short output buffer results in a failure. Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16789)
2021-10-07 21:15:47 +08:00
|| !TEST_true(len == inlen))
goto done;
if (tst != 1) {
/*
* Test that supplying a buffer that is too small fails. Doesn't apply
* to HMAC with a zero length key
*/
Multi-variant ML-KEM This introduces support for ML-KEM-512 and ML-KEM-1024 using the same underlying implementation parameterised by a few macros for the associated types and constants. KAT tests are added for ML-KEM 512 and 1024, to complement the previous tests for ML-KEM-768. MLKEM{512,768,1024} TLS "group" codepoints are updated to match the final IANA assigments and to make the additional KEMs known to the TLS layer. The pure-QC MLKEMs are not in the default list of supported groups, and need to be explicitly enabled by the application. Future work will introduce support for hybrids, and for more fine-grained policy of which keyshares a client should send by default, and when a server should request (HRR) a new mutually-supported group that was not sent. Tests for ML-KEM key exchange added to sslapitest to make sure that our TLS client MLKEM{512,768,1024} implementations interoperate with our TLS server, and that MLKEM* are not negotiated in TLS 1.2. Tests also added to excercise non-derandomised ML-KEM APIs, both directly (bypassing the provider layer), and through the generic EVP KEM API (exercising the provider). These make sure that RNG input is used correctly (KAT tests bypass the RNG by specifying seeds). The API interface to the provider takes an "const ML_KEM_VINFO" pointer, (obtained from ossl_ml_kem_get_vinfo()). This checks input and output buffer sizes before passing control to internal code that assumes correctly sized (for each variant) buffers. The original BoringSSL API was refactored to eliminate the opaque public/private key structure wrappers, since these structures are an internal detail between libcrypto and the provider, they are not part of the public (EVP) API. New "clangover" counter-measures added, refined with much appreciated input from David Benjamin (Chromium). The internal steps of "encrypt_cpa" were reordered to reduce the working-set size of the algorithm, now needs space for just two temporary "vectors" rather than three. The "decap" function now process the decrypted message in one call, rather than three separate calls to scalar_decode_1, scalar_decompress and scalar_add. Some loops were unrolled, improving performance of en/decapsulate (pre-expanded vectors and matrix) by around 5%. To handle, however unlikely, the SHA3 primitives not behaving like "pure" functions and failing, the implementation of `decap` was modifed: - To use the KDF to compute the Fujisaki-Okamoto (FO) failure secret first thing, and if that fails, bail out returning an error, a shared secret is still returned at random from the RNG, but it is OK for the caller to not use it. - If any of the subsequently used hash primitives fail, use the computed FO failure secret (OK, despite no longer constant-time) and return success (otherwise the RNG would replace the result). - We quite reasonably assume that chosen-ciphertext attacks (of the correct length) cannot cause hash functions to fail in a manner the depends on the private key content. Support for ML-KEM-512 required adding a centered binomial distribution helper function to deal with η_1 == 3 in just that variant. Some additional comments were added to highlight how the code relates to the ML-KEM specification in FIPS 203. Reviewed-by: Paul Dale <ppzgs1@gmail.com> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26172)
2024-11-30 22:20:58 +08:00
if ((!pub && !TEST_false(EVP_PKEY_get_raw_private_key(pkey, shortbuf,
Test short buffers Test that calling EVP_DigestSign(), EVP_DigestSignFinal(), EVP_PKEY_sign(), EVP_PKEY_get_raw_private_key(), or EVP_PKEY_get_raw_public_key() with a short output buffer results in a failure. Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16789)
2021-10-07 21:15:47 +08:00
&shortlen)))
Multi-variant ML-KEM This introduces support for ML-KEM-512 and ML-KEM-1024 using the same underlying implementation parameterised by a few macros for the associated types and constants. KAT tests are added for ML-KEM 512 and 1024, to complement the previous tests for ML-KEM-768. MLKEM{512,768,1024} TLS "group" codepoints are updated to match the final IANA assigments and to make the additional KEMs known to the TLS layer. The pure-QC MLKEMs are not in the default list of supported groups, and need to be explicitly enabled by the application. Future work will introduce support for hybrids, and for more fine-grained policy of which keyshares a client should send by default, and when a server should request (HRR) a new mutually-supported group that was not sent. Tests for ML-KEM key exchange added to sslapitest to make sure that our TLS client MLKEM{512,768,1024} implementations interoperate with our TLS server, and that MLKEM* are not negotiated in TLS 1.2. Tests also added to excercise non-derandomised ML-KEM APIs, both directly (bypassing the provider layer), and through the generic EVP KEM API (exercising the provider). These make sure that RNG input is used correctly (KAT tests bypass the RNG by specifying seeds). The API interface to the provider takes an "const ML_KEM_VINFO" pointer, (obtained from ossl_ml_kem_get_vinfo()). This checks input and output buffer sizes before passing control to internal code that assumes correctly sized (for each variant) buffers. The original BoringSSL API was refactored to eliminate the opaque public/private key structure wrappers, since these structures are an internal detail between libcrypto and the provider, they are not part of the public (EVP) API. New "clangover" counter-measures added, refined with much appreciated input from David Benjamin (Chromium). The internal steps of "encrypt_cpa" were reordered to reduce the working-set size of the algorithm, now needs space for just two temporary "vectors" rather than three. The "decap" function now process the decrypted message in one call, rather than three separate calls to scalar_decode_1, scalar_decompress and scalar_add. Some loops were unrolled, improving performance of en/decapsulate (pre-expanded vectors and matrix) by around 5%. To handle, however unlikely, the SHA3 primitives not behaving like "pure" functions and failing, the implementation of `decap` was modifed: - To use the KDF to compute the Fujisaki-Okamoto (FO) failure secret first thing, and if that fails, bail out returning an error, a shared secret is still returned at random from the RNG, but it is OK for the caller to not use it. - If any of the subsequently used hash primitives fail, use the computed FO failure secret (OK, despite no longer constant-time) and return success (otherwise the RNG would replace the result). - We quite reasonably assume that chosen-ciphertext attacks (of the correct length) cannot cause hash functions to fail in a manner the depends on the private key content. Support for ML-KEM-512 required adding a centered binomial distribution helper function to deal with η_1 == 3 in just that variant. Some additional comments were added to highlight how the code relates to the ML-KEM specification in FIPS 203. Reviewed-by: Paul Dale <ppzgs1@gmail.com> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26172)
2024-11-30 22:20:58 +08:00
|| (pub && !TEST_false(EVP_PKEY_get_raw_public_key(pkey, shortbuf,
Test short buffers Test that calling EVP_DigestSign(), EVP_DigestSignFinal(), EVP_PKEY_sign(), EVP_PKEY_get_raw_private_key(), or EVP_PKEY_get_raw_public_key() with a short output buffer results in a failure. Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16789)
2021-10-07 21:15:47 +08:00
&shortlen))))
goto done;
}
Multi-variant ML-KEM This introduces support for ML-KEM-512 and ML-KEM-1024 using the same underlying implementation parameterised by a few macros for the associated types and constants. KAT tests are added for ML-KEM 512 and 1024, to complement the previous tests for ML-KEM-768. MLKEM{512,768,1024} TLS "group" codepoints are updated to match the final IANA assigments and to make the additional KEMs known to the TLS layer. The pure-QC MLKEMs are not in the default list of supported groups, and need to be explicitly enabled by the application. Future work will introduce support for hybrids, and for more fine-grained policy of which keyshares a client should send by default, and when a server should request (HRR) a new mutually-supported group that was not sent. Tests for ML-KEM key exchange added to sslapitest to make sure that our TLS client MLKEM{512,768,1024} implementations interoperate with our TLS server, and that MLKEM* are not negotiated in TLS 1.2. Tests also added to excercise non-derandomised ML-KEM APIs, both directly (bypassing the provider layer), and through the generic EVP KEM API (exercising the provider). These make sure that RNG input is used correctly (KAT tests bypass the RNG by specifying seeds). The API interface to the provider takes an "const ML_KEM_VINFO" pointer, (obtained from ossl_ml_kem_get_vinfo()). This checks input and output buffer sizes before passing control to internal code that assumes correctly sized (for each variant) buffers. The original BoringSSL API was refactored to eliminate the opaque public/private key structure wrappers, since these structures are an internal detail between libcrypto and the provider, they are not part of the public (EVP) API. New "clangover" counter-measures added, refined with much appreciated input from David Benjamin (Chromium). The internal steps of "encrypt_cpa" were reordered to reduce the working-set size of the algorithm, now needs space for just two temporary "vectors" rather than three. The "decap" function now process the decrypted message in one call, rather than three separate calls to scalar_decode_1, scalar_decompress and scalar_add. Some loops were unrolled, improving performance of en/decapsulate (pre-expanded vectors and matrix) by around 5%. To handle, however unlikely, the SHA3 primitives not behaving like "pure" functions and failing, the implementation of `decap` was modifed: - To use the KDF to compute the Fujisaki-Okamoto (FO) failure secret first thing, and if that fails, bail out returning an error, a shared secret is still returned at random from the RNG, but it is OK for the caller to not use it. - If any of the subsequently used hash primitives fail, use the computed FO failure secret (OK, despite no longer constant-time) and return success (otherwise the RNG would replace the result). - We quite reasonably assume that chosen-ciphertext attacks (of the correct length) cannot cause hash functions to fail in a manner the depends on the private key content. Support for ML-KEM-512 required adding a centered binomial distribution helper function to deal with η_1 == 3 in just that variant. Some additional comments were added to highlight how the code relates to the ML-KEM specification in FIPS 203. Reviewed-by: Paul Dale <ppzgs1@gmail.com> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26172)
2024-11-30 22:20:58 +08:00
if (!TEST_ptr(buf = OPENSSL_zalloc(len <= 80 ? 80 : len)))
goto done;
Test short buffers Test that calling EVP_DigestSign(), EVP_DigestSignFinal(), EVP_PKEY_sign(), EVP_PKEY_get_raw_private_key(), or EVP_PKEY_get_raw_public_key() with a short output buffer results in a failure. Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16789)
2021-10-07 21:15:47 +08:00
if ((!pub && !TEST_true(EVP_PKEY_get_raw_private_key(pkey, buf, &len)))
Add a test for the raw private/public key getters Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6394)
2018-06-01 22:06:52 +08:00
|| (pub && !TEST_true(EVP_PKEY_get_raw_public_key(pkey, buf, &len)))
|| !TEST_mem_eq(in, inlen, buf, len))
goto done;
ret = 1;
done:
Improved import and export - On import, if a seed is provided, the keys are regenerated. - The seed is exported as a separate "seed" parameter, when available. The "ml-kem.retain_seed" parameter is also exported, when false. - The seed is optionally dropped after key generation. * When the "ml-kem.retain_seed" keygen parameter is set to zero. * When the "ml-kem.retain_seed" keygen parameter is not set to 1, and the "ml-kem.retain_seed" provider config property is set explictly false. - The exported private key parameter "priv" is always the FIPS 203 |dk|. - Private key decoding from PKCS#8 produces a transient "seed-only" form of the key, in which "retain_seed" is set to false when the "ml-kem.retain_seed" provider config property is set explictly false. The full key is generated during "load" and the seed is retained or not as specified. - Import honours the "ml-kem.retain_seed" parameter when specified, or otherwise honours the provider's "ml-kem.retain_seed" property. Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26512)
2025-01-18 00:28:51 +08:00
OPENSSL_free(privalloc);
Multi-variant ML-KEM This introduces support for ML-KEM-512 and ML-KEM-1024 using the same underlying implementation parameterised by a few macros for the associated types and constants. KAT tests are added for ML-KEM 512 and 1024, to complement the previous tests for ML-KEM-768. MLKEM{512,768,1024} TLS "group" codepoints are updated to match the final IANA assigments and to make the additional KEMs known to the TLS layer. The pure-QC MLKEMs are not in the default list of supported groups, and need to be explicitly enabled by the application. Future work will introduce support for hybrids, and for more fine-grained policy of which keyshares a client should send by default, and when a server should request (HRR) a new mutually-supported group that was not sent. Tests for ML-KEM key exchange added to sslapitest to make sure that our TLS client MLKEM{512,768,1024} implementations interoperate with our TLS server, and that MLKEM* are not negotiated in TLS 1.2. Tests also added to excercise non-derandomised ML-KEM APIs, both directly (bypassing the provider layer), and through the generic EVP KEM API (exercising the provider). These make sure that RNG input is used correctly (KAT tests bypass the RNG by specifying seeds). The API interface to the provider takes an "const ML_KEM_VINFO" pointer, (obtained from ossl_ml_kem_get_vinfo()). This checks input and output buffer sizes before passing control to internal code that assumes correctly sized (for each variant) buffers. The original BoringSSL API was refactored to eliminate the opaque public/private key structure wrappers, since these structures are an internal detail between libcrypto and the provider, they are not part of the public (EVP) API. New "clangover" counter-measures added, refined with much appreciated input from David Benjamin (Chromium). The internal steps of "encrypt_cpa" were reordered to reduce the working-set size of the algorithm, now needs space for just two temporary "vectors" rather than three. The "decap" function now process the decrypted message in one call, rather than three separate calls to scalar_decode_1, scalar_decompress and scalar_add. Some loops were unrolled, improving performance of en/decapsulate (pre-expanded vectors and matrix) by around 5%. To handle, however unlikely, the SHA3 primitives not behaving like "pure" functions and failing, the implementation of `decap` was modifed: - To use the KDF to compute the Fujisaki-Okamoto (FO) failure secret first thing, and if that fails, bail out returning an error, a shared secret is still returned at random from the RNG, but it is OK for the caller to not use it. - If any of the subsequently used hash primitives fail, use the computed FO failure secret (OK, despite no longer constant-time) and return success (otherwise the RNG would replace the result). - We quite reasonably assume that chosen-ciphertext attacks (of the correct length) cannot cause hash functions to fail in a manner the depends on the private key content. Support for ML-KEM-512 required adding a centered binomial distribution helper function to deal with η_1 == 3 in just that variant. Some additional comments were added to highlight how the code relates to the ML-KEM specification in FIPS 203. Reviewed-by: Paul Dale <ppzgs1@gmail.com> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26172)
2024-11-30 22:20:58 +08:00
OPENSSL_free(buf);
Add a test for the raw private/public key getters Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6394)
2018-06-01 22:06:52 +08:00
EVP_PKEY_free(pkey);
return ret;
}
static int test_set_get_raw_keys(int tst)
{
Trivial shortcuts for EVP_PKEY_eq() Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14942)
2021-04-20 22:39:00 +08:00
return (nullprov != NULL || test_set_get_raw_keys_int(tst, 0, 0))
Add some tests for the newly added raw private/public key functions We already had soem tests for the older raw private/public key functions so we expand those to call the new versions as well and pass in a libctx. Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/11635)
2020-04-28 00:17:05 +08:00
&& test_set_get_raw_keys_int(tst, 0, 1)
Trivial shortcuts for EVP_PKEY_eq() Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14942)
2021-04-20 22:39:00 +08:00
&& (nullprov != NULL || test_set_get_raw_keys_int(tst, 1, 0))
Add some tests for the newly added raw private/public key functions We already had soem tests for the older raw private/public key functions so we expand those to call the new versions as well and pass in a libctx. Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/11635)
2020-04-28 00:17:05 +08:00
&& test_set_get_raw_keys_int(tst, 1, 1);
Add a test for the raw private/public key getters Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6394)
2018-06-01 22:06:52 +08:00
}
EVP: deprecate the EVP_X_meth_ functions. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/11082)
2020-02-13 09:00:57 +08:00
#ifndef OPENSSL_NO_DEPRECATED_3_0
Support key check in EVP interface A new method is added to EVP_PKEY_METH as: int (*check) (EVP_PKEY_CTX *ctx); and to EVP_PKEY_ASN1_METHOD as: int (*pkey_check) (EVP_PKEY_CTX *ctx); This is used to check the validity of a specific key. The order of calls is: EVP_PKEY_check -> pmeth.check -> ameth.pkey_check. Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/4337)
2017-09-04 22:02:59 +08:00
static int pkey_custom_check(EVP_PKEY *pkey)
{
return 0xbeef;
}
Support public key and param check in EVP interface EVP_PKEY_public_check() and EVP_PKEY_param_check() Doc and test cases are added Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/4647)
2017-11-01 00:45:24 +08:00
static int pkey_custom_pub_check(EVP_PKEY *pkey)
{
return 0xbeef;
}
static int pkey_custom_param_check(EVP_PKEY *pkey)
{
return 0xbeef;
}
Support key check in EVP interface A new method is added to EVP_PKEY_METH as: int (*check) (EVP_PKEY_CTX *ctx); and to EVP_PKEY_ASN1_METHOD as: int (*pkey_check) (EVP_PKEY_CTX *ctx); This is used to check the validity of a specific key. The order of calls is: EVP_PKEY_check -> pmeth.check -> ameth.pkey_check. Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/4337)
2017-09-04 22:02:59 +08:00
static EVP_PKEY_METHOD *custom_pmeth;
EVP: deprecate the EVP_X_meth_ functions. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/11082)
2020-02-13 09:00:57 +08:00
#endif
Support key check in EVP interface A new method is added to EVP_PKEY_METH as: int (*check) (EVP_PKEY_CTX *ctx); and to EVP_PKEY_ASN1_METHOD as: int (*pkey_check) (EVP_PKEY_CTX *ctx); This is used to check the validity of a specific key. The order of calls is: EVP_PKEY_check -> pmeth.check -> ameth.pkey_check. Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/4337)
2017-09-04 22:02:59 +08:00
static int test_EVP_PKEY_check(int i)
{
int ret = 0;
EVP_PKEY *pkey = NULL;
EVP_PKEY_CTX *ctx = NULL;
EVP: deprecate the EVP_X_meth_ functions. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/11082)
2020-02-13 09:00:57 +08:00
#ifndef OPENSSL_NO_DEPRECATED_3_0
Support key check in EVP interface A new method is added to EVP_PKEY_METH as: int (*check) (EVP_PKEY_CTX *ctx); and to EVP_PKEY_ASN1_METHOD as: int (*pkey_check) (EVP_PKEY_CTX *ctx); This is used to check the validity of a specific key. The order of calls is: EVP_PKEY_check -> pmeth.check -> ameth.pkey_check. Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/4337)
2017-09-04 22:02:59 +08:00
EVP_PKEY_CTX *ctx2 = NULL;
EVP: deprecate the EVP_X_meth_ functions. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/11082)
2020-02-13 09:00:57 +08:00
#endif
Support key check in EVP interface A new method is added to EVP_PKEY_METH as: int (*check) (EVP_PKEY_CTX *ctx); and to EVP_PKEY_ASN1_METHOD as: int (*pkey_check) (EVP_PKEY_CTX *ctx); This is used to check the validity of a specific key. The order of calls is: EVP_PKEY_check -> pmeth.check -> ameth.pkey_check. Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/4337)
2017-09-04 22:02:59 +08:00
const APK_DATA *ak = &keycheckdata[i];
const unsigned char *input = ak->kder;
size_t input_len = ak->size;
int expected_id = ak->evptype;
int expected_check = ak->check;
Support public key and param check in EVP interface EVP_PKEY_public_check() and EVP_PKEY_param_check() Doc and test cases are added Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/4647)
2017-11-01 00:45:24 +08:00
int expected_pub_check = ak->pub_check;
int expected_param_check = ak->param_check;
int type = ak->type;
Deprecate EC_KEY + Update ec apps to use EVP_PKEY Co-author: Richard Levitte <levitte@openssl.org> Co-author: Tomas Mraz <tmraz@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13139)
2020-10-15 11:41:59 +08:00
if (!TEST_ptr(pkey = load_example_key(ak->keytype, input, input_len)))
goto done;
if (type == 0
Rename all getters to use get/get0 in name For functions that exist in 1.1.1 provide a simple aliases via #define. Fixes #15236 Functions with OSSL_DECODER_, OSSL_ENCODER_, OSSL_STORE_LOADER_, EVP_KEYEXCH_, EVP_KEM_, EVP_ASYM_CIPHER_, EVP_SIGNATURE_, EVP_KEYMGMT_, EVP_RAND_, EVP_MAC_, EVP_KDF_, EVP_PKEY_, EVP_MD_, and EVP_CIPHER_ prefixes are renamed. Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/15405)
2021-05-21 22:58:08 +08:00
&& !TEST_int_eq(EVP_PKEY_get_id(pkey), expected_id))
Deprecate EC_KEY + Update ec apps to use EVP_PKEY Co-author: Richard Levitte <levitte@openssl.org> Co-author: Tomas Mraz <tmraz@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13139)
2020-10-15 11:41:59 +08:00
goto done;
Support key check in EVP interface A new method is added to EVP_PKEY_METH as: int (*check) (EVP_PKEY_CTX *ctx); and to EVP_PKEY_ASN1_METHOD as: int (*pkey_check) (EVP_PKEY_CTX *ctx); This is used to check the validity of a specific key. The order of calls is: EVP_PKEY_check -> pmeth.check -> ameth.pkey_check. Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/4337)
2017-09-04 22:02:59 +08:00
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
if (!TEST_ptr(ctx = EVP_PKEY_CTX_new_from_pkey(testctx, pkey, testpropq)))
Support key check in EVP interface A new method is added to EVP_PKEY_METH as: int (*check) (EVP_PKEY_CTX *ctx); and to EVP_PKEY_ASN1_METHOD as: int (*pkey_check) (EVP_PKEY_CTX *ctx); This is used to check the validity of a specific key. The order of calls is: EVP_PKEY_check -> pmeth.check -> ameth.pkey_check. Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/4337)
2017-09-04 22:02:59 +08:00
goto done;
Support public key and param check in EVP interface EVP_PKEY_public_check() and EVP_PKEY_param_check() Doc and test cases are added Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/4647)
2017-11-01 00:45:24 +08:00
if (!TEST_int_eq(EVP_PKEY_check(ctx), expected_check))
goto done;
if (!TEST_int_eq(EVP_PKEY_public_check(ctx), expected_pub_check))
goto done;
if (!TEST_int_eq(EVP_PKEY_param_check(ctx), expected_param_check))
goto done;
Support key check in EVP interface A new method is added to EVP_PKEY_METH as: int (*check) (EVP_PKEY_CTX *ctx); and to EVP_PKEY_ASN1_METHOD as: int (*pkey_check) (EVP_PKEY_CTX *ctx); This is used to check the validity of a specific key. The order of calls is: EVP_PKEY_check -> pmeth.check -> ameth.pkey_check. Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/4337)
2017-09-04 22:02:59 +08:00
EVP: deprecate the EVP_X_meth_ functions. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/11082)
2020-02-13 09:00:57 +08:00
#ifndef OPENSSL_NO_DEPRECATED_3_0
Support key check in EVP interface A new method is added to EVP_PKEY_METH as: int (*check) (EVP_PKEY_CTX *ctx); and to EVP_PKEY_ASN1_METHOD as: int (*pkey_check) (EVP_PKEY_CTX *ctx); This is used to check the validity of a specific key. The order of calls is: EVP_PKEY_check -> pmeth.check -> ameth.pkey_check. Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/4337)
2017-09-04 22:02:59 +08:00
ctx2 = EVP_PKEY_CTX_new_id(0xdefaced, NULL);
/* assign the pkey directly, as an internal test */
EVP_PKEY_up_ref(pkey);
ctx2->pkey = pkey;
if (!TEST_int_eq(EVP_PKEY_check(ctx2), 0xbeef))
goto done;
Support public key and param check in EVP interface EVP_PKEY_public_check() and EVP_PKEY_param_check() Doc and test cases are added Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/4647)
2017-11-01 00:45:24 +08:00
if (!TEST_int_eq(EVP_PKEY_public_check(ctx2), 0xbeef))
goto done;
if (!TEST_int_eq(EVP_PKEY_param_check(ctx2), 0xbeef))
goto done;
EVP: deprecate the EVP_X_meth_ functions. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/11082)
2020-02-13 09:00:57 +08:00
#endif
Support public key and param check in EVP interface EVP_PKEY_public_check() and EVP_PKEY_param_check() Doc and test cases are added Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/4647)
2017-11-01 00:45:24 +08:00
Support key check in EVP interface A new method is added to EVP_PKEY_METH as: int (*check) (EVP_PKEY_CTX *ctx); and to EVP_PKEY_ASN1_METHOD as: int (*pkey_check) (EVP_PKEY_CTX *ctx); This is used to check the validity of a specific key. The order of calls is: EVP_PKEY_check -> pmeth.check -> ameth.pkey_check. Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/4337)
2017-09-04 22:02:59 +08:00
ret = 1;
done:
EVP_PKEY_CTX_free(ctx);
EVP: deprecate the EVP_X_meth_ functions. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/11082)
2020-02-13 09:00:57 +08:00
#ifndef OPENSSL_NO_DEPRECATED_3_0
Support key check in EVP interface A new method is added to EVP_PKEY_METH as: int (*check) (EVP_PKEY_CTX *ctx); and to EVP_PKEY_ASN1_METHOD as: int (*pkey_check) (EVP_PKEY_CTX *ctx); This is used to check the validity of a specific key. The order of calls is: EVP_PKEY_check -> pmeth.check -> ameth.pkey_check. Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/4337)
2017-09-04 22:02:59 +08:00
EVP_PKEY_CTX_free(ctx2);
EVP: deprecate the EVP_X_meth_ functions. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/11082)
2020-02-13 09:00:57 +08:00
#endif
Support key check in EVP interface A new method is added to EVP_PKEY_METH as: int (*check) (EVP_PKEY_CTX *ctx); and to EVP_PKEY_ASN1_METHOD as: int (*pkey_check) (EVP_PKEY_CTX *ctx); This is used to check the validity of a specific key. The order of calls is: EVP_PKEY_check -> pmeth.check -> ameth.pkey_check. Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/4337)
2017-09-04 22:02:59 +08:00
EVP_PKEY_free(pkey);
return ret;
}
test/evp_extra_test.c: Add OPENSSL_NO_CMAC around CMAC test Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/11833)
2020-05-14 23:15:05 +08:00
#ifndef OPENSSL_NO_CMAC
Extend test_CMAC_keygen in evp_extra_test The test only setup the generation of a key. It did not complete that generation. We extend the test to complete the generation and to use the generated key. We also compare the result with a key generated in a different way. Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/12637)
2020-08-12 21:41:12 +08:00
static int get_cmac_val(EVP_PKEY *pkey, unsigned char *mac)
{
EVP_MD_CTX *mdctx = EVP_MD_CTX_new();
const char msg[] = "Hello World";
Fix test_CMAC_keygen Make sure we correctly pass through the size of the buffer to EVP_DigestSignFinal Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16789)
2021-10-11 20:12:49 +08:00
size_t maclen = AES_BLOCK_SIZE;
Extend test_CMAC_keygen in evp_extra_test The test only setup the generation of a key. It did not complete that generation. We extend the test to complete the generation and to use the generated key. We also compare the result with a key generated in a different way. Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/12637)
2020-08-12 21:41:12 +08:00
int ret = 1;
if (!TEST_ptr(mdctx)
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
|| !TEST_true(EVP_DigestSignInit_ex(mdctx, NULL, NULL, testctx,
testpropq, pkey, NULL))
Extend test_CMAC_keygen in evp_extra_test The test only setup the generation of a key. It did not complete that generation. We extend the test to complete the generation and to use the generated key. We also compare the result with a key generated in a different way. Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/12637)
2020-08-12 21:41:12 +08:00
|| !TEST_true(EVP_DigestSignUpdate(mdctx, msg, sizeof(msg)))
|| !TEST_true(EVP_DigestSignFinal(mdctx, mac, &maclen))
|| !TEST_size_t_eq(maclen, AES_BLOCK_SIZE))
ret = 0;
EVP_MD_CTX_free(mdctx);
return ret;
}
test/evp_extra_test.c: Add test for CMAC keygen with a NULL engine Verifies that #11671 is fixed Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/11674)
2020-05-12 21:27:32 +08:00
static int test_CMAC_keygen(void)
{
Extend test_CMAC_keygen in evp_extra_test The test only setup the generation of a key. It did not complete that generation. We extend the test to complete the generation and to use the generated key. We also compare the result with a key generated in a different way. Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/12637)
2020-08-12 21:41:12 +08:00
static unsigned char key[] = {
0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08, 0x09, 0x0a, 0x0b,
0x0c, 0x0d, 0x0e, 0x0f, 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17,
0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f
};
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
EVP_PKEY_CTX *kctx = NULL;
test/evp_extra_test.c: Add test for CMAC keygen with a NULL engine Verifies that #11671 is fixed Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/11674)
2020-05-12 21:27:32 +08:00
int ret = 0;
Extend test_CMAC_keygen in evp_extra_test The test only setup the generation of a key. It did not complete that generation. We extend the test to complete the generation and to use the generated key. We also compare the result with a key generated in a different way. Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/12637)
2020-08-12 21:41:12 +08:00
EVP_PKEY *pkey = NULL;
Deprecate EVP_KEY_new_CMAC_key and EVP_PKEY_new_CMAC_key_ex EVP_KEY_new_CMAC_key_ex was in the pre-release 3.0 only, so is safe to remove. Restore 1.1.1 version of EVP_PKEY_new_CMAC_key documentation. Also make testing of EVP_PKEY_new_CMAC_key properly #ifdef'd. Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13829)
2020-12-08 23:13:54 +08:00
unsigned char mac[AES_BLOCK_SIZE];
# if !defined(OPENSSL_NO_DEPRECATED_3_0)
unsigned char mac2[AES_BLOCK_SIZE];
# endif
Extend test_CMAC_keygen in evp_extra_test The test only setup the generation of a key. It did not complete that generation. We extend the test to complete the generation and to use the generated key. We also compare the result with a key generated in a different way. Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/12637)
2020-08-12 21:41:12 +08:00
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
if (nullprov != NULL)
return TEST_skip("Test does not support a non-default library context");
/*
* This is a legacy method for CMACs, but should still work.
* This verifies that it works without an ENGINE.
*/
kctx = EVP_PKEY_CTX_new_id(EVP_PKEY_CMAC, NULL);
Extend test_CMAC_keygen in evp_extra_test The test only setup the generation of a key. It did not complete that generation. We extend the test to complete the generation and to use the generated key. We also compare the result with a key generated in a different way. Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/12637)
2020-08-12 21:41:12 +08:00
/* Test a CMAC key created using the "generated" method */
if (!TEST_int_gt(EVP_PKEY_keygen_init(kctx), 0)
|| !TEST_int_gt(EVP_PKEY_CTX_ctrl(kctx, -1, EVP_PKEY_OP_KEYGEN,
EVP_PKEY_CTRL_CIPHER,
test_CMAC_keygen(): Avoid using ECB cipher with CMAC Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/19401)
2022-10-12 18:22:25 +08:00
0, (void *)EVP_aes_256_cbc()), 0)
Extend test_CMAC_keygen in evp_extra_test The test only setup the generation of a key. It did not complete that generation. We extend the test to complete the generation and to use the generated key. We also compare the result with a key generated in a different way. Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/12637)
2020-08-12 21:41:12 +08:00
|| !TEST_int_gt(EVP_PKEY_CTX_ctrl(kctx, -1, EVP_PKEY_OP_KEYGEN,
EVP_PKEY_CTRL_SET_MAC_KEY,
sizeof(key), (void *)key), 0)
|| !TEST_int_gt(EVP_PKEY_keygen(kctx, &pkey), 0)
|| !TEST_ptr(pkey)
|| !TEST_true(get_cmac_val(pkey, mac)))
goto done;
test/evp_extra_test.c: Add test for CMAC keygen with a NULL engine Verifies that #11671 is fixed Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/11674)
2020-05-12 21:27:32 +08:00
Deprecate EVP_KEY_new_CMAC_key and EVP_PKEY_new_CMAC_key_ex EVP_KEY_new_CMAC_key_ex was in the pre-release 3.0 only, so is safe to remove. Restore 1.1.1 version of EVP_PKEY_new_CMAC_key documentation. Also make testing of EVP_PKEY_new_CMAC_key properly #ifdef'd. Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13829)
2020-12-08 23:13:54 +08:00
# if !defined(OPENSSL_NO_DEPRECATED_3_0)
Extend test_CMAC_keygen in evp_extra_test The test only setup the generation of a key. It did not complete that generation. We extend the test to complete the generation and to use the generated key. We also compare the result with a key generated in a different way. Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/12637)
2020-08-12 21:41:12 +08:00
EVP_PKEY_free(pkey);
/*
* Test a CMAC key using the direct method, and compare with the mac
* created above.
*/
test_CMAC_keygen(): Avoid using ECB cipher with CMAC Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/19401)
2022-10-12 18:22:25 +08:00
pkey = EVP_PKEY_new_CMAC_key(NULL, key, sizeof(key), EVP_aes_256_cbc());
Extend test_CMAC_keygen in evp_extra_test The test only setup the generation of a key. It did not complete that generation. We extend the test to complete the generation and to use the generated key. We also compare the result with a key generated in a different way. Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/12637)
2020-08-12 21:41:12 +08:00
if (!TEST_ptr(pkey)
|| !TEST_true(get_cmac_val(pkey, mac2))
|| !TEST_mem_eq(mac, sizeof(mac), mac2, sizeof(mac2)))
test/evp_extra_test.c: Add test for CMAC keygen with a NULL engine Verifies that #11671 is fixed Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/11674)
2020-05-12 21:27:32 +08:00
goto done;
Deprecate EVP_KEY_new_CMAC_key and EVP_PKEY_new_CMAC_key_ex EVP_KEY_new_CMAC_key_ex was in the pre-release 3.0 only, so is safe to remove. Restore 1.1.1 version of EVP_PKEY_new_CMAC_key documentation. Also make testing of EVP_PKEY_new_CMAC_key properly #ifdef'd. Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13829)
2020-12-08 23:13:54 +08:00
# endif
Extend test_CMAC_keygen in evp_extra_test The test only setup the generation of a key. It did not complete that generation. We extend the test to complete the generation and to use the generated key. We also compare the result with a key generated in a different way. Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/12637)
2020-08-12 21:41:12 +08:00
test/evp_extra_test.c: Add test for CMAC keygen with a NULL engine Verifies that #11671 is fixed Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/11674)
2020-05-12 21:27:32 +08:00
ret = 1;
done:
Extend test_CMAC_keygen in evp_extra_test The test only setup the generation of a key. It did not complete that generation. We extend the test to complete the generation and to use the generated key. We also compare the result with a key generated in a different way. Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/12637)
2020-08-12 21:41:12 +08:00
EVP_PKEY_free(pkey);
test/evp_extra_test.c: Add test for CMAC keygen with a NULL engine Verifies that #11671 is fixed Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/11674)
2020-05-12 21:27:32 +08:00
EVP_PKEY_CTX_free(kctx);
return ret;
}
test/evp_extra_test.c: Add OPENSSL_NO_CMAC around CMAC test Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/11833)
2020-05-14 23:15:05 +08:00
#endif
test/evp_extra_test.c: Add test for CMAC keygen with a NULL engine Verifies that #11671 is fixed Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/11674)
2020-05-12 21:27:32 +08:00
Add a test where we reuse the EVP_PKEY_CTX for two HKDF test runs Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7501)
2018-10-26 19:45:27 +08:00
static int test_HKDF(void)
{
EVP_PKEY_CTX *pctx;
unsigned char out[20];
size_t outlen;
int i, ret = 0;
unsigned char salt[] = "0123456789";
unsigned char key[] = "012345678901234567890123456789";
unsigned char info[] = "infostring";
const unsigned char expected[] = {
0xe5, 0x07, 0x70, 0x7f, 0xc6, 0x78, 0xd6, 0x54, 0x32, 0x5f, 0x7e, 0xc5,
0x7b, 0x59, 0x3e, 0xd8, 0x03, 0x6b, 0xed, 0xca
};
size_t expectedlen = sizeof(expected);
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
if (!TEST_ptr(pctx = EVP_PKEY_CTX_new_from_name(testctx, "HKDF", testpropq)))
Add a test where we reuse the EVP_PKEY_CTX for two HKDF test runs Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7501)
2018-10-26 19:45:27 +08:00
goto done;
/* We do this twice to test reuse of the EVP_PKEY_CTX */
for (i = 0; i < 2; i++) {
outlen = sizeof(out);
memset(out, 0, outlen);
if (!TEST_int_gt(EVP_PKEY_derive_init(pctx), 0)
|| !TEST_int_gt(EVP_PKEY_CTX_set_hkdf_md(pctx, EVP_sha256()), 0)
|| !TEST_int_gt(EVP_PKEY_CTX_set1_hkdf_salt(pctx, salt,
sizeof(salt) - 1), 0)
|| !TEST_int_gt(EVP_PKEY_CTX_set1_hkdf_key(pctx, key,
sizeof(key) - 1), 0)
|| !TEST_int_gt(EVP_PKEY_CTX_add1_hkdf_info(pctx, info,
sizeof(info) - 1), 0)
|| !TEST_int_gt(EVP_PKEY_derive(pctx, out, &outlen), 0)
|| !TEST_mem_eq(out, outlen, expected, expectedlen))
goto done;
}
ret = 1;
done:
EVP_PKEY_CTX_free(pctx);
return ret;
}
test HKDF with empty IKM Add an extra EVP test that provides empty input key material. It currently fails, since we lose the information about "key present but zero length" as we deserialize parameters in the provider. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/11920)
2020-05-22 03:53:59 +08:00
static int test_emptyikm_HKDF(void)
{
EVP_PKEY_CTX *pctx;
unsigned char out[20];
size_t outlen;
int ret = 0;
unsigned char salt[] = "9876543210";
unsigned char key[] = "";
unsigned char info[] = "stringinfo";
const unsigned char expected[] = {
0x68, 0x81, 0xa5, 0x3e, 0x5b, 0x9c, 0x7b, 0x6f, 0x2e, 0xec, 0xc8, 0x47,
0x7c, 0xfa, 0x47, 0x35, 0x66, 0x82, 0x15, 0x30
};
size_t expectedlen = sizeof(expected);
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
if (!TEST_ptr(pctx = EVP_PKEY_CTX_new_from_name(testctx, "HKDF", testpropq)))
test HKDF with empty IKM Add an extra EVP test that provides empty input key material. It currently fails, since we lose the information about "key present but zero length" as we deserialize parameters in the provider. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/11920)
2020-05-22 03:53:59 +08:00
goto done;
outlen = sizeof(out);
memset(out, 0, outlen);
if (!TEST_int_gt(EVP_PKEY_derive_init(pctx), 0)
|| !TEST_int_gt(EVP_PKEY_CTX_set_hkdf_md(pctx, EVP_sha256()), 0)
|| !TEST_int_gt(EVP_PKEY_CTX_set1_hkdf_salt(pctx, salt,
sizeof(salt) - 1), 0)
|| !TEST_int_gt(EVP_PKEY_CTX_set1_hkdf_key(pctx, key,
sizeof(key) - 1), 0)
|| !TEST_int_gt(EVP_PKEY_CTX_add1_hkdf_info(pctx, info,
sizeof(info) - 1), 0)
|| !TEST_int_gt(EVP_PKEY_derive(pctx, out, &outlen), 0)
|| !TEST_mem_eq(out, outlen, expected, expectedlen))
goto done;
ret = 1;
done:
EVP_PKEY_CTX_free(pctx);
return ret;
}
Handle empty param in EVP_PKEY_CTX_add1_hkdf_info Fixes #24130 The regression was introduced in PR #23456. Reviewed-by: Paul Dale <ppzgs1@gmail.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/24141)
2024-04-15 17:13:14 +08:00
static int test_empty_salt_info_HKDF(void)
{
EVP_PKEY_CTX *pctx;
unsigned char out[20];
size_t outlen;
int ret = 0;
unsigned char salt[] = "";
unsigned char key[] = "012345678901234567890123456789";
unsigned char info[] = "";
const unsigned char expected[] = {
0x67, 0x12, 0xf9, 0x27, 0x8a, 0x8a, 0x3a, 0x8f, 0x7d, 0x2c, 0xa3, 0x6a,
0xaa, 0xe9, 0xb3, 0xb9, 0x52, 0x5f, 0xe0, 0x06,
};
size_t expectedlen = sizeof(expected);
if (!TEST_ptr(pctx = EVP_PKEY_CTX_new_from_name(testctx, "HKDF", testpropq)))
goto done;
outlen = sizeof(out);
memset(out, 0, outlen);
if (!TEST_int_gt(EVP_PKEY_derive_init(pctx), 0)
|| !TEST_int_gt(EVP_PKEY_CTX_set_hkdf_md(pctx, EVP_sha256()), 0)
|| !TEST_int_gt(EVP_PKEY_CTX_set1_hkdf_salt(pctx, salt,
sizeof(salt) - 1), 0)
|| !TEST_int_gt(EVP_PKEY_CTX_set1_hkdf_key(pctx, key,
sizeof(key) - 1), 0)
|| !TEST_int_gt(EVP_PKEY_CTX_add1_hkdf_info(pctx, info,
sizeof(info) - 1), 0)
|| !TEST_int_gt(EVP_PKEY_derive(pctx, out, &outlen), 0)
|| !TEST_mem_eq(out, outlen, expected, expectedlen))
goto done;
ret = 1;
done:
EVP_PKEY_CTX_free(pctx);
return ret;
}
Fix a crash in reuse of i2d_X509_PUBKEY If the second PUBKEY is malformed there is use after free. Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/8122)
2019-01-30 23:20:31 +08:00
#ifndef OPENSSL_NO_EC
static int test_X509_PUBKEY_inplace(void)
{
test: add test case for X508_PUBKEY_dup() function Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14642)
2021-03-25 10:05:54 +08:00
int ret = 0;
Fix evp_extra_test to use libctx in an X509_PUBKEY Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/15591)
2021-05-21 22:50:43 +08:00
X509_PUBKEY *xp = X509_PUBKEY_new_ex(testctx, testpropq);
test: add test case for X508_PUBKEY_dup() function Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14642)
2021-03-25 10:05:54 +08:00
const unsigned char *p = kExampleECPubKeyDER;
size_t input_len = sizeof(kExampleECPubKeyDER);
Fix a crash in reuse of i2d_X509_PUBKEY If the second PUBKEY is malformed there is use after free. Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/8122)
2019-01-30 23:20:31 +08:00
Fix evp_extra_test to use libctx in an X509_PUBKEY Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/15591)
2021-05-21 22:50:43 +08:00
if (!TEST_ptr(xp))
goto done;
if (!TEST_ptr(d2i_X509_PUBKEY(&xp, &p, input_len)))
test: add test case for X508_PUBKEY_dup() function Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14642)
2021-03-25 10:05:54 +08:00
goto done;
Fix a crash in reuse of i2d_X509_PUBKEY If the second PUBKEY is malformed there is use after free. Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/8122)
2019-01-30 23:20:31 +08:00
test: add test case for X508_PUBKEY_dup() function Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14642)
2021-03-25 10:05:54 +08:00
if (!TEST_ptr(X509_PUBKEY_get0(xp)))
goto done;
Fix a crash in reuse of i2d_X509_PUBKEY If the second PUBKEY is malformed there is use after free. Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/8122)
2019-01-30 23:20:31 +08:00
test: add test case for X508_PUBKEY_dup() function Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14642)
2021-03-25 10:05:54 +08:00
p = kExampleBadECPubKeyDER;
input_len = sizeof(kExampleBadECPubKeyDER);
Fix a crash in reuse of i2d_X509_PUBKEY If the second PUBKEY is malformed there is use after free. Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/8122)
2019-01-30 23:20:31 +08:00
test: add test case for X508_PUBKEY_dup() function Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14642)
2021-03-25 10:05:54 +08:00
if (!TEST_ptr(xp = d2i_X509_PUBKEY(&xp, &p, input_len)))
goto done;
Fix a crash in reuse of i2d_X509_PUBKEY If the second PUBKEY is malformed there is use after free. Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/8122)
2019-01-30 23:20:31 +08:00
test: add test case for X508_PUBKEY_dup() function Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14642)
2021-03-25 10:05:54 +08:00
if (!TEST_true(X509_PUBKEY_get0(xp) == NULL))
goto done;
Fix a crash in reuse of i2d_X509_PUBKEY If the second PUBKEY is malformed there is use after free. Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/8122)
2019-01-30 23:20:31 +08:00
test: add test case for X508_PUBKEY_dup() function Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14642)
2021-03-25 10:05:54 +08:00
ret = 1;
Fix a crash in reuse of i2d_X509_PUBKEY If the second PUBKEY is malformed there is use after free. Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/8122)
2019-01-30 23:20:31 +08:00
test: add test case for X508_PUBKEY_dup() function Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14642)
2021-03-25 10:05:54 +08:00
done:
X509_PUBKEY_free(xp);
return ret;
}
static int test_X509_PUBKEY_dup(void)
{
int ret = 0;
X509_PUBKEY *xp = NULL, *xq = NULL;
const unsigned char *p = kExampleECPubKeyDER;
size_t input_len = sizeof(kExampleECPubKeyDER);
Fix evp_extra_test to use libctx in an X509_PUBKEY Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/15591)
2021-05-21 22:50:43 +08:00
xp = X509_PUBKEY_new_ex(testctx, testpropq);
if (!TEST_ptr(xp)
|| !TEST_ptr(d2i_X509_PUBKEY(&xp, &p, input_len))
test: add test case for X508_PUBKEY_dup() function Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14642)
2021-03-25 10:05:54 +08:00
|| !TEST_ptr(xq = X509_PUBKEY_dup(xp))
|| !TEST_ptr_ne(xp, xq))
goto done;
if (!TEST_ptr(X509_PUBKEY_get0(xq))
|| !TEST_ptr(X509_PUBKEY_get0(xp))
X509_PUBKEY_dup: Do not just up-ref the EVP_PKEY We try EVP_PKEY_dup() and if it fails we re-decode it using the legacy method as provided keys should be duplicable. Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16648)
2021-10-22 20:22:57 +08:00
|| !TEST_ptr_ne(X509_PUBKEY_get0(xq), X509_PUBKEY_get0(xp)))
test: add test case for X508_PUBKEY_dup() function Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14642)
2021-03-25 10:05:54 +08:00
goto done;
X509_PUBKEY_free(xq);
xq = NULL;
p = kExampleBadECPubKeyDER;
input_len = sizeof(kExampleBadECPubKeyDER);
if (!TEST_ptr(xp = d2i_X509_PUBKEY(&xp, &p, input_len))
|| !TEST_ptr(xq = X509_PUBKEY_dup(xp)))
goto done;
X509_PUBKEY_free(xp);
xp = NULL;
if (!TEST_true(X509_PUBKEY_get0(xq) == NULL))
goto done;
ret = 1;
done:
X509_PUBKEY_free(xp);
X509_PUBKEY_free(xq);
return ret;
Fix a crash in reuse of i2d_X509_PUBKEY If the second PUBKEY is malformed there is use after free. Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/8122)
2019-01-30 23:20:31 +08:00
}
Add fips module integrity check Add environment variable for setting CONF .include path Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/9769)
2019-09-15 17:55:10 +08:00
#endif /* OPENSSL_NO_EC */
cipher cleanups. Add test to evp_test_extra for ciphers (that is similiar to the digest_fetch). Move some of the aes and gcm methods that can be shared with other ciphers into ciphers_common.c Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/9580)
2019-08-19 07:30:59 +08:00
Revise EVP_PKEY param handling We add new functions for getting parameters and discovering the gettable and settable parameters. We also make EVP_PKEY_CTX_get_signature_md() a function and implement it in terms of the new functions. This enables applications to discover the set of parameters that are supported for a given algorithm implementation. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/9753)
2019-09-04 19:46:02 +08:00
/* Test getting and setting parameters on an EVP_PKEY_CTX */
test/evp_extra_test.c: adapt for RSA signature tests Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/10557)
2019-12-02 18:26:15 +08:00
static int test_EVP_PKEY_CTX_get_set_params(EVP_PKEY *pkey)
Revise EVP_PKEY param handling We add new functions for getting parameters and discovering the gettable and settable parameters. We also make EVP_PKEY_CTX_get_signature_md() a function and implement it in terms of the new functions. This enables applications to discover the set of parameters that are supported for a given algorithm implementation. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/9753)
2019-09-04 19:46:02 +08:00
{
Add a test for the newly added md params code Previous commits added code for routing md related parameters via and EVP_SIGNATURE implementation during a DigestSign operation. This adds a test to make sure this works as expected. Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/10013)
2019-09-25 18:49:04 +08:00
EVP_MD_CTX *mdctx = NULL;
Revise EVP_PKEY param handling We add new functions for getting parameters and discovering the gettable and settable parameters. We also make EVP_PKEY_CTX_get_signature_md() a function and implement it in terms of the new functions. This enables applications to discover the set of parameters that are supported for a given algorithm implementation. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/9753)
2019-09-04 19:46:02 +08:00
EVP_PKEY_CTX *ctx = NULL;
const OSSL_PARAM *params;
Don't pass a digest-size to signature implementations It turns out this was never necessary, as the implementation should always check the default digest size anyway. Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/10947)
2020-02-02 19:55:05 +08:00
OSSL_PARAM ourparams[2], *param = ourparams, *param_md;
Revise EVP_PKEY param handling We add new functions for getting parameters and discovering the gettable and settable parameters. We also make EVP_PKEY_CTX_get_signature_md() a function and implement it in terms of the new functions. This enables applications to discover the set of parameters that are supported for a given algorithm implementation. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/9753)
2019-09-04 19:46:02 +08:00
int ret = 0;
const EVP_MD *md;
Don't pass a digest-size to signature implementations It turns out this was never necessary, as the implementation should always check the default digest size anyway. Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/10947)
2020-02-02 19:55:05 +08:00
char mdname[OSSL_MAX_NAME_SIZE];
Add a test for the newly added md params code Previous commits added code for routing md related parameters via and EVP_SIGNATURE implementation during a DigestSign operation. This adds a test to make sure this works as expected. Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/10013)
2019-09-25 18:49:04 +08:00
char ssl3ms[48];
Revise EVP_PKEY param handling We add new functions for getting parameters and discovering the gettable and settable parameters. We also make EVP_PKEY_CTX_get_signature_md() a function and implement it in terms of the new functions. This enables applications to discover the set of parameters that are supported for a given algorithm implementation. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/9753)
2019-09-04 19:46:02 +08:00
/* Initialise a sign operation */
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
ctx = EVP_PKEY_CTX_new_from_pkey(testctx, pkey, testpropq);
Revise EVP_PKEY param handling We add new functions for getting parameters and discovering the gettable and settable parameters. We also make EVP_PKEY_CTX_get_signature_md() a function and implement it in terms of the new functions. This enables applications to discover the set of parameters that are supported for a given algorithm implementation. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/9753)
2019-09-04 19:46:02 +08:00
if (!TEST_ptr(ctx)
Minimal adaptation of tests back to how it was before Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10303)
2019-10-31 00:06:48 +08:00
|| !TEST_int_gt(EVP_PKEY_sign_init(ctx), 0))
Revise EVP_PKEY param handling We add new functions for getting parameters and discovering the gettable and settable parameters. We also make EVP_PKEY_CTX_get_signature_md() a function and implement it in terms of the new functions. This enables applications to discover the set of parameters that are supported for a given algorithm implementation. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/9753)
2019-09-04 19:46:02 +08:00
goto err;
/*
test/evp_extra_test.c: adapt for RSA signature tests Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/10557)
2019-12-02 18:26:15 +08:00
* We should be able to query the parameters now.
Revise EVP_PKEY param handling We add new functions for getting parameters and discovering the gettable and settable parameters. We also make EVP_PKEY_CTX_get_signature_md() a function and implement it in terms of the new functions. This enables applications to discover the set of parameters that are supported for a given algorithm implementation. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/9753)
2019-09-04 19:46:02 +08:00
*/
params = EVP_PKEY_CTX_settable_params(ctx);
if (!TEST_ptr(params)
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)
2020-01-23 03:59:56 +08:00
|| !TEST_ptr(OSSL_PARAM_locate_const(params,
OSSL_SIGNATURE_PARAM_DIGEST)))
Revise EVP_PKEY param handling We add new functions for getting parameters and discovering the gettable and settable parameters. We also make EVP_PKEY_CTX_get_signature_md() a function and implement it in terms of the new functions. This enables applications to discover the set of parameters that are supported for a given algorithm implementation. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/9753)
2019-09-04 19:46:02 +08:00
goto err;
params = EVP_PKEY_CTX_gettable_params(ctx);
if (!TEST_ptr(params)
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)
2020-01-23 03:59:56 +08:00
|| !TEST_ptr(OSSL_PARAM_locate_const(params,
OSSL_SIGNATURE_PARAM_ALGORITHM_ID))
|| !TEST_ptr(OSSL_PARAM_locate_const(params,
OSSL_SIGNATURE_PARAM_DIGEST)))
Revise EVP_PKEY param handling We add new functions for getting parameters and discovering the gettable and settable parameters. We also make EVP_PKEY_CTX_get_signature_md() a function and implement it in terms of the new functions. This enables applications to discover the set of parameters that are supported for a given algorithm implementation. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/9753)
2019-09-04 19:46:02 +08:00
goto err;
/*
* Test getting and setting params via EVP_PKEY_CTX_set_params() and
* EVP_PKEY_CTX_get_params()
*/
Don't pass a digest-size to signature implementations It turns out this was never necessary, as the implementation should always check the default digest size anyway. Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/10947)
2020-02-02 19:55:05 +08:00
strcpy(mdname, "SHA512");
param_md = param;
*param++ = OSSL_PARAM_construct_utf8_string(OSSL_SIGNATURE_PARAM_DIGEST,
mdname, 0);
Revise EVP_PKEY param handling We add new functions for getting parameters and discovering the gettable and settable parameters. We also make EVP_PKEY_CTX_get_signature_md() a function and implement it in terms of the new functions. This enables applications to discover the set of parameters that are supported for a given algorithm implementation. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/9753)
2019-09-04 19:46:02 +08:00
*param++ = OSSL_PARAM_construct_end();
if (!TEST_true(EVP_PKEY_CTX_set_params(ctx, ourparams)))
goto err;
Don't pass a digest-size to signature implementations It turns out this was never necessary, as the implementation should always check the default digest size anyway. Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/10947)
2020-02-02 19:55:05 +08:00
mdname[0] = '\0';
*param_md = OSSL_PARAM_construct_utf8_string(OSSL_SIGNATURE_PARAM_DIGEST,
mdname, sizeof(mdname));
Revise EVP_PKEY param handling We add new functions for getting parameters and discovering the gettable and settable parameters. We also make EVP_PKEY_CTX_get_signature_md() a function and implement it in terms of the new functions. This enables applications to discover the set of parameters that are supported for a given algorithm implementation. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/9753)
2019-09-04 19:46:02 +08:00
if (!TEST_true(EVP_PKEY_CTX_get_params(ctx, ourparams))
Don't pass a digest-size to signature implementations It turns out this was never necessary, as the implementation should always check the default digest size anyway. Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/10947)
2020-02-02 19:55:05 +08:00
|| !TEST_str_eq(mdname, "SHA512"))
Revise EVP_PKEY param handling We add new functions for getting parameters and discovering the gettable and settable parameters. We also make EVP_PKEY_CTX_get_signature_md() a function and implement it in terms of the new functions. This enables applications to discover the set of parameters that are supported for a given algorithm implementation. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/9753)
2019-09-04 19:46:02 +08:00
goto err;
/*
* Test the TEST_PKEY_CTX_set_signature_md() and
* TEST_PKEY_CTX_get_signature_md() functions
*/
if (!TEST_int_gt(EVP_PKEY_CTX_set_signature_md(ctx, EVP_sha256()), 0)
|| !TEST_int_gt(EVP_PKEY_CTX_get_signature_md(ctx, &md), 0)
|| !TEST_ptr_eq(md, EVP_sha256()))
goto err;
Add a test for the newly added md params code Previous commits added code for routing md related parameters via and EVP_SIGNATURE implementation during a DigestSign operation. This adds a test to make sure this works as expected. Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/10013)
2019-09-25 18:49:04 +08:00
/*
* Test getting MD parameters via an associated EVP_PKEY_CTX
*/
mdctx = EVP_MD_CTX_new();
if (!TEST_ptr(mdctx)
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
|| !TEST_true(EVP_DigestSignInit_ex(mdctx, NULL, "SHA1", testctx, testpropq,
test: support params arguments to init functions Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/14383)
2021-03-02 20:41:24 +08:00
pkey, NULL)))
Add a test for the newly added md params code Previous commits added code for routing md related parameters via and EVP_SIGNATURE implementation during a DigestSign operation. This adds a test to make sure this works as expected. Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/10013)
2019-09-25 18:49:04 +08:00
goto err;
/*
* We now have an EVP_MD_CTX with an EVP_PKEY_CTX inside it. We should be
* able to obtain the digest's settable parameters from the provider.
*/
params = EVP_MD_CTX_settable_params(mdctx);
if (!TEST_ptr(params)
|| !TEST_int_eq(strcmp(params[0].key, OSSL_DIGEST_PARAM_SSL3_MS), 0)
/* The final key should be NULL */
|| !TEST_ptr_null(params[1].key))
goto err;
param = ourparams;
memset(ssl3ms, 0, sizeof(ssl3ms));
*param++ = OSSL_PARAM_construct_octet_string(OSSL_DIGEST_PARAM_SSL3_MS,
ssl3ms, sizeof(ssl3ms));
*param++ = OSSL_PARAM_construct_end();
if (!TEST_true(EVP_MD_CTX_set_params(mdctx, ourparams)))
goto err;
Revise EVP_PKEY param handling We add new functions for getting parameters and discovering the gettable and settable parameters. We also make EVP_PKEY_CTX_get_signature_md() a function and implement it in terms of the new functions. This enables applications to discover the set of parameters that are supported for a given algorithm implementation. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/9753)
2019-09-04 19:46:02 +08:00
ret = 1;
err:
Add a test for the newly added md params code Previous commits added code for routing md related parameters via and EVP_SIGNATURE implementation during a DigestSign operation. This adds a test to make sure this works as expected. Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/10013)
2019-09-25 18:49:04 +08:00
EVP_MD_CTX_free(mdctx);
Revise EVP_PKEY param handling We add new functions for getting parameters and discovering the gettable and settable parameters. We also make EVP_PKEY_CTX_get_signature_md() a function and implement it in terms of the new functions. This enables applications to discover the set of parameters that are supported for a given algorithm implementation. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/9753)
2019-09-04 19:46:02 +08:00
EVP_PKEY_CTX_free(ctx);
test/evp_extra_test.c: adapt for RSA signature tests Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/10557)
2019-12-02 18:26:15 +08:00
return ret;
}
#ifndef OPENSSL_NO_DSA
static int test_DSA_get_set_params(void)
{
Fix 'no-deprecated' Some of the handling of no-deprecated stuff wasn't quite complete, or even plain wrong. This restores i2d_PublicKey() to be able to handle EVP_PKEYs with legacy internal keys. This also refactors the DSA key tests in test/evp_extra_test.c to use EVP functionality entirely. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/13706)
2020-12-18 20:17:33 +08:00
OSSL_PARAM_BLD *bld = NULL;
OSSL_PARAM *params = NULL;
test/evp_extra_test.c: adapt for RSA signature tests Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/10557)
2019-12-02 18:26:15 +08:00
BIGNUM *p = NULL, *q = NULL, *g = NULL, *pub = NULL, *priv = NULL;
Fix 'no-deprecated' Some of the handling of no-deprecated stuff wasn't quite complete, or even plain wrong. This restores i2d_PublicKey() to be able to handle EVP_PKEYs with legacy internal keys. This also refactors the DSA key tests in test/evp_extra_test.c to use EVP functionality entirely. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/13706)
2020-12-18 20:17:33 +08:00
EVP_PKEY_CTX *pctx = NULL;
test/evp_extra_test.c: adapt for RSA signature tests Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/10557)
2019-12-02 18:26:15 +08:00
EVP_PKEY *pkey = NULL;
int ret = 0;
/*
* Setup the parameters for our DSA object. For our purposes they don't
* have to actually be *valid* parameters. We just need to set something.
*/
Fix 'no-deprecated' Some of the handling of no-deprecated stuff wasn't quite complete, or even plain wrong. This restores i2d_PublicKey() to be able to handle EVP_PKEYs with legacy internal keys. This also refactors the DSA key tests in test/evp_extra_test.c to use EVP functionality entirely. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/13706)
2020-12-18 20:17:33 +08:00
if (!TEST_ptr(pctx = EVP_PKEY_CTX_new_from_name(testctx, "DSA", NULL))
|| !TEST_ptr(bld = OSSL_PARAM_BLD_new())
|| !TEST_ptr(p = BN_new())
|| !TEST_ptr(q = BN_new())
|| !TEST_ptr(g = BN_new())
|| !TEST_ptr(pub = BN_new())
|| !TEST_ptr(priv = BN_new()))
goto err;
if (!TEST_true(OSSL_PARAM_BLD_push_BN(bld, OSSL_PKEY_PARAM_FFC_P, p))
|| !TEST_true(OSSL_PARAM_BLD_push_BN(bld, OSSL_PKEY_PARAM_FFC_Q, q))
|| !TEST_true(OSSL_PARAM_BLD_push_BN(bld, OSSL_PKEY_PARAM_FFC_G, g))
|| !TEST_true(OSSL_PARAM_BLD_push_BN(bld, OSSL_PKEY_PARAM_PUB_KEY,
pub))
|| !TEST_true(OSSL_PARAM_BLD_push_BN(bld, OSSL_PKEY_PARAM_PRIV_KEY,
priv)))
goto err;
if (!TEST_ptr(params = OSSL_PARAM_BLD_to_param(bld)))
test/evp_extra_test.c: adapt for RSA signature tests Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/10557)
2019-12-02 18:26:15 +08:00
goto err;
Simplify the EVP_PKEY_XXX_fromdata_XX methods. The existing names such as EVP_PKEY_param_fromdata_settable were a bit confusing since the 'param' referred to key params not OSSL_PARAM. To simplify the interface a 'selection' parameter will be passed instead. The changes are: (1) EVP_PKEY_fromdata_init() replaces both EVP_PKEY_key_fromdata_init() and EVP_PKEY_param_fromdata_init(). (2) EVP_PKEY_fromdata() has an additional selection parameter. (3) EVP_PKEY_fromdata_settable() replaces EVP_PKEY_key_fromdata_settable() and EVP_PKEY_param_fromdata_settable(). EVP_PKEY_fromdata_settable() also uses a selection parameter. Fixes #12989 Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14076)
2021-02-05 11:55:50 +08:00
if (!TEST_int_gt(EVP_PKEY_fromdata_init(pctx), 0)
|| !TEST_int_gt(EVP_PKEY_fromdata(pctx, &pkey, EVP_PKEY_KEYPAIR,
params), 0))
test/evp_extra_test.c: adapt for RSA signature tests Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/10557)
2019-12-02 18:26:15 +08:00
goto err;
Fix 'no-deprecated' Some of the handling of no-deprecated stuff wasn't quite complete, or even plain wrong. This restores i2d_PublicKey() to be able to handle EVP_PKEYs with legacy internal keys. This also refactors the DSA key tests in test/evp_extra_test.c to use EVP functionality entirely. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/13706)
2020-12-18 20:17:33 +08:00
if (!TEST_ptr(pkey))
goto err;
test/evp_extra_test.c: adapt for RSA signature tests Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/10557)
2019-12-02 18:26:15 +08:00
ret = test_EVP_PKEY_CTX_get_set_params(pkey);
err:
Revise EVP_PKEY param handling We add new functions for getting parameters and discovering the gettable and settable parameters. We also make EVP_PKEY_CTX_get_signature_md() a function and implement it in terms of the new functions. This enables applications to discover the set of parameters that are supported for a given algorithm implementation. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/9753)
2019-09-04 19:46:02 +08:00
EVP_PKEY_free(pkey);
Fix 'no-deprecated' Some of the handling of no-deprecated stuff wasn't quite complete, or even plain wrong. This restores i2d_PublicKey() to be able to handle EVP_PKEYs with legacy internal keys. This also refactors the DSA key tests in test/evp_extra_test.c to use EVP functionality entirely. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/13706)
2020-12-18 20:17:33 +08:00
EVP_PKEY_CTX_free(pctx);
Replace OSSL_PARAM_BLD_free_params() with OSSL_PARAM_free(). Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14785)
2021-04-07 11:45:19 +08:00
OSSL_PARAM_free(params);
Fix 'no-deprecated' Some of the handling of no-deprecated stuff wasn't quite complete, or even plain wrong. This restores i2d_PublicKey() to be able to handle EVP_PKEYs with legacy internal keys. This also refactors the DSA key tests in test/evp_extra_test.c to use EVP functionality entirely. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/13706)
2020-12-18 20:17:33 +08:00
OSSL_PARAM_BLD_free(bld);
Revise EVP_PKEY param handling We add new functions for getting parameters and discovering the gettable and settable parameters. We also make EVP_PKEY_CTX_get_signature_md() a function and implement it in terms of the new functions. This enables applications to discover the set of parameters that are supported for a given algorithm implementation. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/9753)
2019-09-04 19:46:02 +08:00
BN_free(p);
BN_free(q);
BN_free(g);
Don't abuse the API when that's not what is tested test_EVP_PKEY_CTX_get_set_params() in test/evp_extra_test.c abused previously sloppy checking in EVP_PKEY_sign_init_ex(), by passing a "key to sign with" that was really just domain parameters. Now that underlying provider import of key payload has become a bit more strict, that leads to errors, so we need to provide at least a public part (even though fake), and because this is a signing operation, a private part as well. Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10169)
2019-10-15 17:35:09 +08:00
BN_free(pub);
BN_free(priv);
Revise EVP_PKEY param handling We add new functions for getting parameters and discovering the gettable and settable parameters. We also make EVP_PKEY_CTX_get_signature_md() a function and implement it in terms of the new functions. This enables applications to discover the set of parameters that are supported for a given algorithm implementation. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/9753)
2019-09-04 19:46:02 +08:00
return ret;
}
Allow EVP_PKEY private key objects to be created without a public component Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/13855)
2021-01-13 12:10:51 +08:00
/*
Fix no-dh and no-dsa Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13915)
2021-01-21 17:19:16 +08:00
* Test combinations of private, public, missing and private + public key
Allow EVP_PKEY private key objects to be created without a public component Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/13855)
2021-01-13 12:10:51 +08:00
* params to ensure they are all accepted
*/
static int test_DSA_priv_pub(void)
{
return test_EVP_PKEY_ffc_priv_pub("DSA");
}
#endif /* !OPENSSL_NO_DSA */
Revise EVP_PKEY param handling We add new functions for getting parameters and discovering the gettable and settable parameters. We also make EVP_PKEY_CTX_get_signature_md() a function and implement it in terms of the new functions. This enables applications to discover the set of parameters that are supported for a given algorithm implementation. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/9753)
2019-09-04 19:46:02 +08:00
test/evp_extra_test.c: adapt for RSA signature tests Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/10557)
2019-12-02 18:26:15 +08:00
static int test_RSA_get_set_params(void)
{
Deprecate RSA harder This deprecates all functions that deal with the types RSA and RSA_METHOD Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/13096)
2020-10-04 22:34:31 +08:00
OSSL_PARAM_BLD *bld = NULL;
OSSL_PARAM *params = NULL;
test/evp_extra_test.c: adapt for RSA signature tests Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/10557)
2019-12-02 18:26:15 +08:00
BIGNUM *n = NULL, *e = NULL, *d = NULL;
Deprecate RSA harder This deprecates all functions that deal with the types RSA and RSA_METHOD Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/13096)
2020-10-04 22:34:31 +08:00
EVP_PKEY_CTX *pctx = NULL;
test/evp_extra_test.c: adapt for RSA signature tests Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/10557)
2019-12-02 18:26:15 +08:00
EVP_PKEY *pkey = NULL;
int ret = 0;
/*
* Setup the parameters for our RSA object. For our purposes they don't
* have to actually be *valid* parameters. We just need to set something.
*/
Deprecate RSA harder This deprecates all functions that deal with the types RSA and RSA_METHOD Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/13096)
2020-10-04 22:34:31 +08:00
if (!TEST_ptr(pctx = EVP_PKEY_CTX_new_from_name(testctx, "RSA", NULL))
|| !TEST_ptr(bld = OSSL_PARAM_BLD_new())
|| !TEST_ptr(n = BN_new())
|| !TEST_ptr(e = BN_new())
|| !TEST_ptr(d = BN_new()))
goto err;
if (!TEST_true(OSSL_PARAM_BLD_push_BN(bld, OSSL_PKEY_PARAM_RSA_N, n))
|| !TEST_true(OSSL_PARAM_BLD_push_BN(bld, OSSL_PKEY_PARAM_RSA_E, e))
|| !TEST_true(OSSL_PARAM_BLD_push_BN(bld, OSSL_PKEY_PARAM_RSA_D, d)))
goto err;
if (!TEST_ptr(params = OSSL_PARAM_BLD_to_param(bld)))
test/evp_extra_test.c: adapt for RSA signature tests Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/10557)
2019-12-02 18:26:15 +08:00
goto err;
Simplify the EVP_PKEY_XXX_fromdata_XX methods. The existing names such as EVP_PKEY_param_fromdata_settable were a bit confusing since the 'param' referred to key params not OSSL_PARAM. To simplify the interface a 'selection' parameter will be passed instead. The changes are: (1) EVP_PKEY_fromdata_init() replaces both EVP_PKEY_key_fromdata_init() and EVP_PKEY_param_fromdata_init(). (2) EVP_PKEY_fromdata() has an additional selection parameter. (3) EVP_PKEY_fromdata_settable() replaces EVP_PKEY_key_fromdata_settable() and EVP_PKEY_param_fromdata_settable(). EVP_PKEY_fromdata_settable() also uses a selection parameter. Fixes #12989 Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14076)
2021-02-05 11:55:50 +08:00
if (!TEST_int_gt(EVP_PKEY_fromdata_init(pctx), 0)
|| !TEST_int_gt(EVP_PKEY_fromdata(pctx, &pkey, EVP_PKEY_KEYPAIR,
params), 0))
test/evp_extra_test.c: adapt for RSA signature tests Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/10557)
2019-12-02 18:26:15 +08:00
goto err;
Deprecate RSA harder This deprecates all functions that deal with the types RSA and RSA_METHOD Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/13096)
2020-10-04 22:34:31 +08:00
if (!TEST_ptr(pkey))
goto err;
test/evp_extra_test.c: adapt for RSA signature tests Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/10557)
2019-12-02 18:26:15 +08:00
ret = test_EVP_PKEY_CTX_get_set_params(pkey);
err:
EVP_PKEY_free(pkey);
Deprecate RSA harder This deprecates all functions that deal with the types RSA and RSA_METHOD Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/13096)
2020-10-04 22:34:31 +08:00
EVP_PKEY_CTX_free(pctx);
Replace OSSL_PARAM_BLD_free_params() with OSSL_PARAM_free(). Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14785)
2021-04-07 11:45:19 +08:00
OSSL_PARAM_free(params);
Deprecate RSA harder This deprecates all functions that deal with the types RSA and RSA_METHOD Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/13096)
2020-10-04 22:34:31 +08:00
OSSL_PARAM_BLD_free(bld);
test/evp_extra_test.c: adapt for RSA signature tests Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/10557)
2019-12-02 18:26:15 +08:00
BN_free(n);
BN_free(e);
BN_free(d);
return ret;
}
set MGF1 digest correctly Fixes #19290 update rsa_set_ctx_params() so that the digest function used in the MGF1 construction is set correctly. Add a test for this to evp_extra_test.c based on the code scaro-axway provided in #19290. Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com> Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/19342)
2022-10-04 22:48:43 +08:00
static int test_RSA_OAEP_set_get_params(void)
{
int ret = 0;
EVP_PKEY *key = NULL;
EVP_PKEY_CTX *key_ctx = NULL;
if (nullprov != NULL)
return TEST_skip("Test does not support a non-default library context");
if (!TEST_ptr(key = load_example_rsa_key())
|| !TEST_ptr(key_ctx = EVP_PKEY_CTX_new_from_pkey(0, key, 0)))
goto err;
{
int padding = RSA_PKCS1_OAEP_PADDING;
OSSL_PARAM params[4];
params[0] = OSSL_PARAM_construct_int(OSSL_SIGNATURE_PARAM_PAD_MODE, &padding);
params[1] = OSSL_PARAM_construct_utf8_string(OSSL_ASYM_CIPHER_PARAM_OAEP_DIGEST,
OSSL_DIGEST_NAME_SHA2_256, 0);
params[2] = OSSL_PARAM_construct_utf8_string(OSSL_ASYM_CIPHER_PARAM_MGF1_DIGEST,
OSSL_DIGEST_NAME_SHA1, 0);
params[3] = OSSL_PARAM_construct_end();
if (!TEST_int_gt(EVP_PKEY_encrypt_init_ex(key_ctx, params),0))
goto err;
}
{
OSSL_PARAM params[3];
char oaepmd[30] = { '\0' };
char mgf1md[30] = { '\0' };
params[0] = OSSL_PARAM_construct_utf8_string(OSSL_ASYM_CIPHER_PARAM_OAEP_DIGEST,
oaepmd, sizeof(oaepmd));
params[1] = OSSL_PARAM_construct_utf8_string(OSSL_ASYM_CIPHER_PARAM_MGF1_DIGEST,
mgf1md, sizeof(mgf1md));
params[2] = OSSL_PARAM_construct_end();
if (!TEST_true(EVP_PKEY_CTX_get_params(key_ctx, params)))
goto err;
if (!TEST_str_eq(oaepmd, OSSL_DIGEST_NAME_SHA2_256)
|| !TEST_str_eq(mgf1md, OSSL_DIGEST_NAME_SHA1))
goto err;
}
ret = 1;
err:
EVP_PKEY_free(key);
EVP_PKEY_CTX_free(key_ctx);
return ret;
}
rsa: Accept NULL OAEP label for backward compatibility According to the manual page, EVP_PKEY_CTX_set0_rsa_oaep_label() should accept NULL as the label argument, though the function currently rejects it while setting the corresponding octet string parameter with OSSL_PARAM_construct_octet_string, which expects non-NULL input. This adds a workaround to the caller for backward compatibility. Signed-off-by: Daiki Ueno <dueno@redhat.com> Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/22397)
2023-10-16 13:42:12 +08:00
/* https://github.com/openssl/openssl/issues/21288 */
static int test_RSA_OAEP_set_null_label(void)
{
int ret = 0;
EVP_PKEY *key = NULL;
EVP_PKEY_CTX *key_ctx = NULL;
if (!TEST_ptr(key = load_example_rsa_key())
|| !TEST_ptr(key_ctx = EVP_PKEY_CTX_new_from_pkey(testctx, key, NULL))
|| !TEST_true(EVP_PKEY_encrypt_init(key_ctx)))
goto err;
if (!TEST_true(EVP_PKEY_CTX_set_rsa_padding(key_ctx, RSA_PKCS1_OAEP_PADDING)))
goto err;
if (!TEST_true(EVP_PKEY_CTX_set0_rsa_oaep_label(key_ctx, OPENSSL_strdup("foo"), 0)))
goto err;
if (!TEST_true(EVP_PKEY_CTX_set0_rsa_oaep_label(key_ctx, NULL, 0)))
goto err;
ret = 1;
err:
EVP_PKEY_free(key);
EVP_PKEY_CTX_free(key_ctx);
return ret;
}
Augment RSA provider to generate CRT coefficients on EVP_PKEY_fromdata() It would be helpful to be able to generate RSA's dmp1/dmq1/iqmp values when not provided in the param list to EVP_PKEY_fromdata. Augment the provider in ossl_rsa_fromdata to preform this generation iff: a) At least p q n e and e are provided b) the new parameter OSSL_PARAM_RSA_DERIVE_PQ is set to 1 Fixes #21826 Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/21875)
2023-08-28 20:48:34 +08:00
#ifndef OPENSSL_NO_DEPRECATED_3_0
static int test_RSA_legacy(void)
{
int ret = 0;
BIGNUM *p = NULL;
BIGNUM *q = NULL;
BIGNUM *n = NULL;
BIGNUM *e = NULL;
BIGNUM *d = NULL;
const EVP_MD *md = EVP_sha256();
EVP_MD_CTX *ctx = NULL;
EVP_PKEY *pkey = NULL;
RSA *rsa = NULL;
if (nullprov != NULL)
return TEST_skip("Test does not support a non-default library context");
if (!TEST_ptr(p = BN_dup(BN_value_one()))
|| !TEST_ptr(q = BN_dup(BN_value_one()))
|| !TEST_ptr(n = BN_dup(BN_value_one()))
|| !TEST_ptr(e = BN_dup(BN_value_one()))
|| !TEST_ptr(d = BN_dup(BN_value_one())))
goto err;
if (!TEST_ptr(rsa = RSA_new())
|| !TEST_ptr(pkey = EVP_PKEY_new())
|| !TEST_ptr(ctx = EVP_MD_CTX_new()))
goto err;
if (!TEST_true(RSA_set0_factors(rsa, p, q)))
goto err;
p = NULL;
q = NULL;
if (!TEST_true(RSA_set0_key(rsa, n, e, d)))
goto err;
n = NULL;
e = NULL;
d = NULL;
if (!TEST_true(EVP_PKEY_assign_RSA(pkey, rsa)))
goto err;
rsa = NULL;
if (!TEST_true(EVP_DigestSignInit(ctx, NULL, md, NULL, pkey)))
goto err;
ret = 1;
err:
RSA_free(rsa);
EVP_MD_CTX_free(ctx);
EVP_PKEY_free(pkey);
BN_free(p);
BN_free(q);
BN_free(n);
BN_free(e);
BN_free(d);
return ret;
}
#endif
Add a test for NULL chunks in encrypt/decrypt Issue #8675 describes a problem where calling EVP_DecryptUpdate() with an empty chunk causes the result to be different compared to if you do not use an empty chunk. This adds a test for that case. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10530)
2019-05-31 21:32:55 +08:00
#if !defined(OPENSSL_NO_CHACHA) && !defined(OPENSSL_NO_POLY1305)
static int test_decrypt_null_chunks(void)
{
EVP_CIPHER_CTX* ctx = NULL;
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
EVP_CIPHER *cipher = NULL;
Add a test for NULL chunks in encrypt/decrypt Issue #8675 describes a problem where calling EVP_DecryptUpdate() with an empty chunk causes the result to be different compared to if you do not use an empty chunk. This adds a test for that case. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10530)
2019-05-31 21:32:55 +08:00
const unsigned char key[32] = {
0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08, 0x09, 0x0a, 0x0b,
0x0c, 0x0d, 0x0e, 0x0f, 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17,
0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1
};
unsigned char iv[12] = {
0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08, 0x09, 0x0a, 0x0b
};
unsigned char msg[] = "It was the best of times, it was the worst of times";
unsigned char ciphertext[80];
unsigned char plaintext[80];
/* We initialise tmp to a non zero value on purpose */
int ctlen, ptlen, tmp = 99;
int ret = 0;
const int enc_offset = 10, dec_offset = 20;
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
if (!TEST_ptr(cipher = EVP_CIPHER_fetch(testctx, "ChaCha20-Poly1305", testpropq))
|| !TEST_ptr(ctx = EVP_CIPHER_CTX_new())
|| !TEST_true(EVP_EncryptInit_ex(ctx, cipher, NULL,
Add a test for NULL chunks in encrypt/decrypt Issue #8675 describes a problem where calling EVP_DecryptUpdate() with an empty chunk causes the result to be different compared to if you do not use an empty chunk. This adds a test for that case. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10530)
2019-05-31 21:32:55 +08:00
key, iv))
|| !TEST_true(EVP_EncryptUpdate(ctx, ciphertext, &ctlen, msg,
enc_offset))
/* Deliberate add a zero length update */
|| !TEST_true(EVP_EncryptUpdate(ctx, ciphertext + ctlen, &tmp, NULL,
0))
|| !TEST_int_eq(tmp, 0)
|| !TEST_true(EVP_EncryptUpdate(ctx, ciphertext + ctlen, &tmp,
msg + enc_offset,
sizeof(msg) - enc_offset))
|| !TEST_int_eq(ctlen += tmp, sizeof(msg))
|| !TEST_true(EVP_EncryptFinal(ctx, ciphertext + ctlen, &tmp))
|| !TEST_int_eq(tmp, 0))
goto err;
/* Deliberately initialise tmp to a non zero value */
tmp = 99;
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
if (!TEST_true(EVP_DecryptInit_ex(ctx, cipher, NULL, key, iv))
Add a test for NULL chunks in encrypt/decrypt Issue #8675 describes a problem where calling EVP_DecryptUpdate() with an empty chunk causes the result to be different compared to if you do not use an empty chunk. This adds a test for that case. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10530)
2019-05-31 21:32:55 +08:00
|| !TEST_true(EVP_DecryptUpdate(ctx, plaintext, &ptlen, ciphertext,
dec_offset))
/*
* Deliberately add a zero length update. We also deliberately do
* this at a different offset than for encryption.
*/
|| !TEST_true(EVP_DecryptUpdate(ctx, plaintext + ptlen, &tmp, NULL,
0))
|| !TEST_int_eq(tmp, 0)
|| !TEST_true(EVP_DecryptUpdate(ctx, plaintext + ptlen, &tmp,
ciphertext + dec_offset,
ctlen - dec_offset))
|| !TEST_int_eq(ptlen += tmp, sizeof(msg))
|| !TEST_true(EVP_DecryptFinal(ctx, plaintext + ptlen, &tmp))
|| !TEST_int_eq(tmp, 0)
|| !TEST_mem_eq(msg, sizeof(msg), plaintext, ptlen))
goto err;
ret = 1;
err:
EVP_CIPHER_CTX_free(ctx);
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
EVP_CIPHER_free(cipher);
Add a test for NULL chunks in encrypt/decrypt Issue #8675 describes a problem where calling EVP_DecryptUpdate() with an empty chunk causes the result to be different compared to if you do not use an empty chunk. This adds a test for that case. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10530)
2019-05-31 21:32:55 +08:00
return ret;
}
#endif /* !defined(OPENSSL_NO_CHACHA) && !defined(OPENSSL_NO_POLY1305) */
Allow EVP_PKEY private key objects to be created without a public component Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/13855)
2021-01-13 12:10:51 +08:00
#ifndef OPENSSL_NO_DH
/*
Fix no-dh and no-dsa Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13915)
2021-01-21 17:19:16 +08:00
* Test combinations of private, public, missing and private + public key
Allow EVP_PKEY private key objects to be created without a public component Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/13855)
2021-01-13 12:10:51 +08:00
* params to ensure they are all accepted
*/
static int test_DH_priv_pub(void)
{
return test_EVP_PKEY_ffc_priv_pub("DH");
}
# ifndef OPENSSL_NO_DEPRECATED_3_0
Test that EVP_PKEY_set1_DH() correctly identifies the DH type Provide a test to check tat when we assign a DH object we know whether we are dealing with PKCS#3 or X9.42 DH keys. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10593)
2019-12-09 20:03:02 +08:00
static int test_EVP_PKEY_set1_DH(void)
{
Add support for DH 'modp' group parameters (RFC 3526) Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10897)
2020-01-31 06:18:46 +08:00
DH *x942dh = NULL, *noqdh = NULL;
EVP_PKEY *pkey1 = NULL, *pkey2 = NULL;
Test that EVP_PKEY_set1_DH() correctly identifies the DH type Provide a test to check tat when we assign a DH object we know whether we are dealing with PKCS#3 or X9.42 DH keys. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10593)
2019-12-09 20:03:02 +08:00
int ret = 0;
Add support for DH 'modp' group parameters (RFC 3526) Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10897)
2020-01-31 06:18:46 +08:00
BIGNUM *p, *g = NULL;
Multiple fixes for getting pub key from legacy DH PKEY There were multiple issues with getting OSSL_PKEY_PARAM_PUB_KEY from a legacy EVP_PKEY DH and DHX keys. Fixes #16247 Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16253)
2021-08-06 23:25:13 +08:00
BIGNUM *pubkey = NULL;
unsigned char pub[2048 / 8];
size_t len = 0;
Add support for DH 'modp' group parameters (RFC 3526) Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10897)
2020-01-31 06:18:46 +08:00
if (!TEST_ptr(p = BN_new())
|| !TEST_ptr(g = BN_new())
Multiple fixes for getting pub key from legacy DH PKEY There were multiple issues with getting OSSL_PKEY_PARAM_PUB_KEY from a legacy EVP_PKEY DH and DHX keys. Fixes #16247 Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16253)
2021-08-06 23:25:13 +08:00
|| !TEST_ptr(pubkey = BN_new())
|| !TEST_true(BN_set_word(p, 9999))
|| !TEST_true(BN_set_word(g, 2))
|| !TEST_true(BN_set_word(pubkey, 4321))
Add support for DH 'modp' group parameters (RFC 3526) Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10897)
2020-01-31 06:18:46 +08:00
|| !TEST_ptr(noqdh = DH_new())
Multiple fixes for getting pub key from legacy DH PKEY There were multiple issues with getting OSSL_PKEY_PARAM_PUB_KEY from a legacy EVP_PKEY DH and DHX keys. Fixes #16247 Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16253)
2021-08-06 23:25:13 +08:00
|| !TEST_true(DH_set0_pqg(noqdh, p, NULL, g))
|| !TEST_true(DH_set0_key(noqdh, pubkey, NULL))
|| !TEST_ptr(pubkey = BN_new())
|| !TEST_true(BN_set_word(pubkey, 4321)))
Add support for DH 'modp' group parameters (RFC 3526) Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10897)
2020-01-31 06:18:46 +08:00
goto err;
p = g = NULL;
Test that EVP_PKEY_set1_DH() correctly identifies the DH type Provide a test to check tat when we assign a DH object we know whether we are dealing with PKCS#3 or X9.42 DH keys. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10593)
2019-12-09 20:03:02 +08:00
x942dh = DH_get_2048_256();
pkey1 = EVP_PKEY_new();
pkey2 = EVP_PKEY_new();
if (!TEST_ptr(x942dh)
Add support for DH 'modp' group parameters (RFC 3526) Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10897)
2020-01-31 06:18:46 +08:00
|| !TEST_ptr(noqdh)
Test that EVP_PKEY_set1_DH() correctly identifies the DH type Provide a test to check tat when we assign a DH object we know whether we are dealing with PKCS#3 or X9.42 DH keys. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10593)
2019-12-09 20:03:02 +08:00
|| !TEST_ptr(pkey1)
Multiple fixes for getting pub key from legacy DH PKEY There were multiple issues with getting OSSL_PKEY_PARAM_PUB_KEY from a legacy EVP_PKEY DH and DHX keys. Fixes #16247 Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16253)
2021-08-06 23:25:13 +08:00
|| !TEST_ptr(pkey2)
|| !TEST_true(DH_set0_key(x942dh, pubkey, NULL)))
Test that EVP_PKEY_set1_DH() correctly identifies the DH type Provide a test to check tat when we assign a DH object we know whether we are dealing with PKCS#3 or X9.42 DH keys. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10593)
2019-12-09 20:03:02 +08:00
goto err;
Multiple fixes for getting pub key from legacy DH PKEY There were multiple issues with getting OSSL_PKEY_PARAM_PUB_KEY from a legacy EVP_PKEY DH and DHX keys. Fixes #16247 Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16253)
2021-08-06 23:25:13 +08:00
pubkey = NULL;
Test that EVP_PKEY_set1_DH() correctly identifies the DH type Provide a test to check tat when we assign a DH object we know whether we are dealing with PKCS#3 or X9.42 DH keys. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10593)
2019-12-09 20:03:02 +08:00
Multiple fixes for getting pub key from legacy DH PKEY There were multiple issues with getting OSSL_PKEY_PARAM_PUB_KEY from a legacy EVP_PKEY DH and DHX keys. Fixes #16247 Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16253)
2021-08-06 23:25:13 +08:00
if (!TEST_true(EVP_PKEY_set1_DH(pkey1, x942dh))
Rename all getters to use get/get0 in name For functions that exist in 1.1.1 provide a simple aliases via #define. Fixes #15236 Functions with OSSL_DECODER_, OSSL_ENCODER_, OSSL_STORE_LOADER_, EVP_KEYEXCH_, EVP_KEM_, EVP_ASYM_CIPHER_, EVP_SIGNATURE_, EVP_KEYMGMT_, EVP_RAND_, EVP_MAC_, EVP_KDF_, EVP_PKEY_, EVP_MD_, and EVP_CIPHER_ prefixes are renamed. Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/15405)
2021-05-21 22:58:08 +08:00
|| !TEST_int_eq(EVP_PKEY_get_id(pkey1), EVP_PKEY_DHX))
Test that EVP_PKEY_set1_DH() correctly identifies the DH type Provide a test to check tat when we assign a DH object we know whether we are dealing with PKCS#3 or X9.42 DH keys. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10593)
2019-12-09 20:03:02 +08:00
goto err;
Multiple fixes for getting pub key from legacy DH PKEY There were multiple issues with getting OSSL_PKEY_PARAM_PUB_KEY from a legacy EVP_PKEY DH and DHX keys. Fixes #16247 Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16253)
2021-08-06 23:25:13 +08:00
if (!TEST_true(EVP_PKEY_get_bn_param(pkey1, OSSL_PKEY_PARAM_PUB_KEY,
&pubkey))
|| !TEST_ptr(pubkey))
goto err;
if (!TEST_true(EVP_PKEY_set1_DH(pkey2, noqdh))
Rename all getters to use get/get0 in name For functions that exist in 1.1.1 provide a simple aliases via #define. Fixes #15236 Functions with OSSL_DECODER_, OSSL_ENCODER_, OSSL_STORE_LOADER_, EVP_KEYEXCH_, EVP_KEM_, EVP_ASYM_CIPHER_, EVP_SIGNATURE_, EVP_KEYMGMT_, EVP_RAND_, EVP_MAC_, EVP_KDF_, EVP_PKEY_, EVP_MD_, and EVP_CIPHER_ prefixes are renamed. Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/15405)
2021-05-21 22:58:08 +08:00
|| !TEST_int_eq(EVP_PKEY_get_id(pkey2), EVP_PKEY_DH))
Test that EVP_PKEY_set1_DH() correctly identifies the DH type Provide a test to check tat when we assign a DH object we know whether we are dealing with PKCS#3 or X9.42 DH keys. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10593)
2019-12-09 20:03:02 +08:00
goto err;
Multiple fixes for getting pub key from legacy DH PKEY There were multiple issues with getting OSSL_PKEY_PARAM_PUB_KEY from a legacy EVP_PKEY DH and DHX keys. Fixes #16247 Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16253)
2021-08-06 23:25:13 +08:00
if (!TEST_true(EVP_PKEY_get_octet_string_param(pkey2,
OSSL_PKEY_PARAM_PUB_KEY,
pub, sizeof(pub), &len))
|| !TEST_size_t_ne(len, 0))
goto err;
Test that EVP_PKEY_set1_DH() correctly identifies the DH type Provide a test to check tat when we assign a DH object we know whether we are dealing with PKCS#3 or X9.42 DH keys. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10593)
2019-12-09 20:03:02 +08:00
ret = 1;
err:
Add support for DH 'modp' group parameters (RFC 3526) Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10897)
2020-01-31 06:18:46 +08:00
BN_free(p);
BN_free(g);
Multiple fixes for getting pub key from legacy DH PKEY There were multiple issues with getting OSSL_PKEY_PARAM_PUB_KEY from a legacy EVP_PKEY DH and DHX keys. Fixes #16247 Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16253)
2021-08-06 23:25:13 +08:00
BN_free(pubkey);
Test that EVP_PKEY_set1_DH() correctly identifies the DH type Provide a test to check tat when we assign a DH object we know whether we are dealing with PKCS#3 or X9.42 DH keys. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10593)
2019-12-09 20:03:02 +08:00
EVP_PKEY_free(pkey1);
EVP_PKEY_free(pkey2);
DH_free(x942dh);
Add support for DH 'modp' group parameters (RFC 3526) Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10897)
2020-01-31 06:18:46 +08:00
DH_free(noqdh);
Test that EVP_PKEY_set1_DH() correctly identifies the DH type Provide a test to check tat when we assign a DH object we know whether we are dealing with PKCS#3 or X9.42 DH keys. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10593)
2019-12-09 20:03:02 +08:00
return ret;
}
Allow EVP_PKEY private key objects to be created without a public component Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/13855)
2021-01-13 12:10:51 +08:00
# endif /* !OPENSSL_NO_DEPRECATED_3_0 */
#endif /* !OPENSSL_NO_DH */
Test that EVP_PKEY_set1_DH() correctly identifies the DH type Provide a test to check tat when we assign a DH object we know whether we are dealing with PKCS#3 or X9.42 DH keys. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10593)
2019-12-09 20:03:02 +08:00
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)
2020-04-15 19:36:19 +08:00
/*
* We test what happens with an empty template. For the sake of this test,
* the template must be ignored, and we know that's the case for RSA keys
* (this might arguably be a misfeature, but that's what we currently do,
* even in provider code, since that's how the legacy RSA implementation
* does things)
*/
static int test_keygen_with_empty_template(int n)
{
EVP_PKEY_CTX *ctx = NULL;
EVP_PKEY *pkey = NULL;
EVP_PKEY *tkey = NULL;
int ret = 0;
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
if (nullprov != NULL)
return TEST_skip("Test does not support a non-default library context");
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)
2020-04-15 19:36:19 +08:00
switch (n) {
case 0:
/* We do test with no template at all as well */
if (!TEST_ptr(ctx = EVP_PKEY_CTX_new_id(EVP_PKEY_RSA, NULL)))
goto err;
break;
case 1:
/* Here we create an empty RSA key that serves as our template */
if (!TEST_ptr(tkey = EVP_PKEY_new())
|| !TEST_true(EVP_PKEY_set_type(tkey, EVP_PKEY_RSA))
|| !TEST_ptr(ctx = EVP_PKEY_CTX_new(tkey, NULL)))
goto err;
break;
}
if (!TEST_int_gt(EVP_PKEY_keygen_init(ctx), 0)
|| !TEST_int_gt(EVP_PKEY_keygen(ctx, &pkey), 0))
goto err;
ret = 1;
err:
EVP_PKEY_CTX_free(ctx);
EVP_PKEY_free(pkey);
EVP_PKEY_free(tkey);
return ret;
}
Add a test for fetching EVP_PKEY style algs without a provider Following on from the previous commit, add a test to check that we fail to create an EVP_PKEY_CTX if an algorithm is not available in any provider, *unless* it is an algorithm that has no provider support. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/11826)
2020-05-14 18:33:01 +08:00
/*
* Test that we fail if we attempt to use an algorithm that is not available
Fix test/evp_extra_test.c Because EVP_PKEY_CTX_new_from_name() could return a non-NULL context with no value in it, the lack of legacy implementation when OpenSSL was configured with 'no-ec' went through undetected. This adds the necessary guards to skip a test of SM2 in that case. Reviewed-by: Paul Yang <kaishen.yy@antfin.com> (Merged from https://github.com/openssl/openssl/pull/12785)
2020-09-04 16:52:20 +08:00
* in the current library context (unless we are using an algorithm that
* should be made available via legacy codepaths).
*
* 0: RSA
* 1: SM2
Add a test for fetching EVP_PKEY style algs without a provider Following on from the previous commit, add a test to check that we fail to create an EVP_PKEY_CTX if an algorithm is not available in any provider, *unless* it is an algorithm that has no provider support. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/11826)
2020-05-14 18:33:01 +08:00
*/
static int test_pkey_ctx_fail_without_provider(int tst)
{
Rename OPENSSL_CTX prefix to OSSL_LIB_CTX Many of the new types introduced by OpenSSL 3.0 have an OSSL_ prefix, e.g., OSSL_CALLBACK, OSSL_PARAM, OSSL_ALGORITHM, OSSL_SERIALIZER. The OPENSSL_CTX type stands out a little by using a different prefix. For consistency reasons, this type is renamed to OSSL_LIB_CTX. Reviewed-by: Paul Dale <paul.dale@oracle.com> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/12621)
2020-10-15 17:55:50 +08:00
OSSL_LIB_CTX *tmpctx = OSSL_LIB_CTX_new();
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
OSSL_PROVIDER *tmpnullprov = NULL;
Add a test for fetching EVP_PKEY style algs without a provider Following on from the previous commit, add a test to check that we fail to create an EVP_PKEY_CTX if an algorithm is not available in any provider, *unless* it is an algorithm that has no provider support. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/11826)
2020-05-14 18:33:01 +08:00
EVP_PKEY_CTX *pctx = NULL;
Fix test/evp_extra_test.c Because EVP_PKEY_CTX_new_from_name() could return a non-NULL context with no value in it, the lack of legacy implementation when OpenSSL was configured with 'no-ec' went through undetected. This adds the necessary guards to skip a test of SM2 in that case. Reviewed-by: Paul Yang <kaishen.yy@antfin.com> (Merged from https://github.com/openssl/openssl/pull/12785)
2020-09-04 16:52:20 +08:00
const char *keytype = NULL;
int expect_null = 0;
Add a test for fetching EVP_PKEY style algs without a provider Following on from the previous commit, add a test to check that we fail to create an EVP_PKEY_CTX if an algorithm is not available in any provider, *unless* it is an algorithm that has no provider support. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/11826)
2020-05-14 18:33:01 +08:00
int ret = 0;
if (!TEST_ptr(tmpctx))
goto err;
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
tmpnullprov = OSSL_PROVIDER_load(tmpctx, "null");
if (!TEST_ptr(tmpnullprov))
Add a test for fetching EVP_PKEY style algs without a provider Following on from the previous commit, add a test to check that we fail to create an EVP_PKEY_CTX if an algorithm is not available in any provider, *unless* it is an algorithm that has no provider support. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/11826)
2020-05-14 18:33:01 +08:00
goto err;
/*
Fix test/evp_extra_test.c Because EVP_PKEY_CTX_new_from_name() could return a non-NULL context with no value in it, the lack of legacy implementation when OpenSSL was configured with 'no-ec' went through undetected. This adds the necessary guards to skip a test of SM2 in that case. Reviewed-by: Paul Yang <kaishen.yy@antfin.com> (Merged from https://github.com/openssl/openssl/pull/12785)
2020-09-04 16:52:20 +08:00
* We check for certain algos in the null provider.
Fix typos Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/17392)
2022-01-03 07:00:27 +08:00
* If an algo is expected to have a provider keymgmt, constructing an
Fix test/evp_extra_test.c Because EVP_PKEY_CTX_new_from_name() could return a non-NULL context with no value in it, the lack of legacy implementation when OpenSSL was configured with 'no-ec' went through undetected. This adds the necessary guards to skip a test of SM2 in that case. Reviewed-by: Paul Yang <kaishen.yy@antfin.com> (Merged from https://github.com/openssl/openssl/pull/12785)
2020-09-04 16:52:20 +08:00
* EVP_PKEY_CTX is expected to fail (return NULL).
Fix typos Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/17392)
2022-01-03 07:00:27 +08:00
* Otherwise, if it's expected to have legacy support, constructing an
Fix test/evp_extra_test.c Because EVP_PKEY_CTX_new_from_name() could return a non-NULL context with no value in it, the lack of legacy implementation when OpenSSL was configured with 'no-ec' went through undetected. This adds the necessary guards to skip a test of SM2 in that case. Reviewed-by: Paul Yang <kaishen.yy@antfin.com> (Merged from https://github.com/openssl/openssl/pull/12785)
2020-09-04 16:52:20 +08:00
* EVP_PKEY_CTX is expected to succeed (return non-NULL).
Add a test for fetching EVP_PKEY style algs without a provider Following on from the previous commit, add a test to check that we fail to create an EVP_PKEY_CTX if an algorithm is not available in any provider, *unless* it is an algorithm that has no provider support. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/11826)
2020-05-14 18:33:01 +08:00
*/
Fix test/evp_extra_test.c Because EVP_PKEY_CTX_new_from_name() could return a non-NULL context with no value in it, the lack of legacy implementation when OpenSSL was configured with 'no-ec' went through undetected. This adds the necessary guards to skip a test of SM2 in that case. Reviewed-by: Paul Yang <kaishen.yy@antfin.com> (Merged from https://github.com/openssl/openssl/pull/12785)
2020-09-04 16:52:20 +08:00
switch (tst) {
case 0:
keytype = "RSA";
expect_null = 1;
break;
case 1:
keytype = "SM2";
Add SM2 signature algorithm to default provider Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/12536)
2020-03-04 23:49:43 +08:00
expect_null = 1;
Fix test/evp_extra_test.c Because EVP_PKEY_CTX_new_from_name() could return a non-NULL context with no value in it, the lack of legacy implementation when OpenSSL was configured with 'no-ec' went through undetected. This adds the necessary guards to skip a test of SM2 in that case. Reviewed-by: Paul Yang <kaishen.yy@antfin.com> (Merged from https://github.com/openssl/openssl/pull/12785)
2020-09-04 16:52:20 +08:00
#ifdef OPENSSL_NO_EC
TEST_info("EC disable, skipping SM2 check...");
goto end;
#endif
#ifdef OPENSSL_NO_SM2
TEST_info("SM2 disable, skipping SM2 check...");
goto end;
#endif
break;
default:
TEST_error("No test for case %d", tst);
goto err;
}
pctx = EVP_PKEY_CTX_new_from_name(tmpctx, keytype, "");
if (expect_null ? !TEST_ptr_null(pctx) : !TEST_ptr(pctx))
Add a test for fetching EVP_PKEY style algs without a provider Following on from the previous commit, add a test to check that we fail to create an EVP_PKEY_CTX if an algorithm is not available in any provider, *unless* it is an algorithm that has no provider support. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/11826)
2020-05-14 18:33:01 +08:00
goto err;
Fix test/evp_extra_test.c Because EVP_PKEY_CTX_new_from_name() could return a non-NULL context with no value in it, the lack of legacy implementation when OpenSSL was configured with 'no-ec' went through undetected. This adds the necessary guards to skip a test of SM2 in that case. Reviewed-by: Paul Yang <kaishen.yy@antfin.com> (Merged from https://github.com/openssl/openssl/pull/12785)
2020-09-04 16:52:20 +08:00
#if defined(OPENSSL_NO_EC) || defined(OPENSSL_NO_SM2)
end:
#endif
Add a test for fetching EVP_PKEY style algs without a provider Following on from the previous commit, add a test to check that we fail to create an EVP_PKEY_CTX if an algorithm is not available in any provider, *unless* it is an algorithm that has no provider support. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/11826)
2020-05-14 18:33:01 +08:00
ret = 1;
err:
EVP_PKEY_CTX_free(pctx);
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
OSSL_PROVIDER_unload(tmpnullprov);
Rename OPENSSL_CTX prefix to OSSL_LIB_CTX Many of the new types introduced by OpenSSL 3.0 have an OSSL_ prefix, e.g., OSSL_CALLBACK, OSSL_PARAM, OSSL_ALGORITHM, OSSL_SERIALIZER. The OPENSSL_CTX type stands out a little by using a different prefix. For consistency reasons, this type is renamed to OSSL_LIB_CTX. Reviewed-by: Paul Dale <paul.dale@oracle.com> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/12621)
2020-10-15 17:55:50 +08:00
OSSL_LIB_CTX_free(tmpctx);
Add a test for fetching EVP_PKEY style algs without a provider Following on from the previous commit, add a test to check that we fail to create an EVP_PKEY_CTX if an algorithm is not available in any provider, *unless* it is an algorithm that has no provider support. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/11826)
2020-05-14 18:33:01 +08:00
return ret;
}
test: add test for generation of random data in chunks. THe EVP_RAND wrapper works with the underlying RNG to produce the amount of random data requested even if it is larger than the largest single generation the source allows. This test verified that this works. Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> (Merged from https://github.com/openssl/openssl/pull/11682)
2020-06-11 09:07:13 +08:00
static int test_rand_agglomeration(void)
{
EVP_RAND *rand;
EVP_RAND_CTX *ctx;
OSSL_PARAM params[3], *p = params;
int res;
unsigned int step = 7;
static unsigned char seed[] = "It does not matter how slowly you go "
"as long as you do not stop.";
unsigned char out[sizeof(seed)];
if (!TEST_int_ne(sizeof(seed) % step, 0)
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
|| !TEST_ptr(rand = EVP_RAND_fetch(testctx, "TEST-RAND", testpropq)))
test: add test for generation of random data in chunks. THe EVP_RAND wrapper works with the underlying RNG to produce the amount of random data requested even if it is larger than the largest single generation the source allows. This test verified that this works. Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> (Merged from https://github.com/openssl/openssl/pull/11682)
2020-06-11 09:07:13 +08:00
return 0;
ctx = EVP_RAND_CTX_new(rand, NULL);
EVP_RAND_free(rand);
if (!TEST_ptr(ctx))
return 0;
memset(out, 0, sizeof(out));
*p++ = OSSL_PARAM_construct_octet_string(OSSL_RAND_PARAM_TEST_ENTROPY,
seed, sizeof(seed));
test: changes resulting from moving the entropy source out of the FIPS provider 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/13226)
2020-10-30 13:53:47 +08:00
*p++ = OSSL_PARAM_construct_uint(OSSL_RAND_PARAM_MAX_REQUEST, &step);
test: add test for generation of random data in chunks. THe EVP_RAND wrapper works with the underlying RNG to produce the amount of random data requested even if it is larger than the largest single generation the source allows. This test verified that this works. Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> (Merged from https://github.com/openssl/openssl/pull/11682)
2020-06-11 09:07:13 +08:00
*p = OSSL_PARAM_construct_end();
Fix naming for EVP_RAND_CTX_gettable functions. Change: EVP_RAND_gettable_ctx_params -> EVP_RAND_CTX_gettable_params EVP_RAND_settable_ctx_params -> EVP_RAND_CTX_settable_params Which brings them in line with the other similar functions for other algorithm types. Fixes #14880 Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14893)
2021-04-16 07:13:43 +08:00
res = TEST_true(EVP_RAND_CTX_set_params(ctx, params))
test: add test for generation of random data in chunks. THe EVP_RAND wrapper works with the underlying RNG to produce the amount of random data requested even if it is larger than the largest single generation the source allows. This test verified that this works. Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> (Merged from https://github.com/openssl/openssl/pull/11682)
2020-06-11 09:07:13 +08:00
&& TEST_true(EVP_RAND_generate(ctx, out, sizeof(out), 0, 1, NULL, 0))
&& TEST_mem_eq(seed, sizeof(seed), out, sizeof(out));
EVP_RAND_CTX_free(ctx);
return res;
}
Add tests for new EVP_CIPHER_CTX IV accessors Test that EVP_CIPHER_CTX_get_iv() returns the same IV that was given at initialization time, and that EVP_CIPHER_CTX_get_iv_state() returns the expected value after performing an encryption operation (which will differ from the previous value for CBC and OFB modes), for various modes of AES. Do this both for the implicit fetch and explicit fetch paths, at the cost of a slightly more complicated switch statement. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/12233)
2020-06-20 11:42:29 +08:00
/*
* Test that we correctly return the original or "running" IV after
* an encryption operation.
* Run multiple times for some different relevant algorithms/modes.
*/
Add test for the IV handling of DES based ciphers Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/14882)
2021-04-15 17:53:42 +08:00
static int test_evp_iv_aes(int idx)
Add tests for new EVP_CIPHER_CTX IV accessors Test that EVP_CIPHER_CTX_get_iv() returns the same IV that was given at initialization time, and that EVP_CIPHER_CTX_get_iv_state() returns the expected value after performing an encryption operation (which will differ from the previous value for CBC and OFB modes), for various modes of AES. Do this both for the implicit fetch and explicit fetch paths, at the cost of a slightly more complicated switch statement. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/12233)
2020-06-20 11:42:29 +08:00
{
int ret = 0;
EVP_CIPHER_CTX *ctx = NULL;
that open brace { should be on the previous line Found by running the checkpatch.pl Linux script to enforce coding style. Reviewed-by: Neil Horman <nhorman@openssl.org> Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/22097)
2023-09-14 02:32:54 +08:00
unsigned char key[16] = {
0x4c, 0x43, 0xdb, 0xdd, 0x42, 0x73, 0x47, 0xd1,
0xe5, 0x62, 0x7d, 0xcd, 0x4d, 0x76, 0x4d, 0x57
};
unsigned char init_iv[EVP_MAX_IV_LENGTH] = {
0x57, 0x71, 0x7d, 0xad, 0xdb, 0x9b, 0x98, 0x82,
0x5a, 0x55, 0x91, 0x81, 0x42, 0xa8, 0x89, 0x34
};
static const unsigned char msg[] = {
1, 2, 3, 4, 5, 6, 7, 8,
9, 10, 11, 12, 13, 14, 15, 16
};
Add tests for new EVP_CIPHER_CTX IV accessors Test that EVP_CIPHER_CTX_get_iv() returns the same IV that was given at initialization time, and that EVP_CIPHER_CTX_get_iv_state() returns the expected value after performing an encryption operation (which will differ from the previous value for CBC and OFB modes), for various modes of AES. Do this both for the implicit fetch and explicit fetch paths, at the cost of a slightly more complicated switch statement. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/12233)
2020-06-20 11:42:29 +08:00
unsigned char ciphertext[32], oiv[16], iv[16];
unsigned char *ref_iv;
that open brace { should be on the previous line Found by running the checkpatch.pl Linux script to enforce coding style. Reviewed-by: Neil Horman <nhorman@openssl.org> Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/22097)
2023-09-14 02:32:54 +08:00
unsigned char cbc_state[16] = {
0x10, 0x2f, 0x05, 0xcc, 0xc2, 0x55, 0x72, 0xb9,
0x88, 0xe6, 0x4a, 0x17, 0x10, 0x74, 0x22, 0x5e
};
unsigned char ofb_state[16] = {
0x76, 0xe6, 0x66, 0x61, 0xd0, 0x8a, 0xe4, 0x64,
0xdd, 0x66, 0xbf, 0x00, 0xf0, 0xe3, 0x6f, 0xfd
};
unsigned char cfb_state[16] = {
0x77, 0xe4, 0x65, 0x65, 0xd5, 0x8c, 0xe3, 0x6c,
0xd4, 0x6c, 0xb4, 0x0c, 0xfd, 0xed, 0x60, 0xed
};
unsigned char gcm_state[12] = {
0x57, 0x71, 0x7d, 0xad, 0xdb, 0x9b,
0x98, 0x82, 0x5a, 0x55, 0x91, 0x81
};
unsigned char ccm_state[7] = { 0x57, 0x71, 0x7d, 0xad, 0xdb, 0x9b, 0x98 };
Add tests for new EVP_CIPHER_CTX IV accessors Test that EVP_CIPHER_CTX_get_iv() returns the same IV that was given at initialization time, and that EVP_CIPHER_CTX_get_iv_state() returns the expected value after performing an encryption operation (which will differ from the previous value for CBC and OFB modes), for various modes of AES. Do this both for the implicit fetch and explicit fetch paths, at the cost of a slightly more complicated switch statement. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/12233)
2020-06-20 11:42:29 +08:00
#ifndef OPENSSL_NO_OCB
that open brace { should be on the previous line Found by running the checkpatch.pl Linux script to enforce coding style. Reviewed-by: Neil Horman <nhorman@openssl.org> Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/22097)
2023-09-14 02:32:54 +08:00
unsigned char ocb_state[12] = {
0x57, 0x71, 0x7d, 0xad, 0xdb, 0x9b,
0x98, 0x82, 0x5a, 0x55, 0x91, 0x81
};
Add tests for new EVP_CIPHER_CTX IV accessors Test that EVP_CIPHER_CTX_get_iv() returns the same IV that was given at initialization time, and that EVP_CIPHER_CTX_get_iv_state() returns the expected value after performing an encryption operation (which will differ from the previous value for CBC and OFB modes), for various modes of AES. Do this both for the implicit fetch and explicit fetch paths, at the cost of a slightly more complicated switch statement. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/12233)
2020-06-20 11:42:29 +08:00
#endif
int len = sizeof(ciphertext);
size_t ivlen, ref_len;
const EVP_CIPHER *type = NULL;
Always reset IV for CBC, OFB, and CFB mode on cipher context reinit This is necessary to keep compatibility with 1.1.1 implementation of the CBC, OFB, and CFB mode ciphers. Fixes #14704 Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14811)
2021-04-09 18:01:16 +08:00
int iv_reset = 0;
Add tests for new EVP_CIPHER_CTX IV accessors Test that EVP_CIPHER_CTX_get_iv() returns the same IV that was given at initialization time, and that EVP_CIPHER_CTX_get_iv_state() returns the expected value after performing an encryption operation (which will differ from the previous value for CBC and OFB modes), for various modes of AES. Do this both for the implicit fetch and explicit fetch paths, at the cost of a slightly more complicated switch statement. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/12233)
2020-06-20 11:42:29 +08:00
Always reset IV for CBC, OFB, and CFB mode on cipher context reinit This is necessary to keep compatibility with 1.1.1 implementation of the CBC, OFB, and CFB mode ciphers. Fixes #14704 Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14811)
2021-04-09 18:01:16 +08:00
if (nullprov != NULL && idx < 6)
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
return TEST_skip("Test does not support a non-default library context");
fix some code with obvious wrong coding style Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16918)
2021-10-26 15:16:18 +08:00
switch (idx) {
Add tests for new EVP_CIPHER_CTX IV accessors Test that EVP_CIPHER_CTX_get_iv() returns the same IV that was given at initialization time, and that EVP_CIPHER_CTX_get_iv_state() returns the expected value after performing an encryption operation (which will differ from the previous value for CBC and OFB modes), for various modes of AES. Do this both for the implicit fetch and explicit fetch paths, at the cost of a slightly more complicated switch statement. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/12233)
2020-06-20 11:42:29 +08:00
case 0:
type = EVP_aes_128_cbc();
/* FALLTHROUGH */
Always reset IV for CBC, OFB, and CFB mode on cipher context reinit This is necessary to keep compatibility with 1.1.1 implementation of the CBC, OFB, and CFB mode ciphers. Fixes #14704 Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14811)
2021-04-09 18:01:16 +08:00
case 6:
Add tests for new EVP_CIPHER_CTX IV accessors Test that EVP_CIPHER_CTX_get_iv() returns the same IV that was given at initialization time, and that EVP_CIPHER_CTX_get_iv_state() returns the expected value after performing an encryption operation (which will differ from the previous value for CBC and OFB modes), for various modes of AES. Do this both for the implicit fetch and explicit fetch paths, at the cost of a slightly more complicated switch statement. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/12233)
2020-06-20 11:42:29 +08:00
type = (type != NULL) ? type :
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
EVP_CIPHER_fetch(testctx, "aes-128-cbc", testpropq);
Add tests for new EVP_CIPHER_CTX IV accessors Test that EVP_CIPHER_CTX_get_iv() returns the same IV that was given at initialization time, and that EVP_CIPHER_CTX_get_iv_state() returns the expected value after performing an encryption operation (which will differ from the previous value for CBC and OFB modes), for various modes of AES. Do this both for the implicit fetch and explicit fetch paths, at the cost of a slightly more complicated switch statement. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/12233)
2020-06-20 11:42:29 +08:00
ref_iv = cbc_state;
ref_len = sizeof(cbc_state);
Always reset IV for CBC, OFB, and CFB mode on cipher context reinit This is necessary to keep compatibility with 1.1.1 implementation of the CBC, OFB, and CFB mode ciphers. Fixes #14704 Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14811)
2021-04-09 18:01:16 +08:00
iv_reset = 1;
Add tests for new EVP_CIPHER_CTX IV accessors Test that EVP_CIPHER_CTX_get_iv() returns the same IV that was given at initialization time, and that EVP_CIPHER_CTX_get_iv_state() returns the expected value after performing an encryption operation (which will differ from the previous value for CBC and OFB modes), for various modes of AES. Do this both for the implicit fetch and explicit fetch paths, at the cost of a slightly more complicated switch statement. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/12233)
2020-06-20 11:42:29 +08:00
break;
case 1:
type = EVP_aes_128_ofb();
/* FALLTHROUGH */
Always reset IV for CBC, OFB, and CFB mode on cipher context reinit This is necessary to keep compatibility with 1.1.1 implementation of the CBC, OFB, and CFB mode ciphers. Fixes #14704 Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14811)
2021-04-09 18:01:16 +08:00
case 7:
Add tests for new EVP_CIPHER_CTX IV accessors Test that EVP_CIPHER_CTX_get_iv() returns the same IV that was given at initialization time, and that EVP_CIPHER_CTX_get_iv_state() returns the expected value after performing an encryption operation (which will differ from the previous value for CBC and OFB modes), for various modes of AES. Do this both for the implicit fetch and explicit fetch paths, at the cost of a slightly more complicated switch statement. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/12233)
2020-06-20 11:42:29 +08:00
type = (type != NULL) ? type :
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
EVP_CIPHER_fetch(testctx, "aes-128-ofb", testpropq);
Add tests for new EVP_CIPHER_CTX IV accessors Test that EVP_CIPHER_CTX_get_iv() returns the same IV that was given at initialization time, and that EVP_CIPHER_CTX_get_iv_state() returns the expected value after performing an encryption operation (which will differ from the previous value for CBC and OFB modes), for various modes of AES. Do this both for the implicit fetch and explicit fetch paths, at the cost of a slightly more complicated switch statement. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/12233)
2020-06-20 11:42:29 +08:00
ref_iv = ofb_state;
ref_len = sizeof(ofb_state);
Always reset IV for CBC, OFB, and CFB mode on cipher context reinit This is necessary to keep compatibility with 1.1.1 implementation of the CBC, OFB, and CFB mode ciphers. Fixes #14704 Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14811)
2021-04-09 18:01:16 +08:00
iv_reset = 1;
Add tests for new EVP_CIPHER_CTX IV accessors Test that EVP_CIPHER_CTX_get_iv() returns the same IV that was given at initialization time, and that EVP_CIPHER_CTX_get_iv_state() returns the expected value after performing an encryption operation (which will differ from the previous value for CBC and OFB modes), for various modes of AES. Do this both for the implicit fetch and explicit fetch paths, at the cost of a slightly more complicated switch statement. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/12233)
2020-06-20 11:42:29 +08:00
break;
case 2:
Always reset IV for CBC, OFB, and CFB mode on cipher context reinit This is necessary to keep compatibility with 1.1.1 implementation of the CBC, OFB, and CFB mode ciphers. Fixes #14704 Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14811)
2021-04-09 18:01:16 +08:00
type = EVP_aes_128_cfb();
/* FALLTHROUGH */
case 8:
type = (type != NULL) ? type :
EVP_CIPHER_fetch(testctx, "aes-128-cfb", testpropq);
ref_iv = cfb_state;
ref_len = sizeof(cfb_state);
iv_reset = 1;
break;
case 3:
Add tests for new EVP_CIPHER_CTX IV accessors Test that EVP_CIPHER_CTX_get_iv() returns the same IV that was given at initialization time, and that EVP_CIPHER_CTX_get_iv_state() returns the expected value after performing an encryption operation (which will differ from the previous value for CBC and OFB modes), for various modes of AES. Do this both for the implicit fetch and explicit fetch paths, at the cost of a slightly more complicated switch statement. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/12233)
2020-06-20 11:42:29 +08:00
type = EVP_aes_128_gcm();
/* FALLTHROUGH */
Always reset IV for CBC, OFB, and CFB mode on cipher context reinit This is necessary to keep compatibility with 1.1.1 implementation of the CBC, OFB, and CFB mode ciphers. Fixes #14704 Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14811)
2021-04-09 18:01:16 +08:00
case 9:
Add tests for new EVP_CIPHER_CTX IV accessors Test that EVP_CIPHER_CTX_get_iv() returns the same IV that was given at initialization time, and that EVP_CIPHER_CTX_get_iv_state() returns the expected value after performing an encryption operation (which will differ from the previous value for CBC and OFB modes), for various modes of AES. Do this both for the implicit fetch and explicit fetch paths, at the cost of a slightly more complicated switch statement. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/12233)
2020-06-20 11:42:29 +08:00
type = (type != NULL) ? type :
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
EVP_CIPHER_fetch(testctx, "aes-128-gcm", testpropq);
Add tests for new EVP_CIPHER_CTX IV accessors Test that EVP_CIPHER_CTX_get_iv() returns the same IV that was given at initialization time, and that EVP_CIPHER_CTX_get_iv_state() returns the expected value after performing an encryption operation (which will differ from the previous value for CBC and OFB modes), for various modes of AES. Do this both for the implicit fetch and explicit fetch paths, at the cost of a slightly more complicated switch statement. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/12233)
2020-06-20 11:42:29 +08:00
ref_iv = gcm_state;
ref_len = sizeof(gcm_state);
break;
Always reset IV for CBC, OFB, and CFB mode on cipher context reinit This is necessary to keep compatibility with 1.1.1 implementation of the CBC, OFB, and CFB mode ciphers. Fixes #14704 Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14811)
2021-04-09 18:01:16 +08:00
case 4:
Add tests for new EVP_CIPHER_CTX IV accessors Test that EVP_CIPHER_CTX_get_iv() returns the same IV that was given at initialization time, and that EVP_CIPHER_CTX_get_iv_state() returns the expected value after performing an encryption operation (which will differ from the previous value for CBC and OFB modes), for various modes of AES. Do this both for the implicit fetch and explicit fetch paths, at the cost of a slightly more complicated switch statement. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/12233)
2020-06-20 11:42:29 +08:00
type = EVP_aes_128_ccm();
/* FALLTHROUGH */
Always reset IV for CBC, OFB, and CFB mode on cipher context reinit This is necessary to keep compatibility with 1.1.1 implementation of the CBC, OFB, and CFB mode ciphers. Fixes #14704 Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14811)
2021-04-09 18:01:16 +08:00
case 10:
Add tests for new EVP_CIPHER_CTX IV accessors Test that EVP_CIPHER_CTX_get_iv() returns the same IV that was given at initialization time, and that EVP_CIPHER_CTX_get_iv_state() returns the expected value after performing an encryption operation (which will differ from the previous value for CBC and OFB modes), for various modes of AES. Do this both for the implicit fetch and explicit fetch paths, at the cost of a slightly more complicated switch statement. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/12233)
2020-06-20 11:42:29 +08:00
type = (type != NULL) ? type :
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
EVP_CIPHER_fetch(testctx, "aes-128-ccm", testpropq);
Add tests for new EVP_CIPHER_CTX IV accessors Test that EVP_CIPHER_CTX_get_iv() returns the same IV that was given at initialization time, and that EVP_CIPHER_CTX_get_iv_state() returns the expected value after performing an encryption operation (which will differ from the previous value for CBC and OFB modes), for various modes of AES. Do this both for the implicit fetch and explicit fetch paths, at the cost of a slightly more complicated switch statement. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/12233)
2020-06-20 11:42:29 +08:00
ref_iv = ccm_state;
ref_len = sizeof(ccm_state);
break;
#ifdef OPENSSL_NO_OCB
Always reset IV for CBC, OFB, and CFB mode on cipher context reinit This is necessary to keep compatibility with 1.1.1 implementation of the CBC, OFB, and CFB mode ciphers. Fixes #14704 Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14811)
2021-04-09 18:01:16 +08:00
case 5:
case 11:
Add tests for new EVP_CIPHER_CTX IV accessors Test that EVP_CIPHER_CTX_get_iv() returns the same IV that was given at initialization time, and that EVP_CIPHER_CTX_get_iv_state() returns the expected value after performing an encryption operation (which will differ from the previous value for CBC and OFB modes), for various modes of AES. Do this both for the implicit fetch and explicit fetch paths, at the cost of a slightly more complicated switch statement. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/12233)
2020-06-20 11:42:29 +08:00
return 1;
#else
Always reset IV for CBC, OFB, and CFB mode on cipher context reinit This is necessary to keep compatibility with 1.1.1 implementation of the CBC, OFB, and CFB mode ciphers. Fixes #14704 Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14811)
2021-04-09 18:01:16 +08:00
case 5:
Add tests for new EVP_CIPHER_CTX IV accessors Test that EVP_CIPHER_CTX_get_iv() returns the same IV that was given at initialization time, and that EVP_CIPHER_CTX_get_iv_state() returns the expected value after performing an encryption operation (which will differ from the previous value for CBC and OFB modes), for various modes of AES. Do this both for the implicit fetch and explicit fetch paths, at the cost of a slightly more complicated switch statement. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/12233)
2020-06-20 11:42:29 +08:00
type = EVP_aes_128_ocb();
/* FALLTHROUGH */
Always reset IV for CBC, OFB, and CFB mode on cipher context reinit This is necessary to keep compatibility with 1.1.1 implementation of the CBC, OFB, and CFB mode ciphers. Fixes #14704 Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14811)
2021-04-09 18:01:16 +08:00
case 11:
Add tests for new EVP_CIPHER_CTX IV accessors Test that EVP_CIPHER_CTX_get_iv() returns the same IV that was given at initialization time, and that EVP_CIPHER_CTX_get_iv_state() returns the expected value after performing an encryption operation (which will differ from the previous value for CBC and OFB modes), for various modes of AES. Do this both for the implicit fetch and explicit fetch paths, at the cost of a slightly more complicated switch statement. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/12233)
2020-06-20 11:42:29 +08:00
type = (type != NULL) ? type :
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
EVP_CIPHER_fetch(testctx, "aes-128-ocb", testpropq);
Add tests for new EVP_CIPHER_CTX IV accessors Test that EVP_CIPHER_CTX_get_iv() returns the same IV that was given at initialization time, and that EVP_CIPHER_CTX_get_iv_state() returns the expected value after performing an encryption operation (which will differ from the previous value for CBC and OFB modes), for various modes of AES. Do this both for the implicit fetch and explicit fetch paths, at the cost of a slightly more complicated switch statement. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/12233)
2020-06-20 11:42:29 +08:00
ref_iv = ocb_state;
ref_len = sizeof(ocb_state);
break;
#endif
default:
return 0;
}
if (!TEST_ptr(type)
|| !TEST_ptr((ctx = EVP_CIPHER_CTX_new()))
|| !TEST_true(EVP_EncryptInit_ex(ctx, type, NULL, key, init_iv))
|| !TEST_true(EVP_EncryptUpdate(ctx, ciphertext, &len, msg,
(int)sizeof(msg)))
Rename EVP_CIPHER_CTX_get_iv and EVP_CIPHER_CTX_get_iv_state for clarity To clarify the purpose of these two calls rename them to EVP_CIPHER_CTX_get_original_iv and EVP_CIPHER_CTX_get_updated_iv. Also rename the OSSL_CIPHER_PARAM_IV_STATE to OSSL_CIPHER_PARAM_UPDATED_IV to better align with the function name. Fixes #13411 Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13870)
2021-01-14 22:19:46 +08:00
|| !TEST_true(EVP_CIPHER_CTX_get_original_iv(ctx, oiv, sizeof(oiv)))
|| !TEST_true(EVP_CIPHER_CTX_get_updated_iv(ctx, iv, sizeof(iv)))
Add tests for new EVP_CIPHER_CTX IV accessors Test that EVP_CIPHER_CTX_get_iv() returns the same IV that was given at initialization time, and that EVP_CIPHER_CTX_get_iv_state() returns the expected value after performing an encryption operation (which will differ from the previous value for CBC and OFB modes), for various modes of AES. Do this both for the implicit fetch and explicit fetch paths, at the cost of a slightly more complicated switch statement. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/12233)
2020-06-20 11:42:29 +08:00
|| !TEST_true(EVP_EncryptFinal_ex(ctx, ciphertext, &len)))
goto err;
Rename all getters to use get/get0 in name For functions that exist in 1.1.1 provide a simple aliases via #define. Fixes #15236 Functions with OSSL_DECODER_, OSSL_ENCODER_, OSSL_STORE_LOADER_, EVP_KEYEXCH_, EVP_KEM_, EVP_ASYM_CIPHER_, EVP_SIGNATURE_, EVP_KEYMGMT_, EVP_RAND_, EVP_MAC_, EVP_KDF_, EVP_PKEY_, EVP_MD_, and EVP_CIPHER_ prefixes are renamed. Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/15405)
2021-05-21 22:58:08 +08:00
ivlen = EVP_CIPHER_CTX_get_iv_length(ctx);
Add appropriate NULL checks in EVP_CIPHER api The EVP_CIPHER api currently assumes that calls made into several APIs have already initalized the cipher in a given context via a call to EVP_CipherInit[_ex[2]]. If that hasnt been done, instead of an error, the result is typically a SIGSEGV. Correct that by adding missing NULL checks in the apropriate apis prior to using ctx->cipher Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/22995)
2023-12-10 02:40:01 +08:00
if (!TEST_int_gt(ivlen, 0))
goto err;
Add tests for new EVP_CIPHER_CTX IV accessors Test that EVP_CIPHER_CTX_get_iv() returns the same IV that was given at initialization time, and that EVP_CIPHER_CTX_get_iv_state() returns the expected value after performing an encryption operation (which will differ from the previous value for CBC and OFB modes), for various modes of AES. Do this both for the implicit fetch and explicit fetch paths, at the cost of a slightly more complicated switch statement. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/12233)
2020-06-20 11:42:29 +08:00
if (!TEST_mem_eq(init_iv, ivlen, oiv, ivlen)
|| !TEST_mem_eq(ref_iv, ref_len, iv, ivlen))
goto err;
Always reset IV for CBC, OFB, and CFB mode on cipher context reinit This is necessary to keep compatibility with 1.1.1 implementation of the CBC, OFB, and CFB mode ciphers. Fixes #14704 Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14811)
2021-04-09 18:01:16 +08:00
/* CBC, OFB, and CFB modes: the updated iv must be reset after reinit */
if (!TEST_true(EVP_EncryptInit_ex(ctx, NULL, NULL, NULL, NULL))
|| !TEST_true(EVP_CIPHER_CTX_get_updated_iv(ctx, iv, sizeof(iv))))
goto err;
if (iv_reset) {
if (!TEST_mem_eq(init_iv, ivlen, iv, ivlen))
goto err;
} else {
if (!TEST_mem_eq(ref_iv, ivlen, iv, ivlen))
goto err;
}
Add tests for new EVP_CIPHER_CTX IV accessors Test that EVP_CIPHER_CTX_get_iv() returns the same IV that was given at initialization time, and that EVP_CIPHER_CTX_get_iv_state() returns the expected value after performing an encryption operation (which will differ from the previous value for CBC and OFB modes), for various modes of AES. Do this both for the implicit fetch and explicit fetch paths, at the cost of a slightly more complicated switch statement. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/12233)
2020-06-20 11:42:29 +08:00
ret = 1;
err:
EVP_CIPHER_CTX_free(ctx);
Always reset IV for CBC, OFB, and CFB mode on cipher context reinit This is necessary to keep compatibility with 1.1.1 implementation of the CBC, OFB, and CFB mode ciphers. Fixes #14704 Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14811)
2021-04-09 18:01:16 +08:00
if (idx >= 6)
Add tests for new EVP_CIPHER_CTX IV accessors Test that EVP_CIPHER_CTX_get_iv() returns the same IV that was given at initialization time, and that EVP_CIPHER_CTX_get_iv_state() returns the expected value after performing an encryption operation (which will differ from the previous value for CBC and OFB modes), for various modes of AES. Do this both for the implicit fetch and explicit fetch paths, at the cost of a slightly more complicated switch statement. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/12233)
2020-06-20 11:42:29 +08:00
EVP_CIPHER_free((EVP_CIPHER *)type);
return ret;
}
Add test for the IV handling of DES based ciphers Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/14882)
2021-04-15 17:53:42 +08:00
#ifndef OPENSSL_NO_DES
static int test_evp_iv_des(int idx)
{
int ret = 0;
EVP_CIPHER_CTX *ctx = NULL;
static const unsigned char key[24] = {
0x01, 0x23, 0x45, 0x67, 0x89, 0xab, 0xcd, 0xef,
0xf1, 0xe0, 0xd3, 0xc2, 0xb5, 0xa4, 0x97, 0x86,
0xfe, 0xdc, 0xba, 0x98, 0x76, 0x54, 0x32, 0x10
};
static const unsigned char init_iv[8] = {
0xfe, 0xdc, 0xba, 0x98, 0x76, 0x54, 0x32, 0x10
};
that open brace { should be on the previous line Found by running the checkpatch.pl Linux script to enforce coding style. Reviewed-by: Neil Horman <nhorman@openssl.org> Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/22097)
2023-09-14 02:32:54 +08:00
static const unsigned char msg[] = {
1, 2, 3, 4, 5, 6, 7, 8,
9, 10, 11, 12, 13, 14, 15, 16
};
Add test for the IV handling of DES based ciphers Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/14882)
2021-04-15 17:53:42 +08:00
unsigned char ciphertext[32], oiv[8], iv[8];
unsigned const char *ref_iv;
static const unsigned char cbc_state_des[8] = {
0x4f, 0xa3, 0x85, 0xcd, 0x8b, 0xf3, 0x06, 0x2a
};
static const unsigned char cbc_state_3des[8] = {
0x35, 0x27, 0x7d, 0x65, 0x6c, 0xfb, 0x50, 0xd9
};
static const unsigned char ofb_state_des[8] = {
0xa7, 0x0d, 0x1d, 0x45, 0xf9, 0x96, 0x3f, 0x2c
};
static const unsigned char ofb_state_3des[8] = {
0xab, 0x16, 0x24, 0xbb, 0x5b, 0xac, 0xed, 0x5e
};
static const unsigned char cfb_state_des[8] = {
0x91, 0xeb, 0x6d, 0x29, 0x4b, 0x08, 0xbd, 0x73
};
static const unsigned char cfb_state_3des[8] = {
0x34, 0xdd, 0xfb, 0x47, 0x33, 0x1c, 0x61, 0xf7
};
int len = sizeof(ciphertext);
size_t ivlen, ref_len;
EVP_CIPHER *type = NULL;
if (lgcyprov == NULL && idx < 3)
return TEST_skip("Test requires legacy provider to be loaded");
fix some code with obvious wrong coding style Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16918)
2021-10-26 15:16:18 +08:00
switch (idx) {
Add test for the IV handling of DES based ciphers Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/14882)
2021-04-15 17:53:42 +08:00
case 0:
type = EVP_CIPHER_fetch(testctx, "des-cbc", testpropq);
ref_iv = cbc_state_des;
ref_len = sizeof(cbc_state_des);
break;
case 1:
type = EVP_CIPHER_fetch(testctx, "des-ofb", testpropq);
ref_iv = ofb_state_des;
ref_len = sizeof(ofb_state_des);
break;
case 2:
type = EVP_CIPHER_fetch(testctx, "des-cfb", testpropq);
ref_iv = cfb_state_des;
ref_len = sizeof(cfb_state_des);
break;
case 3:
type = EVP_CIPHER_fetch(testctx, "des-ede3-cbc", testpropq);
ref_iv = cbc_state_3des;
ref_len = sizeof(cbc_state_3des);
break;
case 4:
type = EVP_CIPHER_fetch(testctx, "des-ede3-ofb", testpropq);
ref_iv = ofb_state_3des;
ref_len = sizeof(ofb_state_3des);
break;
case 5:
type = EVP_CIPHER_fetch(testctx, "des-ede3-cfb", testpropq);
ref_iv = cfb_state_3des;
ref_len = sizeof(cfb_state_3des);
break;
default:
return 0;
}
if (!TEST_ptr(type)
|| !TEST_ptr((ctx = EVP_CIPHER_CTX_new()))
|| !TEST_true(EVP_EncryptInit_ex(ctx, type, NULL, key, init_iv))
|| !TEST_true(EVP_EncryptUpdate(ctx, ciphertext, &len, msg,
(int)sizeof(msg)))
|| !TEST_true(EVP_CIPHER_CTX_get_original_iv(ctx, oiv, sizeof(oiv)))
|| !TEST_true(EVP_CIPHER_CTX_get_updated_iv(ctx, iv, sizeof(iv)))
|| !TEST_true(EVP_EncryptFinal_ex(ctx, ciphertext, &len)))
goto err;
Rename all getters to use get/get0 in name For functions that exist in 1.1.1 provide a simple aliases via #define. Fixes #15236 Functions with OSSL_DECODER_, OSSL_ENCODER_, OSSL_STORE_LOADER_, EVP_KEYEXCH_, EVP_KEM_, EVP_ASYM_CIPHER_, EVP_SIGNATURE_, EVP_KEYMGMT_, EVP_RAND_, EVP_MAC_, EVP_KDF_, EVP_PKEY_, EVP_MD_, and EVP_CIPHER_ prefixes are renamed. Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/15405)
2021-05-21 22:58:08 +08:00
ivlen = EVP_CIPHER_CTX_get_iv_length(ctx);
Add appropriate NULL checks in EVP_CIPHER api The EVP_CIPHER api currently assumes that calls made into several APIs have already initalized the cipher in a given context via a call to EVP_CipherInit[_ex[2]]. If that hasnt been done, instead of an error, the result is typically a SIGSEGV. Correct that by adding missing NULL checks in the apropriate apis prior to using ctx->cipher Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/22995)
2023-12-10 02:40:01 +08:00
if (!TEST_int_gt(ivlen, 0))
goto err;
Add test for the IV handling of DES based ciphers Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/14882)
2021-04-15 17:53:42 +08:00
if (!TEST_mem_eq(init_iv, ivlen, oiv, ivlen)
|| !TEST_mem_eq(ref_iv, ref_len, iv, ivlen))
goto err;
if (!TEST_true(EVP_EncryptInit_ex(ctx, NULL, NULL, NULL, NULL))
|| !TEST_true(EVP_CIPHER_CTX_get_updated_iv(ctx, iv, sizeof(iv))))
goto err;
if (!TEST_mem_eq(init_iv, ivlen, iv, ivlen))
goto err;
ret = 1;
err:
EVP_CIPHER_CTX_free(ctx);
EVP_CIPHER_free(type);
return ret;
}
#endif
Test the default key length of the Blowfish ciphers Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/18362)
2022-05-20 22:31:33 +08:00
#ifndef OPENSSL_NO_BF
static int test_evp_bf_default_keylen(int idx)
{
int ret = 0;
static const char *algos[4] = {
"bf-ecb", "bf-cbc", "bf-cfb", "bf-ofb"
};
int ivlen[4] = { 0, 8, 8, 8 };
EVP_CIPHER *cipher = NULL;
if (lgcyprov == NULL)
return TEST_skip("Test requires legacy provider to be loaded");
if (!TEST_ptr(cipher = EVP_CIPHER_fetch(testctx, algos[idx], testpropq))
|| !TEST_int_eq(EVP_CIPHER_get_key_length(cipher), 16)
|| !TEST_int_eq(EVP_CIPHER_get_iv_length(cipher), ivlen[idx]))
goto err;
ret = 1;
err:
EVP_CIPHER_free(cipher);
return ret;
}
#endif
Add test for EC pubkey export/import There seems to be an issue with i2d_provided() in i2d_evp.c that causes us to fail to construct a valid chain of encoders for the "type-specific" output when it's an EC pubkey. This test is designed to exercise that codepath for a variety of curves. Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14291)
2021-01-26 04:19:16 +08:00
#ifndef OPENSSL_NO_EC
Runchecker: fix no-ec2m build which was trying to validate the e2cm curves The evp_extra_test program was trying to validate these curves when they were not build. Fixes #14959 Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14977)
2021-04-22 09:50:15 +08:00
static int ecpub_nids[] = {
NID_brainpoolP256r1, NID_X9_62_prime256v1,
NID_secp384r1, NID_secp521r1,
# ifndef OPENSSL_NO_EC2M
NID_sect233k1, NID_sect233r1, NID_sect283r1,
Add test for EC pubkey export/import There seems to be an issue with i2d_provided() in i2d_evp.c that causes us to fail to construct a valid chain of encoders for the "type-specific" output when it's an EC pubkey. This test is designed to exercise that codepath for a variety of curves. Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14291)
2021-01-26 04:19:16 +08:00
NID_sect409k1, NID_sect409r1, NID_sect571k1, NID_sect571r1,
Runchecker: fix no-ec2m build which was trying to validate the e2cm curves The evp_extra_test program was trying to validate these curves when they were not build. Fixes #14959 Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14977)
2021-04-22 09:50:15 +08:00
# endif
NID_brainpoolP384r1, NID_brainpoolP512r1
};
Add test for EC pubkey export/import There seems to be an issue with i2d_provided() in i2d_evp.c that causes us to fail to construct a valid chain of encoders for the "type-specific" output when it's an EC pubkey. This test is designed to exercise that codepath for a variety of curves. Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14291)
2021-01-26 04:19:16 +08:00
static int test_ecpub(int idx)
{
test_ecpub: verify returned length after encoding Save the length we got from querying how much space was needed, and check that the actual encoding call returned the same length. Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14291)
2021-02-20 05:46:49 +08:00
int ret = 0, len, savelen;
Add test for EC pubkey export/import There seems to be an issue with i2d_provided() in i2d_evp.c that causes us to fail to construct a valid chain of encoders for the "type-specific" output when it's an EC pubkey. This test is designed to exercise that codepath for a variety of curves. Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14291)
2021-01-26 04:19:16 +08:00
int nid;
unsigned char buf[1024];
unsigned char *p;
EVP_PKEY *pkey = NULL;
EVP_PKEY_CTX *ctx = NULL;
test_ecpub: test that we can decode the DER we encoded We should be able to round-trip through the encoded DER form of the EC public key and get back something that compares as equal to the original key. Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14291)
2021-02-20 05:20:00 +08:00
# ifndef OPENSSL_NO_DEPRECATED_3_0
const unsigned char *q;
EVP_PKEY *pkey2 = NULL;
EC_KEY *ec = NULL;
# endif
Add test for EC pubkey export/import There seems to be an issue with i2d_provided() in i2d_evp.c that causes us to fail to construct a valid chain of encoders for the "type-specific" output when it's an EC pubkey. This test is designed to exercise that codepath for a variety of curves. Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14291)
2021-01-26 04:19:16 +08:00
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
if (nullprov != NULL)
return TEST_skip("Test does not support a non-default library context");
Add test for EC pubkey export/import There seems to be an issue with i2d_provided() in i2d_evp.c that causes us to fail to construct a valid chain of encoders for the "type-specific" output when it's an EC pubkey. This test is designed to exercise that codepath for a variety of curves. Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14291)
2021-01-26 04:19:16 +08:00
nid = ecpub_nids[idx];
ctx = EVP_PKEY_CTX_new_id(EVP_PKEY_EC, NULL);
if (!TEST_ptr(ctx)
EVP_PKEY_keygen_init: fix return check Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/17031)
2021-11-15 00:05:04 +08:00
|| !TEST_int_gt(EVP_PKEY_keygen_init(ctx), 0)
Fix the checks of EVP_PKEY_CTX_set/get_* functions Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/18399)
2022-05-24 23:57:33 +08:00
|| !TEST_int_gt(EVP_PKEY_CTX_set_ec_paramgen_curve_nid(ctx, nid), 0)
Add test for EC pubkey export/import There seems to be an issue with i2d_provided() in i2d_evp.c that causes us to fail to construct a valid chain of encoders for the "type-specific" output when it's an EC pubkey. This test is designed to exercise that codepath for a variety of curves. Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14291)
2021-01-26 04:19:16 +08:00
|| !TEST_true(EVP_PKEY_keygen(ctx, &pkey)))
goto done;
len = i2d_PublicKey(pkey, NULL);
test_ecpub: verify returned length after encoding Save the length we got from querying how much space was needed, and check that the actual encoding call returned the same length. Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14291)
2021-02-20 05:46:49 +08:00
savelen = len;
Add test for EC pubkey export/import There seems to be an issue with i2d_provided() in i2d_evp.c that causes us to fail to construct a valid chain of encoders for the "type-specific" output when it's an EC pubkey. This test is designed to exercise that codepath for a variety of curves. Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14291)
2021-01-26 04:19:16 +08:00
if (!TEST_int_ge(len, 1)
|| !TEST_int_lt(len, 1024))
goto done;
p = buf;
len = i2d_PublicKey(pkey, &p);
test_ecpub: verify returned length after encoding Save the length we got from querying how much space was needed, and check that the actual encoding call returned the same length. Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14291)
2021-02-20 05:46:49 +08:00
if (!TEST_int_ge(len, 1)
|| !TEST_int_eq(len, savelen))
Add test for EC pubkey export/import There seems to be an issue with i2d_provided() in i2d_evp.c that causes us to fail to construct a valid chain of encoders for the "type-specific" output when it's an EC pubkey. This test is designed to exercise that codepath for a variety of curves. Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14291)
2021-01-26 04:19:16 +08:00
goto done;
test_ecpub: test that we can decode the DER we encoded We should be able to round-trip through the encoded DER form of the EC public key and get back something that compares as equal to the original key. Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14291)
2021-02-20 05:20:00 +08:00
# ifndef OPENSSL_NO_DEPRECATED_3_0
/* Now try to decode the just-created DER. */
q = buf;
if (!TEST_ptr((pkey2 = EVP_PKEY_new()))
|| !TEST_ptr((ec = EC_KEY_new_by_curve_name(nid)))
|| !TEST_true(EVP_PKEY_assign_EC_KEY(pkey2, ec)))
goto done;
/* EC_KEY ownership transferred */
ec = NULL;
if (!TEST_ptr(d2i_PublicKey(EVP_PKEY_EC, &pkey2, &q, savelen)))
goto done;
/* The keys should match. */
Deprecate EVP_PKEY_cmp() and EVP_PKEY_cmp_parameters(). The replacement functions EVP_PKEY_eq() and EVP_PKEY_parameters_eq() already exist. 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/14997)
2021-04-23 08:53:03 +08:00
if (!TEST_int_eq(EVP_PKEY_eq(pkey, pkey2), 1))
test_ecpub: test that we can decode the DER we encoded We should be able to round-trip through the encoded DER form of the EC public key and get back something that compares as equal to the original key. Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14291)
2021-02-20 05:20:00 +08:00
goto done;
# endif
Add test for EC pubkey export/import There seems to be an issue with i2d_provided() in i2d_evp.c that causes us to fail to construct a valid chain of encoders for the "type-specific" output when it's an EC pubkey. This test is designed to exercise that codepath for a variety of curves. Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14291)
2021-01-26 04:19:16 +08:00
ret = 1;
done:
EVP_PKEY_CTX_free(ctx);
EVP_PKEY_free(pkey);
test_ecpub: test that we can decode the DER we encoded We should be able to round-trip through the encoded DER form of the EC public key and get back something that compares as equal to the original key. Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14291)
2021-02-20 05:20:00 +08:00
# ifndef OPENSSL_NO_DEPRECATED_3_0
EVP_PKEY_free(pkey2);
EC_KEY_free(ec);
# endif
Add test for EC pubkey export/import There seems to be an issue with i2d_provided() in i2d_evp.c that causes us to fail to construct a valid chain of encoders for the "type-specific" output when it's an EC pubkey. This test is designed to exercise that codepath for a variety of curves. Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14291)
2021-01-26 04:19:16 +08:00
return ret;
}
#endif
EVP: fix keygen for EVP_PKEY_RSA_PSS This commit attempts to fix the an issue when generating a key of type EVP_PKEY_RSA_PSS. Currently, EVP_PKEY_CTX_set_rsa_keygen_bits will return -1 if the key id is not of type EVP_PKEY_RSA. This commit adds EVP_PKEY_RSA_PSS to also be accepted. The macro EVP_PKEY_CTX_set_rsa_pss_keygen_md si converted into a function and it is now called in legacy_ctrl_to_param. Fixes #12384 Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13099)
2020-10-09 12:07:43 +08:00
static int test_EVP_rsa_pss_with_keygen_bits(void)
{
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
int ret = 0;
EVP_PKEY_CTX *ctx = NULL;
EVP_PKEY *pkey = NULL;
EVP_MD *md;
EVP: fix keygen for EVP_PKEY_RSA_PSS This commit attempts to fix the an issue when generating a key of type EVP_PKEY_RSA_PSS. Currently, EVP_PKEY_CTX_set_rsa_keygen_bits will return -1 if the key id is not of type EVP_PKEY_RSA. This commit adds EVP_PKEY_RSA_PSS to also be accepted. The macro EVP_PKEY_CTX_set_rsa_pss_keygen_md si converted into a function and it is now called in legacy_ctrl_to_param. Fixes #12384 Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13099)
2020-10-09 12:07:43 +08:00
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
md = EVP_MD_fetch(testctx, "sha256", testpropq);
ret = TEST_ptr(md)
Fix the checks of EVP_PKEY_CTX_set/get_* functions Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/18399)
2022-05-24 23:57:33 +08:00
&& TEST_ptr((ctx = EVP_PKEY_CTX_new_from_name(testctx, "RSA-PSS", testpropq)))
EVP_PKEY_keygen_init: fix return check Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/17031)
2021-11-15 00:05:04 +08:00
&& TEST_int_gt(EVP_PKEY_keygen_init(ctx), 0)
EVP: fix keygen for EVP_PKEY_RSA_PSS This commit attempts to fix the an issue when generating a key of type EVP_PKEY_RSA_PSS. Currently, EVP_PKEY_CTX_set_rsa_keygen_bits will return -1 if the key id is not of type EVP_PKEY_RSA. This commit adds EVP_PKEY_RSA_PSS to also be accepted. The macro EVP_PKEY_CTX_set_rsa_pss_keygen_md si converted into a function and it is now called in legacy_ctrl_to_param. Fixes #12384 Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13099)
2020-10-09 12:07:43 +08:00
&& TEST_int_gt(EVP_PKEY_CTX_set_rsa_keygen_bits(ctx, 512), 0)
Fix the checks of EVP_PKEY_CTX_set/get_* functions Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/18399)
2022-05-24 23:57:33 +08:00
&& TEST_int_gt(EVP_PKEY_CTX_set_rsa_pss_keygen_md(ctx, md), 0)
EVP: fix keygen for EVP_PKEY_RSA_PSS This commit attempts to fix the an issue when generating a key of type EVP_PKEY_RSA_PSS. Currently, EVP_PKEY_CTX_set_rsa_keygen_bits will return -1 if the key id is not of type EVP_PKEY_RSA. This commit adds EVP_PKEY_RSA_PSS to also be accepted. The macro EVP_PKEY_CTX_set_rsa_pss_keygen_md si converted into a function and it is now called in legacy_ctrl_to_param. Fixes #12384 Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13099)
2020-10-09 12:07:43 +08:00
&& TEST_true(EVP_PKEY_keygen(ctx, &pkey));
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
EVP_MD_free(md);
EVP: fix keygen for EVP_PKEY_RSA_PSS This commit attempts to fix the an issue when generating a key of type EVP_PKEY_RSA_PSS. Currently, EVP_PKEY_CTX_set_rsa_keygen_bits will return -1 if the key id is not of type EVP_PKEY_RSA. This commit adds EVP_PKEY_RSA_PSS to also be accepted. The macro EVP_PKEY_CTX_set_rsa_pss_keygen_md si converted into a function and it is now called in legacy_ctrl_to_param. Fixes #12384 Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13099)
2020-10-09 12:07:43 +08:00
EVP_PKEY_free(pkey);
EVP_PKEY_CTX_free(ctx);
return ret;
}
Fix EVP_PKEY_CTX_get_rsa_pss_saltlen() not returning a value When an integer value was specified, it was not being passed back via the orig_p2 weirdness. Regression test included. Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/17136)
2021-11-25 23:49:26 +08:00
static int test_EVP_rsa_pss_set_saltlen(void)
{
int ret = 0;
EVP_PKEY *pkey = NULL;
EVP_PKEY_CTX *pkey_ctx = NULL;
EVP_MD *sha256 = NULL;
EVP_MD_CTX *sha256_ctx = NULL;
int saltlen = 9999; /* buggy EVP_PKEY_CTX_get_rsa_pss_saltlen() didn't update this */
const int test_value = 32;
ret = TEST_ptr(pkey = load_example_rsa_key())
&& TEST_ptr(sha256 = EVP_MD_fetch(testctx, "sha256", NULL))
&& TEST_ptr(sha256_ctx = EVP_MD_CTX_new())
&& TEST_true(EVP_DigestSignInit(sha256_ctx, &pkey_ctx, sha256, NULL, pkey))
&& TEST_true(EVP_PKEY_CTX_set_rsa_padding(pkey_ctx, RSA_PKCS1_PSS_PADDING))
Fix the checks of EVP_PKEY_CTX_get/set_rsa_pss_saltlen Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/18399)
2022-05-24 23:38:39 +08:00
&& TEST_int_gt(EVP_PKEY_CTX_set_rsa_pss_saltlen(pkey_ctx, test_value), 0)
&& TEST_int_gt(EVP_PKEY_CTX_get_rsa_pss_saltlen(pkey_ctx, &saltlen), 0)
Fix EVP_PKEY_CTX_get_rsa_pss_saltlen() not returning a value When an integer value was specified, it was not being passed back via the orig_p2 weirdness. Regression test included. Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/17136)
2021-11-25 23:49:26 +08:00
&& TEST_int_eq(saltlen, test_value);
EVP_MD_CTX_free(sha256_ctx);
EVP_PKEY_free(pkey);
EVP_MD_free(sha256);
return ret;
}
RSA decoder should check also sanity of p, q, e, d ... with respect to n This issue has been discovered by osss-fuzzer [1]. The test function decodes RSA key created by fuzzer and calls EVP_PKEY_pairwise_check() which proceeds to ossl_bn_miller_rabin_is_prime() check which takes too long exceeding timeout (45secs). The idea is to fix OSSL_DECODER_from_data() code path so invalid RSA keys will be refused. [1] https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=69134 Test case generated by the fuzzer is added. Reviewed-by: Neil Horman <nhorman@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/25190)
2024-08-15 02:07:29 +08:00
static int test_EVP_rsa_invalid_key(void)
{
int ret = 0;
EVP_PKEY *pkey = NULL;
pkey = load_example_key("RSA", kInvalidRSAKeyDER, sizeof(kInvalidRSAKeyDER));
/* we expect to fail to load bogus key */
ret = !TEST_ptr(pkey);
EVP_PKEY_free(pkey);
return ret;
}
Add a test for a names_do_all function Make sure that if we change the namemap part way through calling a names_do_all function it still works. Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14250)
2021-02-20 01:47:21 +08:00
static int success = 1;
static void md_names(const char *name, void *vctx)
{
OSSL_LIB_CTX *ctx = (OSSL_LIB_CTX *)vctx;
/* Force a namemap update */
EVP_CIPHER *aes128 = EVP_CIPHER_fetch(ctx, "AES-128-CBC", NULL);
if (!TEST_ptr(aes128))
success = 0;
EVP_CIPHER_free(aes128);
}
/*
* Test that changing the namemap in a user callback works in a names_do_all
* function.
*/
static int test_names_do_all(void)
{
/* We use a custom libctx so that we know the state of the namemap */
OSSL_LIB_CTX *ctx = OSSL_LIB_CTX_new();
EVP_MD *sha256 = NULL;
int testresult = 0;
if (!TEST_ptr(ctx))
goto err;
sha256 = EVP_MD_fetch(ctx, "SHA2-256", NULL);
if (!TEST_ptr(sha256))
goto err;
/*
* We loop through all the names for a given digest. This should still work
* even if the namemap changes part way through.
*/
if (!TEST_true(EVP_MD_names_do_all(sha256, md_names, ctx)))
goto err;
if (!TEST_true(success))
goto err;
testresult = 1;
err:
EVP_MD_free(sha256);
OSSL_LIB_CTX_free(ctx);
return testresult;
}
EVP: fix keygen for EVP_PKEY_RSA_PSS This commit attempts to fix the an issue when generating a key of type EVP_PKEY_RSA_PSS. Currently, EVP_PKEY_CTX_set_rsa_keygen_bits will return -1 if the key id is not of type EVP_PKEY_RSA. This commit adds EVP_PKEY_RSA_PSS to also be accepted. The macro EVP_PKEY_CTX_set_rsa_pss_keygen_md si converted into a function and it is now called in legacy_ctrl_to_param. Fixes #12384 Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13099)
2020-10-09 12:07:43 +08:00
Test EVP_CipherInit sequences and resets Various EVP_CipherInit sequences including partial inits and initializations with different "enc" flags caused problems on s390x. Similarly, cipher reinitialization and especially GCM reinitialization with different tag length led to wrong results. Add some unit tests to cover these rather exotic use cases. Signed-off-by: Juergen Christ <jchrist@linux.ibm.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/15521)
2021-06-03 01:33:50 +08:00
typedef struct {
const char *cipher;
const unsigned char *key;
const unsigned char *iv;
const unsigned char *input;
const unsigned char *expected;
const unsigned char *tag;
size_t ivlen; /* 0 if we do not need to set a specific IV len */
size_t inlen;
size_t expectedlen;
size_t taglen;
int keyfirst;
int initenc;
int finalenc;
} EVP_INIT_TEST_st;
static const EVP_INIT_TEST_st evp_init_tests[] = {
{
"aes-128-cfb", kCFBDefaultKey, iCFBIV, cfbPlaintext,
cfbCiphertext, NULL, 0, sizeof(cfbPlaintext), sizeof(cfbCiphertext),
0, 1, 0, 1
},
{
"aes-256-gcm", kGCMDefaultKey, iGCMDefaultIV, gcmDefaultPlaintext,
gcmDefaultCiphertext, gcmDefaultTag, sizeof(iGCMDefaultIV),
sizeof(gcmDefaultPlaintext), sizeof(gcmDefaultCiphertext),
sizeof(gcmDefaultTag), 1, 0, 1
},
{
"aes-128-cfb", kCFBDefaultKey, iCFBIV, cfbPlaintext,
cfbCiphertext, NULL, 0, sizeof(cfbPlaintext), sizeof(cfbCiphertext),
0, 0, 0, 1
},
{
"aes-256-gcm", kGCMDefaultKey, iGCMDefaultIV, gcmDefaultPlaintext,
gcmDefaultCiphertext, gcmDefaultTag, sizeof(iGCMDefaultIV),
sizeof(gcmDefaultPlaintext), sizeof(gcmDefaultCiphertext),
sizeof(gcmDefaultTag), 0, 0, 1
},
{
"aes-128-cfb", kCFBDefaultKey, iCFBIV, cfbCiphertext,
cfbPlaintext, NULL, 0, sizeof(cfbCiphertext), sizeof(cfbPlaintext),
0, 1, 1, 0
},
{
"aes-256-gcm", kGCMDefaultKey, iGCMDefaultIV, gcmDefaultCiphertext,
gcmDefaultPlaintext, gcmDefaultTag, sizeof(iGCMDefaultIV),
sizeof(gcmDefaultCiphertext), sizeof(gcmDefaultPlaintext),
sizeof(gcmDefaultTag), 1, 1, 0
},
{
"aes-128-cfb", kCFBDefaultKey, iCFBIV, cfbCiphertext,
cfbPlaintext, NULL, 0, sizeof(cfbCiphertext), sizeof(cfbPlaintext),
0, 0, 1, 0
},
{
"aes-256-gcm", kGCMDefaultKey, iGCMDefaultIV, gcmDefaultCiphertext,
gcmDefaultPlaintext, gcmDefaultTag, sizeof(iGCMDefaultIV),
sizeof(gcmDefaultCiphertext), sizeof(gcmDefaultPlaintext),
sizeof(gcmDefaultTag), 0, 1, 0
}
};
Add tests for re-using cipher contexts Add test case for re-using a cipher context with the same key, iv and cipher. It detects, if the hardware-specific cipher context is reset correctly, like reported in issue #23175. This test has encrypt and decrypt iterations for cfb128 and ofb128. All iteations use the same key, iv and plaintext. Signed-off-by: Holger Dengler <dengler@linux.ibm.com> Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/23201)
2024-01-05 02:25:08 +08:00
/* use same key, iv and plaintext for cfb and ofb */
static const EVP_INIT_TEST_st evp_reinit_tests[] = {
{
"aes-128-cfb", kCFBDefaultKey, iCFBIV, cfbPlaintext_partial,
cfbCiphertext_partial, NULL, 0, sizeof(cfbPlaintext_partial),
sizeof(cfbCiphertext_partial), 0, 0, 1, 0
},
{
"aes-128-cfb", kCFBDefaultKey, iCFBIV, cfbCiphertext_partial,
cfbPlaintext_partial, NULL, 0, sizeof(cfbCiphertext_partial),
sizeof(cfbPlaintext_partial), 0, 0, 0, 0
},
{
"aes-128-ofb", kCFBDefaultKey, iCFBIV, cfbPlaintext_partial,
ofbCiphertext_partial, NULL, 0, sizeof(cfbPlaintext_partial),
sizeof(ofbCiphertext_partial), 0, 0, 1, 0
},
{
"aes-128-ofb", kCFBDefaultKey, iCFBIV, ofbCiphertext_partial,
cfbPlaintext_partial, NULL, 0, sizeof(ofbCiphertext_partial),
sizeof(cfbPlaintext_partial), 0, 0, 0, 0
},
};
Test EVP_CipherInit sequences and resets Various EVP_CipherInit sequences including partial inits and initializations with different "enc" flags caused problems on s390x. Similarly, cipher reinitialization and especially GCM reinitialization with different tag length led to wrong results. Add some unit tests to cover these rather exotic use cases. Signed-off-by: Juergen Christ <jchrist@linux.ibm.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/15521)
2021-06-03 01:33:50 +08:00
static int evp_init_seq_set_iv(EVP_CIPHER_CTX *ctx, const EVP_INIT_TEST_st *t)
{
int res = 0;
fix some code with obvious wrong coding style Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16918)
2021-10-26 15:16:18 +08:00
Test EVP_CipherInit sequences and resets Various EVP_CipherInit sequences including partial inits and initializations with different "enc" flags caused problems on s390x. Similarly, cipher reinitialization and especially GCM reinitialization with different tag length led to wrong results. Add some unit tests to cover these rather exotic use cases. Signed-off-by: Juergen Christ <jchrist@linux.ibm.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/15521)
2021-06-03 01:33:50 +08:00
if (t->ivlen != 0) {
Fix check of EVP_CIPHER_CTX_ctrl Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/18368)
2022-05-21 16:17:23 +08:00
if (!TEST_int_gt(EVP_CIPHER_CTX_ctrl(ctx, EVP_CTRL_GCM_SET_IVLEN, t->ivlen, NULL), 0))
Test EVP_CipherInit sequences and resets Various EVP_CipherInit sequences including partial inits and initializations with different "enc" flags caused problems on s390x. Similarly, cipher reinitialization and especially GCM reinitialization with different tag length led to wrong results. Add some unit tests to cover these rather exotic use cases. Signed-off-by: Juergen Christ <jchrist@linux.ibm.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/15521)
2021-06-03 01:33:50 +08:00
goto err;
}
if (!TEST_true(EVP_CipherInit_ex(ctx, NULL, NULL, NULL, t->iv, -1)))
goto err;
res = 1;
err:
return res;
}
/*
* Test step-wise cipher initialization via EVP_CipherInit_ex where the
* arguments are given one at a time and a final adjustment to the enc
* parameter sets the correct operation.
*/
static int test_evp_init_seq(int idx)
{
int outlen1, outlen2;
int testresult = 0;
unsigned char outbuf[1024];
unsigned char tag[16];
const EVP_INIT_TEST_st *t = &evp_init_tests[idx];
EVP_CIPHER_CTX *ctx = NULL;
EVP_CIPHER *type = NULL;
size_t taglen = sizeof(tag);
char *errmsg = NULL;
ctx = EVP_CIPHER_CTX_new();
if (ctx == NULL) {
errmsg = "CTX_ALLOC";
goto err;
}
if (!TEST_ptr(type = EVP_CIPHER_fetch(testctx, t->cipher, testpropq))) {
errmsg = "CIPHER_FETCH";
goto err;
}
if (!TEST_true(EVP_CipherInit_ex(ctx, type, NULL, NULL, NULL, t->initenc))) {
errmsg = "EMPTY_ENC_INIT";
goto err;
}
if (!TEST_true(EVP_CIPHER_CTX_set_padding(ctx, 0))) {
errmsg = "PADDING";
goto err;
}
if (t->keyfirst && !TEST_true(EVP_CipherInit_ex(ctx, NULL, NULL, t->key, NULL, -1))) {
errmsg = "KEY_INIT (before iv)";
goto err;
}
if (!evp_init_seq_set_iv(ctx, t)) {
errmsg = "IV_INIT";
goto err;
}
if (t->keyfirst == 0 && !TEST_true(EVP_CipherInit_ex(ctx, NULL, NULL, t->key, NULL, -1))) {
errmsg = "KEY_INIT (after iv)";
goto err;
}
if (!TEST_true(EVP_CipherInit_ex(ctx, NULL, NULL, NULL, NULL, t->finalenc))) {
errmsg = "FINAL_ENC_INIT";
goto err;
}
if (!TEST_true(EVP_CipherUpdate(ctx, outbuf, &outlen1, t->input, t->inlen))) {
errmsg = "CIPHER_UPDATE";
goto err;
}
if (t->finalenc == 0 && t->tag != NULL) {
/* Set expected tag */
Fix check of EVP_CIPHER_CTX_ctrl Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/18368)
2022-05-21 16:17:23 +08:00
if (!TEST_int_gt(EVP_CIPHER_CTX_ctrl(ctx, EVP_CTRL_AEAD_SET_TAG,
t->taglen, (void *)t->tag), 0)) {
Test EVP_CipherInit sequences and resets Various EVP_CipherInit sequences including partial inits and initializations with different "enc" flags caused problems on s390x. Similarly, cipher reinitialization and especially GCM reinitialization with different tag length led to wrong results. Add some unit tests to cover these rather exotic use cases. Signed-off-by: Juergen Christ <jchrist@linux.ibm.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/15521)
2021-06-03 01:33:50 +08:00
errmsg = "SET_TAG";
goto err;
}
}
if (!TEST_true(EVP_CipherFinal_ex(ctx, outbuf + outlen1, &outlen2))) {
errmsg = "CIPHER_FINAL";
goto err;
}
if (!TEST_mem_eq(t->expected, t->expectedlen, outbuf, outlen1 + outlen2)) {
errmsg = "WRONG_RESULT";
goto err;
}
if (t->finalenc != 0 && t->tag != NULL) {
Fix check of EVP_CIPHER_CTX_ctrl Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/18368)
2022-05-21 16:17:23 +08:00
if (!TEST_int_gt(EVP_CIPHER_CTX_ctrl(ctx, EVP_CTRL_AEAD_GET_TAG, taglen, tag), 0)) {
Test EVP_CipherInit sequences and resets Various EVP_CipherInit sequences including partial inits and initializations with different "enc" flags caused problems on s390x. Similarly, cipher reinitialization and especially GCM reinitialization with different tag length led to wrong results. Add some unit tests to cover these rather exotic use cases. Signed-off-by: Juergen Christ <jchrist@linux.ibm.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/15521)
2021-06-03 01:33:50 +08:00
errmsg = "GET_TAG";
goto err;
}
if (!TEST_mem_eq(t->tag, t->taglen, tag, taglen)) {
errmsg = "TAG_ERROR";
goto err;
}
}
testresult = 1;
err:
if (errmsg != NULL)
TEST_info("evp_init_test %d: %s", idx, errmsg);
EVP_CIPHER_CTX_free(ctx);
EVP_CIPHER_free(type);
return testresult;
}
Add tests for re-using cipher contexts Add test case for re-using a cipher context with the same key, iv and cipher. It detects, if the hardware-specific cipher context is reset correctly, like reported in issue #23175. This test has encrypt and decrypt iterations for cfb128 and ofb128. All iteations use the same key, iv and plaintext. Signed-off-by: Holger Dengler <dengler@linux.ibm.com> Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/23201)
2024-01-05 02:25:08 +08:00
/*
* Test re-initialization of cipher context without changing key or iv.
* The result of both iteration should be the same.
*/
static int test_evp_reinit_seq(int idx)
{
int outlen1, outlen2, outlen_final;
int testresult = 0;
unsigned char outbuf1[1024];
unsigned char outbuf2[1024];
const EVP_INIT_TEST_st *t = &evp_reinit_tests[idx];
EVP_CIPHER_CTX *ctx = NULL;
EVP_CIPHER *type = NULL;
if (!TEST_ptr(ctx = EVP_CIPHER_CTX_new())
|| !TEST_ptr(type = EVP_CIPHER_fetch(testctx, t->cipher, testpropq))
/* setup cipher context */
|| !TEST_true(EVP_CipherInit_ex2(ctx, type, t->key, t->iv, t->initenc, NULL))
/* first iteration */
|| !TEST_true(EVP_CipherUpdate(ctx, outbuf1, &outlen1, t->input, t->inlen))
|| !TEST_true(EVP_CipherFinal_ex(ctx, outbuf1, &outlen_final))
/* check test results iteration 1 */
|| !TEST_mem_eq(t->expected, t->expectedlen, outbuf1, outlen1 + outlen_final)
/* now re-init the context (same cipher, key and iv) */
|| !TEST_true(EVP_CipherInit_ex2(ctx, NULL, NULL, NULL, -1, NULL))
/* second iteration */
|| !TEST_true(EVP_CipherUpdate(ctx, outbuf2, &outlen2, t->input, t->inlen))
|| !TEST_true(EVP_CipherFinal_ex(ctx, outbuf2, &outlen_final))
/* check test results iteration 2 */
|| !TEST_mem_eq(t->expected, t->expectedlen, outbuf2, outlen2 + outlen_final))
goto err;
testresult = 1;
err:
EVP_CIPHER_CTX_free(ctx);
EVP_CIPHER_free(type);
return testresult;
}
Test EVP_CipherInit sequences and resets Various EVP_CipherInit sequences including partial inits and initializations with different "enc" flags caused problems on s390x. Similarly, cipher reinitialization and especially GCM reinitialization with different tag length led to wrong results. Add some unit tests to cover these rather exotic use cases. Signed-off-by: Juergen Christ <jchrist@linux.ibm.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/15521)
2021-06-03 01:33:50 +08:00
typedef struct {
const unsigned char *input;
const unsigned char *expected;
size_t inlen;
size_t expectedlen;
int enc;
} EVP_RESET_TEST_st;
static const EVP_RESET_TEST_st evp_reset_tests[] = {
{
cfbPlaintext, cfbCiphertext,
sizeof(cfbPlaintext), sizeof(cfbCiphertext), 1
},
{
cfbCiphertext, cfbPlaintext,
sizeof(cfbCiphertext), sizeof(cfbPlaintext), 0
}
};
/*
* Test a reset of a cipher via EVP_CipherInit_ex after the cipher has already
* been used.
*/
static int test_evp_reset(int idx)
{
const EVP_RESET_TEST_st *t = &evp_reset_tests[idx];
int outlen1, outlen2;
int testresult = 0;
unsigned char outbuf[1024];
EVP_CIPHER_CTX *ctx = NULL;
EVP_CIPHER *type = NULL;
char *errmsg = NULL;
if (!TEST_ptr(ctx = EVP_CIPHER_CTX_new())) {
errmsg = "CTX_ALLOC";
goto err;
}
if (!TEST_ptr(type = EVP_CIPHER_fetch(testctx, "aes-128-cfb", testpropq))) {
errmsg = "CIPHER_FETCH";
goto err;
}
if (!TEST_true(EVP_CipherInit_ex(ctx, type, NULL, kCFBDefaultKey, iCFBIV, t->enc))) {
errmsg = "CIPHER_INIT";
goto err;
}
if (!TEST_true(EVP_CIPHER_CTX_set_padding(ctx, 0))) {
errmsg = "PADDING";
goto err;
}
if (!TEST_true(EVP_CipherUpdate(ctx, outbuf, &outlen1, t->input, t->inlen))) {
errmsg = "CIPHER_UPDATE";
goto err;
}
if (!TEST_true(EVP_CipherFinal_ex(ctx, outbuf + outlen1, &outlen2))) {
errmsg = "CIPHER_FINAL";
goto err;
}
if (!TEST_mem_eq(t->expected, t->expectedlen, outbuf, outlen1 + outlen2)) {
errmsg = "WRONG_RESULT";
goto err;
}
if (!TEST_true(EVP_CipherInit_ex(ctx, NULL, NULL, NULL, NULL, -1))) {
errmsg = "CIPHER_REINIT";
goto err;
}
if (!TEST_true(EVP_CipherUpdate(ctx, outbuf, &outlen1, t->input, t->inlen))) {
errmsg = "CIPHER_UPDATE (reinit)";
goto err;
}
if (!TEST_true(EVP_CipherFinal_ex(ctx, outbuf + outlen1, &outlen2))) {
errmsg = "CIPHER_FINAL (reinit)";
goto err;
}
if (!TEST_mem_eq(t->expected, t->expectedlen, outbuf, outlen1 + outlen2)) {
errmsg = "WRONG_RESULT (reinit)";
goto err;
}
testresult = 1;
err:
if (errmsg != NULL)
TEST_info("test_evp_reset %d: %s", idx, errmsg);
EVP_CIPHER_CTX_free(ctx);
EVP_CIPHER_free(type);
Add negative test for iv length change Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/22590)
2023-11-02 01:14:09 +08:00
return testresult;
Test EVP_CipherInit sequences and resets Various EVP_CipherInit sequences including partial inits and initializations with different "enc" flags caused problems on s390x. Similarly, cipher reinitialization and especially GCM reinitialization with different tag length led to wrong results. Add some unit tests to cover these rather exotic use cases. Signed-off-by: Juergen Christ <jchrist@linux.ibm.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/15521)
2021-06-03 01:33:50 +08:00
}
Test EVP Cipher updating the context's IV Ensure that an EVP_CipherUpdate operation updates the context's IV for AES CBC, CFB, OFB, and CTR. An application can get the updated IV via EVP_CIPHER_CTX_iv(). The s390x implementation of the CFB and OFB ciphers in e_aes.c did not update the IV in the context, but only within its s390x specific context data. Signed-off-by: Ingo Franzki <ifranzki@linux.ibm.com> Reviewed-by: Patrick Steuer <patrick.steuer@de.ibm.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16291)
2021-08-11 19:04:52 +08:00
typedef struct {
const char *cipher;
int enc;
} EVP_UPDATED_IV_TEST_st;
static const EVP_UPDATED_IV_TEST_st evp_updated_iv_tests[] = {
{
"aes-128-cfb", 1
},
{
"aes-128-cfb", 0
},
{
"aes-128-cfb1", 1
},
{
"aes-128-cfb1", 0
},
{
"aes-128-cfb8", 1
},
{
"aes-128-cfb8", 0
},
{
"aes-128-ofb", 1
},
{
"aes-128-ofb", 0
},
{
"aes-128-ctr", 1
},
{
"aes-128-ctr", 0
},
{
"aes-128-cbc", 1
},
{
"aes-128-cbc", 0
}
};
/*
* Test that the IV in the context is updated during a crypto operation for CFB
* and OFB.
*/
static int test_evp_updated_iv(int idx)
{
const EVP_UPDATED_IV_TEST_st *t = &evp_updated_iv_tests[idx];
int outlen1, outlen2;
int testresult = 0;
unsigned char outbuf[1024];
EVP_CIPHER_CTX *ctx = NULL;
EVP_CIPHER *type = NULL;
unsigned char updated_iv[EVP_MAX_IV_LENGTH];
int iv_len;
char *errmsg = NULL;
if (!TEST_ptr(ctx = EVP_CIPHER_CTX_new())) {
errmsg = "CTX_ALLOC";
goto err;
}
if ((type = EVP_CIPHER_fetch(testctx, t->cipher, testpropq)) == NULL) {
TEST_info("cipher %s not supported, skipping", t->cipher);
goto ok;
}
if (!TEST_true(EVP_CipherInit_ex(ctx, type, NULL, kCFBDefaultKey, iCFBIV, t->enc))) {
errmsg = "CIPHER_INIT";
goto err;
}
if (!TEST_true(EVP_CIPHER_CTX_set_padding(ctx, 0))) {
errmsg = "PADDING";
goto err;
}
if (!TEST_true(EVP_CipherUpdate(ctx, outbuf, &outlen1, cfbPlaintext, sizeof(cfbPlaintext)))) {
errmsg = "CIPHER_UPDATE";
goto err;
}
if (!TEST_true(EVP_CIPHER_CTX_get_updated_iv(ctx, updated_iv, sizeof(updated_iv)))) {
errmsg = "CIPHER_CTX_GET_UPDATED_IV";
goto err;
}
Add appropriate NULL checks in EVP_CIPHER api The EVP_CIPHER api currently assumes that calls made into several APIs have already initalized the cipher in a given context via a call to EVP_CipherInit[_ex[2]]. If that hasnt been done, instead of an error, the result is typically a SIGSEGV. Correct that by adding missing NULL checks in the apropriate apis prior to using ctx->cipher Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/22995)
2023-12-10 02:40:01 +08:00
iv_len = EVP_CIPHER_CTX_get_iv_length(ctx);
if (!TEST_int_ge(iv_len,0)) {
Test EVP Cipher updating the context's IV Ensure that an EVP_CipherUpdate operation updates the context's IV for AES CBC, CFB, OFB, and CTR. An application can get the updated IV via EVP_CIPHER_CTX_iv(). The s390x implementation of the CFB and OFB ciphers in e_aes.c did not update the IV in the context, but only within its s390x specific context data. Signed-off-by: Ingo Franzki <ifranzki@linux.ibm.com> Reviewed-by: Patrick Steuer <patrick.steuer@de.ibm.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16291)
2021-08-11 19:04:52 +08:00
errmsg = "CIPHER_CTX_GET_IV_LEN";
goto err;
}
if (!TEST_mem_ne(iCFBIV, sizeof(iCFBIV), updated_iv, iv_len)) {
errmsg = "IV_NOT_UPDATED";
goto err;
}
if (!TEST_true(EVP_CipherFinal_ex(ctx, outbuf + outlen1, &outlen2))) {
errmsg = "CIPHER_FINAL";
goto err;
}
ok:
testresult = 1;
err:
if (errmsg != NULL)
TEST_info("test_evp_updated_iv %d: %s", idx, errmsg);
EVP_CIPHER_CTX_free(ctx);
EVP_CIPHER_free(type);
return testresult;
}
Test EVP_CipherInit sequences and resets Various EVP_CipherInit sequences including partial inits and initializations with different "enc" flags caused problems on s390x. Similarly, cipher reinitialization and especially GCM reinitialization with different tag length led to wrong results. Add some unit tests to cover these rather exotic use cases. Signed-off-by: Juergen Christ <jchrist@linux.ibm.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/15521)
2021-06-03 01:33:50 +08:00
typedef struct {
const unsigned char *iv1;
const unsigned char *iv2;
const unsigned char *expected1;
const unsigned char *expected2;
const unsigned char *tag1;
const unsigned char *tag2;
size_t ivlen1;
size_t ivlen2;
size_t expectedlen1;
size_t expectedlen2;
} TEST_GCM_IV_REINIT_st;
static const TEST_GCM_IV_REINIT_st gcm_reinit_tests[] = {
{
iGCMResetIV1, iGCMResetIV2, gcmResetCiphertext1, gcmResetCiphertext2,
gcmResetTag1, gcmResetTag2, sizeof(iGCMResetIV1), sizeof(iGCMResetIV2),
sizeof(gcmResetCiphertext1), sizeof(gcmResetCiphertext2)
},
{
iGCMResetIV2, iGCMResetIV1, gcmResetCiphertext2, gcmResetCiphertext1,
gcmResetTag2, gcmResetTag1, sizeof(iGCMResetIV2), sizeof(iGCMResetIV1),
sizeof(gcmResetCiphertext2), sizeof(gcmResetCiphertext1)
}
};
static int test_gcm_reinit(int idx)
{
int outlen1, outlen2, outlen3;
int testresult = 0;
unsigned char outbuf[1024];
unsigned char tag[16];
const TEST_GCM_IV_REINIT_st *t = &gcm_reinit_tests[idx];
EVP_CIPHER_CTX *ctx = NULL;
EVP_CIPHER *type = NULL;
size_t taglen = sizeof(tag);
char *errmsg = NULL;
if (!TEST_ptr(ctx = EVP_CIPHER_CTX_new())) {
errmsg = "CTX_ALLOC";
goto err;
}
if (!TEST_ptr(type = EVP_CIPHER_fetch(testctx, "aes-256-gcm", testpropq))) {
errmsg = "CIPHER_FETCH";
goto err;
}
if (!TEST_true(EVP_CipherInit_ex(ctx, type, NULL, NULL, NULL, 1))) {
errmsg = "ENC_INIT";
goto err;
}
Fix check of EVP_CIPHER_CTX_ctrl Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/18368)
2022-05-21 16:17:23 +08:00
if (!TEST_int_gt(EVP_CIPHER_CTX_ctrl(ctx, EVP_CTRL_GCM_SET_IVLEN, t->ivlen1, NULL), 0)) {
Test EVP_CipherInit sequences and resets Various EVP_CipherInit sequences including partial inits and initializations with different "enc" flags caused problems on s390x. Similarly, cipher reinitialization and especially GCM reinitialization with different tag length led to wrong results. Add some unit tests to cover these rather exotic use cases. Signed-off-by: Juergen Christ <jchrist@linux.ibm.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/15521)
2021-06-03 01:33:50 +08:00
errmsg = "SET_IVLEN1";
goto err;
}
if (!TEST_true(EVP_CipherInit_ex(ctx, NULL, NULL, kGCMResetKey, t->iv1, 1))) {
errmsg = "SET_IV1";
goto err;
}
if (!TEST_true(EVP_CipherUpdate(ctx, NULL, &outlen3, gcmAAD, sizeof(gcmAAD)))) {
errmsg = "AAD1";
goto err;
}
EVP_CIPHER_CTX_set_padding(ctx, 0);
if (!TEST_true(EVP_CipherUpdate(ctx, outbuf, &outlen1, gcmResetPlaintext,
sizeof(gcmResetPlaintext)))) {
errmsg = "CIPHER_UPDATE1";
goto err;
}
if (!TEST_true(EVP_CipherFinal_ex(ctx, outbuf + outlen1, &outlen2))) {
errmsg = "CIPHER_FINAL1";
goto err;
}
if (!TEST_mem_eq(t->expected1, t->expectedlen1, outbuf, outlen1 + outlen2)) {
errmsg = "WRONG_RESULT1";
goto err;
}
Fix check of EVP_CIPHER_CTX_ctrl Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/18368)
2022-05-21 16:17:23 +08:00
if (!TEST_int_gt(EVP_CIPHER_CTX_ctrl(ctx, EVP_CTRL_AEAD_GET_TAG, taglen, tag), 0)) {
Test EVP_CipherInit sequences and resets Various EVP_CipherInit sequences including partial inits and initializations with different "enc" flags caused problems on s390x. Similarly, cipher reinitialization and especially GCM reinitialization with different tag length led to wrong results. Add some unit tests to cover these rather exotic use cases. Signed-off-by: Juergen Christ <jchrist@linux.ibm.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/15521)
2021-06-03 01:33:50 +08:00
errmsg = "GET_TAG1";
goto err;
}
if (!TEST_mem_eq(t->tag1, taglen, tag, taglen)) {
errmsg = "TAG_ERROR1";
goto err;
}
/* Now reinit */
Fix check of EVP_CIPHER_CTX_ctrl Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/18368)
2022-05-21 16:17:23 +08:00
if (!TEST_int_gt(EVP_CIPHER_CTX_ctrl(ctx, EVP_CTRL_GCM_SET_IVLEN, t->ivlen2, NULL), 0)) {
Test EVP_CipherInit sequences and resets Various EVP_CipherInit sequences including partial inits and initializations with different "enc" flags caused problems on s390x. Similarly, cipher reinitialization and especially GCM reinitialization with different tag length led to wrong results. Add some unit tests to cover these rather exotic use cases. Signed-off-by: Juergen Christ <jchrist@linux.ibm.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/15521)
2021-06-03 01:33:50 +08:00
errmsg = "SET_IVLEN2";
goto err;
}
if (!TEST_true(EVP_CipherInit_ex(ctx, NULL, NULL, NULL, t->iv2, -1))) {
errmsg = "SET_IV2";
goto err;
}
if (!TEST_true(EVP_CipherUpdate(ctx, NULL, &outlen3, gcmAAD, sizeof(gcmAAD)))) {
errmsg = "AAD2";
goto err;
}
if (!TEST_true(EVP_CipherUpdate(ctx, outbuf, &outlen1, gcmResetPlaintext,
sizeof(gcmResetPlaintext)))) {
errmsg = "CIPHER_UPDATE2";
goto err;
}
if (!TEST_true(EVP_CipherFinal_ex(ctx, outbuf + outlen1, &outlen2))) {
errmsg = "CIPHER_FINAL2";
goto err;
}
if (!TEST_mem_eq(t->expected2, t->expectedlen2, outbuf, outlen1 + outlen2)) {
errmsg = "WRONG_RESULT2";
goto err;
}
Fix check of EVP_CIPHER_CTX_ctrl Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/18368)
2022-05-21 16:17:23 +08:00
if (!TEST_int_gt(EVP_CIPHER_CTX_ctrl(ctx, EVP_CTRL_AEAD_GET_TAG, taglen, tag), 0)) {
Test EVP_CipherInit sequences and resets Various EVP_CipherInit sequences including partial inits and initializations with different "enc" flags caused problems on s390x. Similarly, cipher reinitialization and especially GCM reinitialization with different tag length led to wrong results. Add some unit tests to cover these rather exotic use cases. Signed-off-by: Juergen Christ <jchrist@linux.ibm.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/15521)
2021-06-03 01:33:50 +08:00
errmsg = "GET_TAG2";
goto err;
}
if (!TEST_mem_eq(t->tag2, taglen, tag, taglen)) {
errmsg = "TAG_ERROR2";
goto err;
}
testresult = 1;
err:
if (errmsg != NULL)
TEST_info("evp_init_test %d: %s", idx, errmsg);
EVP_CIPHER_CTX_free(ctx);
EVP_CIPHER_free(type);
return testresult;
}
Add negative test for iv length change Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/22590)
2023-11-02 01:14:09 +08:00
static const char *ivlen_change_ciphers[] = {
"AES-256-GCM",
#ifndef OPENSSL_NO_OCB
"AES-256-OCB",
#endif
"AES-256-CCM"
};
/* Negative test for ivlen change after iv being set */
static int test_ivlen_change(int idx)
{
int outlen;
int res = 0;
unsigned char outbuf[1024];
static const unsigned char iv[] = {
0x57, 0x71, 0x7d, 0xad, 0xdb, 0x9b, 0x98, 0x82,
0x5a, 0x55, 0x91, 0x81, 0x42, 0xa8, 0x89, 0x34
};
EVP_CIPHER_CTX *ctx = NULL;
EVP_CIPHER *ciph = NULL;
OSSL_PARAM params[] = { OSSL_PARAM_END, OSSL_PARAM_END };
size_t ivlen = 13; /* non-default IV length */
if (!TEST_ptr(ctx = EVP_CIPHER_CTX_new()))
goto err;
if (!TEST_ptr(ciph = EVP_CIPHER_fetch(testctx, ivlen_change_ciphers[idx],
testpropq)))
goto err;
if (!TEST_true(EVP_CipherInit_ex(ctx, ciph, NULL, kGCMDefaultKey, iv, 1)))
goto err;
if (!TEST_true(EVP_CipherUpdate(ctx, outbuf, &outlen, gcmDefaultPlaintext,
sizeof(gcmDefaultPlaintext))))
goto err;
params[0] = OSSL_PARAM_construct_size_t(OSSL_CIPHER_PARAM_AEAD_IVLEN,
&ivlen);
if (!TEST_true(EVP_CIPHER_CTX_set_params(ctx, params)))
goto err;
ERR_set_mark();
if (!TEST_false(EVP_CipherUpdate(ctx, outbuf, &outlen, gcmDefaultPlaintext,
sizeof(gcmDefaultPlaintext)))) {
ERR_clear_last_mark();
goto err;
}
ERR_pop_to_mark();
res = 1;
err:
EVP_CIPHER_CTX_free(ctx);
EVP_CIPHER_free(ciph);
return res;
}
Add negative test for key length change Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/22590)
2023-11-02 01:39:32 +08:00
static const char *keylen_change_ciphers[] = {
#ifndef OPENSSL_NO_BF
"BF-ECB",
#endif
#ifndef OPENSSL_NO_CAST
"CAST5-ECB",
#endif
#ifndef OPENSSL_NO_RC2
"RC2-ECB",
#endif
#ifndef OPENSSL_NO_RC4
"RC4",
#endif
#ifndef OPENSSL_NO_RC5
"RC5-ECB",
#endif
NULL
};
/* Negative test for keylen change after key was set */
static int test_keylen_change(int idx)
{
int outlen;
int res = 0;
unsigned char outbuf[1024];
static const unsigned char key[] = {
0x57, 0x71, 0x7d, 0xad, 0xdb, 0x9b, 0x98, 0x82,
0x5a, 0x55, 0x91, 0x81, 0x42, 0xa8, 0x89, 0x34
};
EVP_CIPHER_CTX *ctx = NULL;
EVP_CIPHER *ciph = NULL;
OSSL_PARAM params[] = { OSSL_PARAM_END, OSSL_PARAM_END };
size_t keylen = 12; /* non-default key length */
if (lgcyprov == NULL)
return TEST_skip("Test requires legacy provider to be loaded");
if (!TEST_ptr(ctx = EVP_CIPHER_CTX_new()))
goto err;
if (!TEST_ptr(ciph = EVP_CIPHER_fetch(testctx, keylen_change_ciphers[idx],
testpropq)))
goto err;
if (!TEST_true(EVP_CipherInit_ex(ctx, ciph, NULL, key, NULL, 1)))
goto err;
if (!TEST_true(EVP_CipherUpdate(ctx, outbuf, &outlen, gcmDefaultPlaintext,
sizeof(gcmDefaultPlaintext))))
goto err;
params[0] = OSSL_PARAM_construct_size_t(OSSL_CIPHER_PARAM_KEYLEN,
&keylen);
if (!TEST_true(EVP_CIPHER_CTX_set_params(ctx, params)))
goto err;
ERR_set_mark();
if (!TEST_false(EVP_CipherUpdate(ctx, outbuf, &outlen, gcmDefaultPlaintext,
sizeof(gcmDefaultPlaintext)))) {
ERR_clear_last_mark();
goto err;
}
ERR_pop_to_mark();
res = 1;
err:
EVP_CIPHER_CTX_free(ctx);
EVP_CIPHER_free(ciph);
return res;
}
Add a test for custom EVP_PKEY_METHODs Adds a test for using custom EVP_PKEY_METHODs without an ENGINE. As part of this we also test having a custom EVP_PKEY_METHOD that wraps a built-in EVP_PKEY_METHOD. We do this for both legacy and provided keys. Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16118)
2021-07-20 23:18:58 +08:00
#ifndef OPENSSL_NO_DEPRECATED_3_0
static EVP_PKEY_METHOD *custom_pmeth = NULL;
static const EVP_PKEY_METHOD *orig_pmeth = NULL;
# define EVP_PKEY_CTRL_MY_COMMAND 9999
static int custom_pmeth_init(EVP_PKEY_CTX *ctx)
{
int (*pinit)(EVP_PKEY_CTX *ctx);
EVP_PKEY_meth_get_init(orig_pmeth, &pinit);
return pinit(ctx);
}
static void custom_pmeth_cleanup(EVP_PKEY_CTX *ctx)
{
void (*pcleanup)(EVP_PKEY_CTX *ctx);
EVP_PKEY_meth_get_cleanup(orig_pmeth, &pcleanup);
pcleanup(ctx);
}
static int custom_pmeth_sign(EVP_PKEY_CTX *ctx, unsigned char *out,
size_t *outlen, const unsigned char *in,
size_t inlen)
{
int (*psign)(EVP_PKEY_CTX *ctx, unsigned char *sig, size_t *siglen,
const unsigned char *tbs, size_t tbslen);
EVP_PKEY_meth_get_sign(orig_pmeth, NULL, &psign);
return psign(ctx, out, outlen, in, inlen);
}
static int custom_pmeth_digestsign(EVP_MD_CTX *ctx, unsigned char *sig,
size_t *siglen, const unsigned char *tbs,
size_t tbslen)
{
int (*pdigestsign)(EVP_MD_CTX *ctx, unsigned char *sig, size_t *siglen,
const unsigned char *tbs, size_t tbslen);
Mark the EVP_PKEY_METHOD arg as const on some EVP_PKEY_meth_get_*() funcs Most EVP_PKEY_meth_get_*() functions mark the EVP_PKEY_METHOD argument as const. But 3 did not. We fix those to be consistent. Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16128)
2021-07-20 23:18:04 +08:00
EVP_PKEY_meth_get_digestsign(orig_pmeth, &pdigestsign);
Add a test for custom EVP_PKEY_METHODs Adds a test for using custom EVP_PKEY_METHODs without an ENGINE. As part of this we also test having a custom EVP_PKEY_METHOD that wraps a built-in EVP_PKEY_METHOD. We do this for both legacy and provided keys. Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16118)
2021-07-20 23:18:58 +08:00
return pdigestsign(ctx, sig, siglen, tbs, tbslen);
}
static int custom_pmeth_derive(EVP_PKEY_CTX *ctx, unsigned char *key,
size_t *keylen)
{
int (*pderive)(EVP_PKEY_CTX *ctx, unsigned char *key, size_t *keylen);
EVP_PKEY_meth_get_derive(orig_pmeth, NULL, &pderive);
return pderive(ctx, key, keylen);
}
static int custom_pmeth_copy(EVP_PKEY_CTX *dst, const EVP_PKEY_CTX *src)
{
int (*pcopy)(EVP_PKEY_CTX *dst, const EVP_PKEY_CTX *src);
EVP_PKEY_meth_get_copy(orig_pmeth, &pcopy);
return pcopy(dst, src);
}
static int ctrl_called;
static int custom_pmeth_ctrl(EVP_PKEY_CTX *ctx, int type, int p1, void *p2)
{
int (*pctrl)(EVP_PKEY_CTX *ctx, int type, int p1, void *p2);
EVP_PKEY_meth_get_ctrl(orig_pmeth, &pctrl, NULL);
if (type == EVP_PKEY_CTRL_MY_COMMAND) {
ctrl_called = 1;
return 1;
}
return pctrl(ctx, type, p1, p2);
}
static int test_custom_pmeth(int idx)
{
EVP_PKEY_CTX *pctx = NULL;
EVP_MD_CTX *ctx = NULL;
EVP_PKEY *pkey = NULL;
int id, orig_id, orig_flags;
int testresult = 0;
size_t reslen;
unsigned char *res = NULL;
unsigned char msg[] = { 'H', 'e', 'l', 'l', 'o' };
const EVP_MD *md = EVP_sha256();
int doderive = 0;
ctrl_called = 0;
/* We call deprecated APIs so this test doesn't support a custom libctx */
if (testctx != NULL)
return 1;
fix some code with obvious wrong coding style Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16918)
2021-10-26 15:16:18 +08:00
switch (idx) {
Add a test for custom EVP_PKEY_METHODs Adds a test for using custom EVP_PKEY_METHODs without an ENGINE. As part of this we also test having a custom EVP_PKEY_METHOD that wraps a built-in EVP_PKEY_METHOD. We do this for both legacy and provided keys. Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16118)
2021-07-20 23:18:58 +08:00
case 0:
case 6:
id = EVP_PKEY_RSA;
pkey = load_example_rsa_key();
break;
case 1:
case 7:
# ifndef OPENSSL_NO_DSA
id = EVP_PKEY_DSA;
pkey = load_example_dsa_key();
break;
# else
return 1;
# endif
case 2:
case 8:
# ifndef OPENSSL_NO_EC
id = EVP_PKEY_EC;
pkey = load_example_ec_key();
break;
# else
return 1;
# endif
case 3:
case 9:
configure: introduce no-ecx to remove ECX related feature This can effectively reduce the binary size for platforms that don't need ECX feature(~100KB). Signed-off-by: Yi Li <yi1.li@intel.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/20781)
2023-04-17 16:20:31 +08:00
# ifndef OPENSSL_NO_ECX
Add a test for custom EVP_PKEY_METHODs Adds a test for using custom EVP_PKEY_METHODs without an ENGINE. As part of this we also test having a custom EVP_PKEY_METHOD that wraps a built-in EVP_PKEY_METHOD. We do this for both legacy and provided keys. Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16118)
2021-07-20 23:18:58 +08:00
id = EVP_PKEY_ED25519;
md = NULL;
pkey = load_example_ed25519_key();
break;
# else
return 1;
# endif
case 4:
case 10:
# ifndef OPENSSL_NO_DH
id = EVP_PKEY_DH;
doderive = 1;
pkey = load_example_dh_key();
break;
# else
return 1;
# endif
case 5:
case 11:
configure: introduce no-ecx to remove ECX related feature This can effectively reduce the binary size for platforms that don't need ECX feature(~100KB). Signed-off-by: Yi Li <yi1.li@intel.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/20781)
2023-04-17 16:20:31 +08:00
# ifndef OPENSSL_NO_ECX
Add a test for custom EVP_PKEY_METHODs Adds a test for using custom EVP_PKEY_METHODs without an ENGINE. As part of this we also test having a custom EVP_PKEY_METHOD that wraps a built-in EVP_PKEY_METHOD. We do this for both legacy and provided keys. Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16118)
2021-07-20 23:18:58 +08:00
id = EVP_PKEY_X25519;
doderive = 1;
pkey = load_example_x25519_key();
break;
# else
return 1;
# endif
default:
TEST_error("Should not happen");
goto err;
}
if (!TEST_ptr(pkey))
goto err;
if (idx < 6) {
if (!TEST_true(evp_pkey_is_provided(pkey)))
goto err;
} else {
EVP_PKEY *tmp = pkey;
/* Convert to a legacy key */
pkey = EVP_PKEY_new();
if (!TEST_ptr(pkey)) {
pkey = tmp;
goto err;
}
if (!TEST_true(evp_pkey_copy_downgraded(&pkey, tmp))) {
EVP_PKEY_free(tmp);
goto err;
}
EVP_PKEY_free(tmp);
if (!TEST_true(evp_pkey_is_legacy(pkey)))
goto err;
}
if (!TEST_ptr(orig_pmeth = EVP_PKEY_meth_find(id))
|| !TEST_ptr(pkey))
goto err;
EVP_PKEY_meth_get0_info(&orig_id, &orig_flags, orig_pmeth);
if (!TEST_int_eq(orig_id, id)
|| !TEST_ptr(custom_pmeth = EVP_PKEY_meth_new(id, orig_flags)))
goto err;
if (id == EVP_PKEY_ED25519) {
EVP_PKEY_meth_set_digestsign(custom_pmeth, custom_pmeth_digestsign);
} if (id == EVP_PKEY_DH || id == EVP_PKEY_X25519) {
EVP_PKEY_meth_set_derive(custom_pmeth, NULL, custom_pmeth_derive);
} else {
EVP_PKEY_meth_set_sign(custom_pmeth, NULL, custom_pmeth_sign);
}
if (id != EVP_PKEY_ED25519 && id != EVP_PKEY_X25519) {
EVP_PKEY_meth_set_init(custom_pmeth, custom_pmeth_init);
EVP_PKEY_meth_set_cleanup(custom_pmeth, custom_pmeth_cleanup);
EVP_PKEY_meth_set_copy(custom_pmeth, custom_pmeth_copy);
}
EVP_PKEY_meth_set_ctrl(custom_pmeth, custom_pmeth_ctrl, NULL);
if (!TEST_true(EVP_PKEY_meth_add0(custom_pmeth)))
goto err;
if (doderive) {
pctx = EVP_PKEY_CTX_new(pkey, NULL);
if (!TEST_ptr(pctx)
|| !TEST_int_eq(EVP_PKEY_derive_init(pctx), 1)
|| !TEST_int_ge(EVP_PKEY_CTX_ctrl(pctx, -1, -1,
EVP_PKEY_CTRL_MY_COMMAND, 0, NULL),
1)
|| !TEST_int_eq(ctrl_called, 1)
|| !TEST_int_ge(EVP_PKEY_derive_set_peer(pctx, pkey), 1)
|| !TEST_int_ge(EVP_PKEY_derive(pctx, NULL, &reslen), 1)
|| !TEST_ptr(res = OPENSSL_malloc(reslen))
|| !TEST_int_ge(EVP_PKEY_derive(pctx, res, &reslen), 1))
goto err;
} else {
ctx = EVP_MD_CTX_new();
reslen = EVP_PKEY_size(pkey);
res = OPENSSL_malloc(reslen);
if (!TEST_ptr(ctx)
|| !TEST_ptr(res)
|| !TEST_true(EVP_DigestSignInit(ctx, &pctx, md, NULL, pkey))
|| !TEST_int_ge(EVP_PKEY_CTX_ctrl(pctx, -1, -1,
EVP_PKEY_CTRL_MY_COMMAND, 0, NULL),
1)
|| !TEST_int_eq(ctrl_called, 1))
goto err;
if (id == EVP_PKEY_ED25519) {
if (!TEST_true(EVP_DigestSign(ctx, res, &reslen, msg, sizeof(msg))))
goto err;
} else {
if (!TEST_true(EVP_DigestUpdate(ctx, msg, sizeof(msg)))
|| !TEST_true(EVP_DigestSignFinal(ctx, res, &reslen)))
goto err;
}
}
testresult = 1;
err:
OPENSSL_free(res);
EVP_MD_CTX_free(ctx);
if (doderive)
EVP_PKEY_CTX_free(pctx);
EVP_PKEY_free(pkey);
EVP_PKEY_meth_remove(custom_pmeth);
EVP_PKEY_meth_free(custom_pmeth);
custom_pmeth = NULL;
return testresult;
}
Add a test case for EVP_MD_meth_dup() and EVP_CIPHER_meth_dup() Check that EVP_MD_meth_free() and EVP_CIPHER_meth_free() does actually free the data. Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16159)
2021-07-27 17:32:49 +08:00
static int test_evp_md_cipher_meth(void)
{
EVP_MD *md = EVP_MD_meth_dup(EVP_sha256());
EVP_CIPHER *ciph = EVP_CIPHER_meth_dup(EVP_aes_128_cbc());
int testresult = 0;
if (!TEST_ptr(md) || !TEST_ptr(ciph))
goto err;
testresult = 1;
err:
EVP_MD_meth_free(md);
EVP_CIPHER_meth_free(ciph);
return testresult;
}
Add tests for ENGINE problems Add some tests which would have caught the issues fixed in the previous 3 commits related to engine handling. Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16846)
2021-10-15 23:30:45 +08:00
Add a test for a custom digest created via EVP_MD_meth_new() We check that the init and cleanup functions for the custom method are called as expected. Based on an original reproducer by Dmitry Belyavsky from issue #17149. Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com> (Merged from https://github.com/openssl/openssl/pull/17255)
2021-12-30 00:39:11 +08:00
typedef struct {
int data;
} custom_dgst_ctx;
static int custom_md_init_called = 0;
static int custom_md_cleanup_called = 0;
static int custom_md_init(EVP_MD_CTX *ctx)
{
custom_dgst_ctx *p = EVP_MD_CTX_md_data(ctx);
if (p == NULL)
return 0;
custom_md_init_called++;
return 1;
}
static int custom_md_cleanup(EVP_MD_CTX *ctx)
{
custom_dgst_ctx *p = EVP_MD_CTX_md_data(ctx);
if (p == NULL)
/* Nothing to do */
return 1;
custom_md_cleanup_called++;
return 1;
}
static int test_custom_md_meth(void)
{
Fix error handling in OBJ_add_object This fixes the possible memory leak in OBJ_add_object when a pre-existing object is replaced by a new one, with identical NID, OID, and/or short/long name. We do not try to delete any orphans, but only mark them as type == -1, because the previously returned pointers from OBJ_nid2obj/OBJ_nid2sn/OBJ_nid2ln may be cached by applications and can thus not be cleaned up before the application terminates. Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/22534)
2023-10-27 18:05:05 +08:00
ASN1_OBJECT *o = NULL;
Add a test for a custom digest created via EVP_MD_meth_new() We check that the init and cleanup functions for the custom method are called as expected. Based on an original reproducer by Dmitry Belyavsky from issue #17149. Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com> (Merged from https://github.com/openssl/openssl/pull/17255)
2021-12-30 00:39:11 +08:00
EVP_MD_CTX *mdctx = NULL;
EVP_MD *tmp = NULL;
char mess[] = "Test Message\n";
unsigned char md_value[EVP_MAX_MD_SIZE];
unsigned int md_len;
int testresult = 0;
int nid;
/*
* We are testing deprecated functions. We don't support a non-default
* library context in this test.
*/
if (testctx != NULL)
Test usage of a custom EVP_CIPHER Test that a custom EVP_CIPHER gets used in EVP_CipherInit_ex() calls. Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/19300)
2022-09-29 19:10:30 +08:00
return TEST_skip("Non-default libctx");
Add a test for a custom digest created via EVP_MD_meth_new() We check that the init and cleanup functions for the custom method are called as expected. Based on an original reproducer by Dmitry Belyavsky from issue #17149. Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com> (Merged from https://github.com/openssl/openssl/pull/17255)
2021-12-30 00:39:11 +08:00
custom_md_init_called = custom_md_cleanup_called = 0;
nid = OBJ_create("1.3.6.1.4.1.16604.998866.1", "custom-md", "custom-md");
if (!TEST_int_ne(nid, NID_undef))
goto err;
tmp = EVP_MD_meth_new(nid, NID_undef);
if (!TEST_ptr(tmp))
goto err;
if (!TEST_true(EVP_MD_meth_set_init(tmp, custom_md_init))
|| !TEST_true(EVP_MD_meth_set_cleanup(tmp, custom_md_cleanup))
|| !TEST_true(EVP_MD_meth_set_app_datasize(tmp,
sizeof(custom_dgst_ctx))))
goto err;
mdctx = EVP_MD_CTX_new();
if (!TEST_ptr(mdctx)
/*
* Initing our custom md and then initing another md should
* result in the init and cleanup functions of the custom md
Test usage of a custom EVP_CIPHER Test that a custom EVP_CIPHER gets used in EVP_CipherInit_ex() calls. Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/19300)
2022-09-29 19:10:30 +08:00
* being called.
Add a test for a custom digest created via EVP_MD_meth_new() We check that the init and cleanup functions for the custom method are called as expected. Based on an original reproducer by Dmitry Belyavsky from issue #17149. Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com> (Merged from https://github.com/openssl/openssl/pull/17255)
2021-12-30 00:39:11 +08:00
*/
|| !TEST_true(EVP_DigestInit_ex(mdctx, tmp, NULL))
|| !TEST_true(EVP_DigestInit_ex(mdctx, EVP_sha256(), NULL))
|| !TEST_true(EVP_DigestUpdate(mdctx, mess, strlen(mess)))
|| !TEST_true(EVP_DigestFinal_ex(mdctx, md_value, &md_len))
|| !TEST_int_eq(custom_md_init_called, 1)
|| !TEST_int_eq(custom_md_cleanup_called, 1))
goto err;
Fix error handling in OBJ_add_object This fixes the possible memory leak in OBJ_add_object when a pre-existing object is replaced by a new one, with identical NID, OID, and/or short/long name. We do not try to delete any orphans, but only mark them as type == -1, because the previously returned pointers from OBJ_nid2obj/OBJ_nid2sn/OBJ_nid2ln may be cached by applications and can thus not be cleaned up before the application terminates. Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/22534)
2023-10-27 18:05:05 +08:00
if (!TEST_int_eq(OBJ_create("1.3.6.1.4.1.16604.998866.1",
"custom-md", "custom-md"), NID_undef)
|| !TEST_int_eq(ERR_GET_LIB(ERR_peek_error()), ERR_LIB_OBJ)
|| !TEST_int_eq(ERR_GET_REASON(ERR_get_error()), OBJ_R_OID_EXISTS))
goto err;
o = ASN1_OBJECT_create(nid, (unsigned char *)
"\53\6\1\4\1\201\201\134\274\373\122\1", 12,
"custom-md", "custom-md");
if (!TEST_int_eq(OBJ_add_object(o), nid))
goto err;
Add a test for a custom digest created via EVP_MD_meth_new() We check that the init and cleanup functions for the custom method are called as expected. Based on an original reproducer by Dmitry Belyavsky from issue #17149. Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com> (Merged from https://github.com/openssl/openssl/pull/17255)
2021-12-30 00:39:11 +08:00
testresult = 1;
err:
Fix error handling in OBJ_add_object This fixes the possible memory leak in OBJ_add_object when a pre-existing object is replaced by a new one, with identical NID, OID, and/or short/long name. We do not try to delete any orphans, but only mark them as type == -1, because the previously returned pointers from OBJ_nid2obj/OBJ_nid2sn/OBJ_nid2ln may be cached by applications and can thus not be cleaned up before the application terminates. Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/22534)
2023-10-27 18:05:05 +08:00
ASN1_OBJECT_free(o);
Add a test for a custom digest created via EVP_MD_meth_new() We check that the init and cleanup functions for the custom method are called as expected. Based on an original reproducer by Dmitry Belyavsky from issue #17149. Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com> (Merged from https://github.com/openssl/openssl/pull/17255)
2021-12-30 00:39:11 +08:00
EVP_MD_CTX_free(mdctx);
EVP_MD_meth_free(tmp);
return testresult;
}
Test usage of a custom EVP_CIPHER Test that a custom EVP_CIPHER gets used in EVP_CipherInit_ex() calls. Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/19300)
2022-09-29 19:10:30 +08:00
typedef struct {
int data;
} custom_ciph_ctx;
static int custom_ciph_init_called = 0;
static int custom_ciph_cleanup_called = 0;
static int custom_ciph_init(EVP_CIPHER_CTX *ctx, const unsigned char *key,
const unsigned char *iv, int enc)
{
custom_ciph_ctx *p = EVP_CIPHER_CTX_get_cipher_data(ctx);
if (p == NULL)
return 0;
custom_ciph_init_called++;
return 1;
}
static int custom_ciph_cleanup(EVP_CIPHER_CTX *ctx)
{
custom_ciph_ctx *p = EVP_CIPHER_CTX_get_cipher_data(ctx);
if (p == NULL)
/* Nothing to do */
return 1;
custom_ciph_cleanup_called++;
return 1;
}
static int test_custom_ciph_meth(void)
{
EVP_CIPHER_CTX *ciphctx = NULL;
EVP_CIPHER *tmp = NULL;
int testresult = 0;
int nid;
/*
* We are testing deprecated functions. We don't support a non-default
* library context in this test.
*/
if (testctx != NULL)
return TEST_skip("Non-default libctx");
custom_ciph_init_called = custom_ciph_cleanup_called = 0;
nid = OBJ_create("1.3.6.1.4.1.16604.998866.2", "custom-ciph", "custom-ciph");
if (!TEST_int_ne(nid, NID_undef))
goto err;
tmp = EVP_CIPHER_meth_new(nid, 16, 16);
if (!TEST_ptr(tmp))
goto err;
if (!TEST_true(EVP_CIPHER_meth_set_init(tmp, custom_ciph_init))
|| !TEST_true(EVP_CIPHER_meth_set_flags(tmp, EVP_CIPH_ALWAYS_CALL_INIT))
|| !TEST_true(EVP_CIPHER_meth_set_cleanup(tmp, custom_ciph_cleanup))
|| !TEST_true(EVP_CIPHER_meth_set_impl_ctx_size(tmp,
sizeof(custom_ciph_ctx))))
goto err;
ciphctx = EVP_CIPHER_CTX_new();
if (!TEST_ptr(ciphctx)
/*
* Initing our custom cipher and then initing another cipher
* should result in the init and cleanup functions of the custom
* cipher being called.
*/
|| !TEST_true(EVP_CipherInit_ex(ciphctx, tmp, NULL, NULL, NULL, 1))
|| !TEST_true(EVP_CipherInit_ex(ciphctx, EVP_aes_128_cbc(), NULL,
NULL, NULL, 1))
|| !TEST_int_eq(custom_ciph_init_called, 1)
|| !TEST_int_eq(custom_ciph_cleanup_called, 1))
goto err;
testresult = 1;
err:
EVP_CIPHER_CTX_free(ciphctx);
EVP_CIPHER_meth_free(tmp);
return testresult;
}
Add tests for ENGINE problems Add some tests which would have caught the issues fixed in the previous 3 commits related to engine handling. Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16846)
2021-10-15 23:30:45 +08:00
# ifndef OPENSSL_NO_DYNAMIC_ENGINE
/* Test we can create a signature keys with an associated ENGINE */
static int test_signatures_with_engine(int tst)
{
ENGINE *e;
const char *engine_id = "dasync";
EVP_PKEY *pkey = NULL;
const unsigned char badcmackey[] = { 0x00, 0x01 };
const unsigned char cmackey[] = {
0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08, 0x09, 0x0a, 0x0b,
0x0c, 0x0d, 0x0e, 0x0f
};
const unsigned char ed25519key[] = {
0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08, 0x09, 0x0a, 0x0b,
0x0c, 0x0d, 0x0e, 0x0f, 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17,
0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f
};
const unsigned char msg[] = { 0x00, 0x01, 0x02, 0x03 };
int testresult = 0;
EVP_MD_CTX *ctx = NULL;
unsigned char *mac = NULL;
size_t maclen = 0;
int ret;
Fix no-cmac Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16894)
2021-10-22 23:09:44 +08:00
# ifdef OPENSSL_NO_CMAC
/* Skip CMAC tests in a no-cmac build */
if (tst <= 1)
return 1;
# endif
configure: introduce no-ecx to remove ECX related feature This can effectively reduce the binary size for platforms that don't need ECX feature(~100KB). Signed-off-by: Yi Li <yi1.li@intel.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/20781)
2023-04-17 16:20:31 +08:00
# ifdef OPENSSL_NO_ECX
/* Skip ECX tests in a no-ecx build */
if (tst == 2)
return 1;
# endif
Fix no-cmac Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16894)
2021-10-22 23:09:44 +08:00
Add tests for ENGINE problems Add some tests which would have caught the issues fixed in the previous 3 commits related to engine handling. Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16846)
2021-10-15 23:30:45 +08:00
if (!TEST_ptr(e = ENGINE_by_id(engine_id)))
return 0;
if (!TEST_true(ENGINE_init(e))) {
ENGINE_free(e);
return 0;
}
switch (tst) {
case 0:
pkey = EVP_PKEY_new_CMAC_key(e, cmackey, sizeof(cmackey),
EVP_aes_128_cbc());
break;
case 1:
pkey = EVP_PKEY_new_CMAC_key(e, badcmackey, sizeof(badcmackey),
EVP_aes_128_cbc());
break;
case 2:
pkey = EVP_PKEY_new_raw_private_key(EVP_PKEY_ED25519, e, ed25519key,
sizeof(ed25519key));
break;
default:
TEST_error("Invalid test case");
goto err;
}
if (!TEST_ptr(pkey))
goto err;
if (!TEST_ptr(ctx = EVP_MD_CTX_new()))
goto err;
ret = EVP_DigestSignInit(ctx, NULL, tst == 2 ? NULL : EVP_sha256(), NULL,
pkey);
if (tst == 0) {
if (!TEST_true(ret))
goto err;
if (!TEST_true(EVP_DigestSignUpdate(ctx, msg, sizeof(msg)))
|| !TEST_true(EVP_DigestSignFinal(ctx, NULL, &maclen)))
goto err;
if (!TEST_ptr(mac = OPENSSL_malloc(maclen)))
goto err;
if (!TEST_true(EVP_DigestSignFinal(ctx, mac, &maclen)))
goto err;
} else {
/* We used a bad key. We expect a failure here */
if (!TEST_false(ret))
goto err;
}
testresult = 1;
err:
EVP_MD_CTX_free(ctx);
OPENSSL_free(mac);
EVP_PKEY_free(pkey);
ENGINE_finish(e);
ENGINE_free(e);
return testresult;
}
static int test_cipher_with_engine(void)
{
ENGINE *e;
const char *engine_id = "dasync";
const unsigned char keyiv[] = {
0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08, 0x09, 0x0a, 0x0b,
0x0c, 0x0d, 0x0e, 0x0f
};
const unsigned char msg[] = { 0x00, 0x01, 0x02, 0x03 };
int testresult = 0;
EVP_CIPHER_CTX *ctx = NULL, *ctx2 = NULL;
unsigned char buf[AES_BLOCK_SIZE];
int len = 0;
if (!TEST_ptr(e = ENGINE_by_id(engine_id)))
return 0;
if (!TEST_true(ENGINE_init(e))) {
ENGINE_free(e);
return 0;
}
if (!TEST_ptr(ctx = EVP_CIPHER_CTX_new())
|| !TEST_ptr(ctx2 = EVP_CIPHER_CTX_new()))
goto err;
if (!TEST_true(EVP_EncryptInit_ex(ctx, EVP_aes_128_cbc(), e, keyiv, keyiv)))
goto err;
/* Copy the ctx, and complete the operation with the new ctx */
if (!TEST_true(EVP_CIPHER_CTX_copy(ctx2, ctx)))
goto err;
if (!TEST_true(EVP_EncryptUpdate(ctx2, buf, &len, msg, sizeof(msg)))
|| !TEST_true(EVP_EncryptFinal_ex(ctx2, buf + len, &len)))
goto err;
testresult = 1;
err:
EVP_CIPHER_CTX_free(ctx);
EVP_CIPHER_CTX_free(ctx2);
ENGINE_finish(e);
ENGINE_free(e);
return testresult;
}
# endif /* OPENSSL_NO_DYNAMIC_ENGINE */
Add a test case for EVP_MD_meth_dup() and EVP_CIPHER_meth_dup() Check that EVP_MD_meth_free() and EVP_CIPHER_meth_free() does actually free the data. Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16159)
2021-07-27 17:32:49 +08:00
#endif /* OPENSSL_NO_DEPRECATED_3_0 */
Test EVP_CipherInit sequences and resets Various EVP_CipherInit sequences including partial inits and initializations with different "enc" flags caused problems on s390x. Similarly, cipher reinitialization and especially GCM reinitialization with different tag length led to wrong results. Add some unit tests to cover these rather exotic use cases. Signed-off-by: Juergen Christ <jchrist@linux.ibm.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/15521)
2021-06-03 01:33:50 +08:00
configure: introduce no-ecx to remove ECX related feature This can effectively reduce the binary size for platforms that don't need ECX feature(~100KB). Signed-off-by: Yi Li <yi1.li@intel.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/20781)
2023-04-17 16:20:31 +08:00
#ifndef OPENSSL_NO_ECX
Add a test for creating ECX private keys that are too short We expect attempting to create such short keys to fail Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/17041)
2021-11-15 20:24:05 +08:00
static int ecxnids[] = {
NID_X25519,
NID_X448,
NID_ED25519,
NID_ED448
};
/* Test that creating ECX keys with a short private key fails as expected */
static int test_ecx_short_keys(int tst)
{
unsigned char ecxkeydata = 1;
EVP_PKEY *pkey;
Check for private key existence before calling eddsa sign functions Fixes #19524 Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/19525)
2022-10-28 11:47:50 +08:00
pkey = EVP_PKEY_new_raw_private_key_ex(testctx, OBJ_nid2sn(ecxnids[tst]),
NULL, &ecxkeydata, 1);
Add a test for creating ECX private keys that are too short We expect attempting to create such short keys to fail Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/17041)
2021-11-15 20:24:05 +08:00
if (!TEST_ptr_null(pkey)) {
EVP_PKEY_free(pkey);
return 0;
}
Check for private key existence before calling eddsa sign functions Fixes #19524 Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/19525)
2022-10-28 11:47:50 +08:00
Add a test for creating ECX private keys that are too short We expect attempting to create such short keys to fail Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/17041)
2021-11-15 20:24:05 +08:00
return 1;
}
configure: introduce no-ecx to remove ECX related feature This can effectively reduce the binary size for platforms that don't need ECX feature(~100KB). Signed-off-by: Yi Li <yi1.li@intel.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/20781)
2023-04-17 16:20:31 +08:00
#endif
Add a test for creating ECX private keys that are too short We expect attempting to create such short keys to fail Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/17041)
2021-11-15 20:24:05 +08:00
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
typedef enum OPTION_choice {
OPT_ERR = -1,
OPT_EOF = 0,
OPT_CONTEXT,
EVP_get_default_properties - tests Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/25434)
2024-09-11 22:48:44 +08:00
OPT_CONFIG_FILE,
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
OPT_TEST_ENUM
} OPTION_CHOICE;
const OPTIONS *test_get_options(void)
Import evp_test.c from BoringSSL. Unfortunately we already have a file called evp_test.c, so I have called this one evp_extra_test.c Reviewed-by: Emilia Käsper <emilia@openssl.org>
2015-02-09 17:45:35 +08:00
{
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
static const OPTIONS options[] = {
OPT_TEST_OPTIONS_DEFAULT_USAGE,
{ "context", OPT_CONTEXT, '-', "Explicitly use a non-default library context" },
EVP_get_default_properties - tests Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/25434)
2024-09-11 22:48:44 +08:00
{ "config", OPT_CONFIG_FILE, '<',
"The configuration file to use for the libctx" },
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
{ NULL }
};
return options;
}
Add some tests for the newly added raw private/public key functions We already had soem tests for the older raw private/public key functions so we expand those to call the new versions as well and pass in a libctx. Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/11635)
2020-04-28 00:17:05 +08:00
configure: introduce no-ecx to remove ECX related feature This can effectively reduce the binary size for platforms that don't need ECX feature(~100KB). Signed-off-by: Yi Li <yi1.li@intel.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/20781)
2023-04-17 16:20:31 +08:00
#ifndef OPENSSL_NO_ECX
Check for private key existence before calling eddsa sign functions Fixes #19524 Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/19525)
2022-10-28 11:47:50 +08:00
/* Test that trying to sign with a public key errors out gracefully */
static int test_ecx_not_private_key(int tst)
{
EVP_PKEY *pkey = NULL;
const unsigned char msg[] = { 0x00, 0x01, 0x02, 0x03 };
int testresult = 0;
EVP_MD_CTX *ctx = NULL;
unsigned char *mac = NULL;
size_t maclen = 0;
Multi-variant ML-KEM This introduces support for ML-KEM-512 and ML-KEM-1024 using the same underlying implementation parameterised by a few macros for the associated types and constants. KAT tests are added for ML-KEM 512 and 1024, to complement the previous tests for ML-KEM-768. MLKEM{512,768,1024} TLS "group" codepoints are updated to match the final IANA assigments and to make the additional KEMs known to the TLS layer. The pure-QC MLKEMs are not in the default list of supported groups, and need to be explicitly enabled by the application. Future work will introduce support for hybrids, and for more fine-grained policy of which keyshares a client should send by default, and when a server should request (HRR) a new mutually-supported group that was not sent. Tests for ML-KEM key exchange added to sslapitest to make sure that our TLS client MLKEM{512,768,1024} implementations interoperate with our TLS server, and that MLKEM* are not negotiated in TLS 1.2. Tests also added to excercise non-derandomised ML-KEM APIs, both directly (bypassing the provider layer), and through the generic EVP KEM API (exercising the provider). These make sure that RNG input is used correctly (KAT tests bypass the RNG by specifying seeds). The API interface to the provider takes an "const ML_KEM_VINFO" pointer, (obtained from ossl_ml_kem_get_vinfo()). This checks input and output buffer sizes before passing control to internal code that assumes correctly sized (for each variant) buffers. The original BoringSSL API was refactored to eliminate the opaque public/private key structure wrappers, since these structures are an internal detail between libcrypto and the provider, they are not part of the public (EVP) API. New "clangover" counter-measures added, refined with much appreciated input from David Benjamin (Chromium). The internal steps of "encrypt_cpa" were reordered to reduce the working-set size of the algorithm, now needs space for just two temporary "vectors" rather than three. The "decap" function now process the decrypted message in one call, rather than three separate calls to scalar_decode_1, scalar_decompress and scalar_add. Some loops were unrolled, improving performance of en/decapsulate (pre-expanded vectors and matrix) by around 5%. To handle, however unlikely, the SHA3 primitives not behaving like "pure" functions and failing, the implementation of `decap` was modifed: - To use the KDF to compute the Fujisaki-Okamoto (FO) failure secret first thing, and if that fails, bail out returning an error, a shared secret is still returned at random from the RNG, but it is OK for the caller to not use it. - If any of the subsequently used hash primitives fail, use the computed FO failure secret (OK, despite no longer constant-time) and return success (otherwise the RNG would replace the result). - We quite reasonably assume that chosen-ciphertext attacks (of the correct length) cannot cause hash functions to fail in a manner the depends on the private key content. Support for ML-KEM-512 required adding a centered binomial distribution helper function to deal with η_1 == 3 in just that variant. Some additional comments were added to highlight how the code relates to the ML-KEM specification in FIPS 203. Reviewed-by: Paul Dale <ppzgs1@gmail.com> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26172)
2024-11-30 22:20:58 +08:00
const uint8_t *pubkey;
Check for private key existence before calling eddsa sign functions Fixes #19524 Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/19525)
2022-10-28 11:47:50 +08:00
size_t pubkeylen;
switch (keys[tst].type) {
case NID_X25519:
case NID_X448:
Multi-variant ML-KEM This introduces support for ML-KEM-512 and ML-KEM-1024 using the same underlying implementation parameterised by a few macros for the associated types and constants. KAT tests are added for ML-KEM 512 and 1024, to complement the previous tests for ML-KEM-768. MLKEM{512,768,1024} TLS "group" codepoints are updated to match the final IANA assigments and to make the additional KEMs known to the TLS layer. The pure-QC MLKEMs are not in the default list of supported groups, and need to be explicitly enabled by the application. Future work will introduce support for hybrids, and for more fine-grained policy of which keyshares a client should send by default, and when a server should request (HRR) a new mutually-supported group that was not sent. Tests for ML-KEM key exchange added to sslapitest to make sure that our TLS client MLKEM{512,768,1024} implementations interoperate with our TLS server, and that MLKEM* are not negotiated in TLS 1.2. Tests also added to excercise non-derandomised ML-KEM APIs, both directly (bypassing the provider layer), and through the generic EVP KEM API (exercising the provider). These make sure that RNG input is used correctly (KAT tests bypass the RNG by specifying seeds). The API interface to the provider takes an "const ML_KEM_VINFO" pointer, (obtained from ossl_ml_kem_get_vinfo()). This checks input and output buffer sizes before passing control to internal code that assumes correctly sized (for each variant) buffers. The original BoringSSL API was refactored to eliminate the opaque public/private key structure wrappers, since these structures are an internal detail between libcrypto and the provider, they are not part of the public (EVP) API. New "clangover" counter-measures added, refined with much appreciated input from David Benjamin (Chromium). The internal steps of "encrypt_cpa" were reordered to reduce the working-set size of the algorithm, now needs space for just two temporary "vectors" rather than three. The "decap" function now process the decrypted message in one call, rather than three separate calls to scalar_decode_1, scalar_decompress and scalar_add. Some loops were unrolled, improving performance of en/decapsulate (pre-expanded vectors and matrix) by around 5%. To handle, however unlikely, the SHA3 primitives not behaving like "pure" functions and failing, the implementation of `decap` was modifed: - To use the KDF to compute the Fujisaki-Okamoto (FO) failure secret first thing, and if that fails, bail out returning an error, a shared secret is still returned at random from the RNG, but it is OK for the caller to not use it. - If any of the subsequently used hash primitives fail, use the computed FO failure secret (OK, despite no longer constant-time) and return success (otherwise the RNG would replace the result). - We quite reasonably assume that chosen-ciphertext attacks (of the correct length) cannot cause hash functions to fail in a manner the depends on the private key content. Support for ML-KEM-512 required adding a centered binomial distribution helper function to deal with η_1 == 3 in just that variant. Some additional comments were added to highlight how the code relates to the ML-KEM specification in FIPS 203. Reviewed-by: Paul Dale <ppzgs1@gmail.com> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26172)
2024-11-30 22:20:58 +08:00
case NID_undef:
return TEST_skip("signing not supported for X25519/X448/ML-KEM");
Check for private key existence before calling eddsa sign functions Fixes #19524 Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/19525)
2022-10-28 11:47:50 +08:00
}
/* Check if this algorithm supports public keys */
if (keys[tst].pub == NULL)
return TEST_skip("no public key present");
Multi-variant ML-KEM This introduces support for ML-KEM-512 and ML-KEM-1024 using the same underlying implementation parameterised by a few macros for the associated types and constants. KAT tests are added for ML-KEM 512 and 1024, to complement the previous tests for ML-KEM-768. MLKEM{512,768,1024} TLS "group" codepoints are updated to match the final IANA assigments and to make the additional KEMs known to the TLS layer. The pure-QC MLKEMs are not in the default list of supported groups, and need to be explicitly enabled by the application. Future work will introduce support for hybrids, and for more fine-grained policy of which keyshares a client should send by default, and when a server should request (HRR) a new mutually-supported group that was not sent. Tests for ML-KEM key exchange added to sslapitest to make sure that our TLS client MLKEM{512,768,1024} implementations interoperate with our TLS server, and that MLKEM* are not negotiated in TLS 1.2. Tests also added to excercise non-derandomised ML-KEM APIs, both directly (bypassing the provider layer), and through the generic EVP KEM API (exercising the provider). These make sure that RNG input is used correctly (KAT tests bypass the RNG by specifying seeds). The API interface to the provider takes an "const ML_KEM_VINFO" pointer, (obtained from ossl_ml_kem_get_vinfo()). This checks input and output buffer sizes before passing control to internal code that assumes correctly sized (for each variant) buffers. The original BoringSSL API was refactored to eliminate the opaque public/private key structure wrappers, since these structures are an internal detail between libcrypto and the provider, they are not part of the public (EVP) API. New "clangover" counter-measures added, refined with much appreciated input from David Benjamin (Chromium). The internal steps of "encrypt_cpa" were reordered to reduce the working-set size of the algorithm, now needs space for just two temporary "vectors" rather than three. The "decap" function now process the decrypted message in one call, rather than three separate calls to scalar_decode_1, scalar_decompress and scalar_add. Some loops were unrolled, improving performance of en/decapsulate (pre-expanded vectors and matrix) by around 5%. To handle, however unlikely, the SHA3 primitives not behaving like "pure" functions and failing, the implementation of `decap` was modifed: - To use the KDF to compute the Fujisaki-Okamoto (FO) failure secret first thing, and if that fails, bail out returning an error, a shared secret is still returned at random from the RNG, but it is OK for the caller to not use it. - If any of the subsequently used hash primitives fail, use the computed FO failure secret (OK, despite no longer constant-time) and return success (otherwise the RNG would replace the result). - We quite reasonably assume that chosen-ciphertext attacks (of the correct length) cannot cause hash functions to fail in a manner the depends on the private key content. Support for ML-KEM-512 required adding a centered binomial distribution helper function to deal with η_1 == 3 in just that variant. Some additional comments were added to highlight how the code relates to the ML-KEM specification in FIPS 203. Reviewed-by: Paul Dale <ppzgs1@gmail.com> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26172)
2024-11-30 22:20:58 +08:00
pubkey = keys[tst].pub;
pubkeylen = keys[tst].publen;
Check for private key existence before calling eddsa sign functions Fixes #19524 Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/19525)
2022-10-28 11:47:50 +08:00
pkey = EVP_PKEY_new_raw_public_key_ex(testctx, OBJ_nid2sn(keys[tst].type),
NULL, pubkey, pubkeylen);
if (!TEST_ptr(pkey))
goto err;
if (!TEST_ptr(ctx = EVP_MD_CTX_new()))
goto err;
if (EVP_DigestSignInit(ctx, NULL, NULL, NULL, pkey) != 1)
goto check_err;
if (EVP_DigestSign(ctx, NULL, &maclen, msg, sizeof(msg)) != 1)
goto check_err;
if (!TEST_ptr(mac = OPENSSL_malloc(maclen)))
goto err;
if (!TEST_int_eq(EVP_DigestSign(ctx, mac, &maclen, msg, sizeof(msg)), 0))
goto err;
check_err:
/*
* Currently only EVP_DigestSign will throw PROV_R_NOT_A_PRIVATE_KEY,
* but we relax the check to allow error also thrown by
* EVP_DigestSignInit and EVP_DigestSign.
*/
if (ERR_GET_REASON(ERR_peek_error()) == PROV_R_NOT_A_PRIVATE_KEY) {
testresult = 1;
ERR_clear_error();
}
err:
EVP_MD_CTX_free(ctx);
OPENSSL_free(mac);
EVP_PKEY_free(pkey);
return testresult;
}
configure: introduce no-ecx to remove ECX related feature This can effectively reduce the binary size for platforms that don't need ECX feature(~100KB). Signed-off-by: Yi Li <yi1.li@intel.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/20781)
2023-04-17 16:20:31 +08:00
#endif /* OPENSSL_NO_ECX */
Check for private key existence before calling eddsa sign functions Fixes #19524 Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/19525)
2022-10-28 11:47:50 +08:00
Add test for context duplication failure This checks that the first operation successfully completes even if context duplication fails. But follwing operations get errors as if the context was finlised. Signed-off-by: Simo Sorce <simo@redhat.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com> (Merged from https://github.com/openssl/openssl/pull/20375)
2023-03-07 07:38:00 +08:00
static int test_sign_continuation(void)
{
OSSL_PROVIDER *fake_rsa = NULL;
int testresult = 0;
EVP_PKEY *pkey = NULL;
EVP_PKEY_CTX *pctx = NULL;
EVP_MD_CTX *mctx = NULL;
const char sigbuf[] = "To Be Signed";
unsigned char signature[256];
size_t siglen = 256;
static int nodupnum = 1;
static const OSSL_PARAM nodup_params[] = {
OSSL_PARAM_int("NO_DUP", &nodupnum),
OSSL_PARAM_END
};
if (!TEST_ptr(fake_rsa = fake_rsa_start(testctx)))
return 0;
/* Construct a pkey using precise propq to use our provider */
if (!TEST_ptr(pctx = EVP_PKEY_CTX_new_from_name(testctx, "RSA",
"provider=fake-rsa"))
|| !TEST_true(EVP_PKEY_fromdata_init(pctx))
|| !TEST_true(EVP_PKEY_fromdata(pctx, &pkey, EVP_PKEY_KEYPAIR, NULL))
|| !TEST_ptr(pkey))
goto end;
/* First test it continues (classic behavior) */
if (!TEST_ptr(mctx = EVP_MD_CTX_new())
|| !TEST_true(EVP_DigestSignInit_ex(mctx, NULL, NULL, testctx,
NULL, pkey, NULL))
|| !TEST_true(EVP_DigestSignUpdate(mctx, sigbuf, sizeof(sigbuf)))
|| !TEST_true(EVP_DigestSignFinal(mctx, signature, &siglen))
|| !TEST_true(EVP_DigestSignUpdate(mctx, sigbuf, sizeof(sigbuf)))
|| !TEST_true(EVP_DigestSignFinal(mctx, signature, &siglen)))
goto end;
EVP_MD_CTX_free(mctx);
/* try again but failing the continuation */
if (!TEST_ptr(mctx = EVP_MD_CTX_new())
|| !TEST_true(EVP_DigestSignInit_ex(mctx, NULL, NULL, testctx,
NULL, pkey, nodup_params))
|| !TEST_true(EVP_DigestSignUpdate(mctx, sigbuf, sizeof(sigbuf)))
|| !TEST_true(EVP_DigestSignFinal(mctx, signature, &siglen))
|| !TEST_false(EVP_DigestSignUpdate(mctx, sigbuf, sizeof(sigbuf)))
|| !TEST_false(EVP_DigestSignFinal(mctx, signature, &siglen)))
goto end;
testresult = 1;
end:
EVP_MD_CTX_free(mctx);
EVP_PKEY_free(pkey);
EVP_PKEY_CTX_free(pctx);
fake_rsa_finish(fake_rsa);
return testresult;
}
test: add unit test for CVE-2023-5363 Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Hugo Landau <hlandau@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org>
2023-10-06 07:32:20 +08:00
static int aes_gcm_encrypt(const unsigned char *gcm_key, size_t gcm_key_s,
const unsigned char *gcm_iv, size_t gcm_ivlen,
const unsigned char *gcm_pt, size_t gcm_pt_s,
const unsigned char *gcm_aad, size_t gcm_aad_s,
const unsigned char *gcm_ct, size_t gcm_ct_s,
const unsigned char *gcm_tag, size_t gcm_tag_s)
{
int ret = 0;
EVP_CIPHER_CTX *ctx;
EVP_CIPHER *cipher = NULL;
int outlen, tmplen;
unsigned char outbuf[1024];
unsigned char outtag[16];
OSSL_PARAM params[2] = {
OSSL_PARAM_END, OSSL_PARAM_END
};
if (!TEST_ptr(ctx = EVP_CIPHER_CTX_new())
|| !TEST_ptr(cipher = EVP_CIPHER_fetch(testctx, "AES-256-GCM", "")))
goto err;
params[0] = OSSL_PARAM_construct_size_t(OSSL_CIPHER_PARAM_AEAD_IVLEN,
&gcm_ivlen);
if (!TEST_true(EVP_EncryptInit_ex2(ctx, cipher, gcm_key, gcm_iv, params))
|| (gcm_aad != NULL
&& !TEST_true(EVP_EncryptUpdate(ctx, NULL, &outlen,
gcm_aad, gcm_aad_s)))
|| !TEST_true(EVP_EncryptUpdate(ctx, outbuf, &outlen,
gcm_pt, gcm_pt_s))
|| !TEST_true(EVP_EncryptFinal_ex(ctx, outbuf, &tmplen)))
goto err;
params[0] = OSSL_PARAM_construct_octet_string(OSSL_CIPHER_PARAM_AEAD_TAG,
outtag, sizeof(outtag));
if (!TEST_true(EVP_CIPHER_CTX_get_params(ctx, params))
|| !TEST_mem_eq(outbuf, outlen, gcm_ct, gcm_ct_s)
|| !TEST_mem_eq(outtag, gcm_tag_s, gcm_tag, gcm_tag_s))
goto err;
ret = 1;
err:
EVP_CIPHER_free(cipher);
EVP_CIPHER_CTX_free(ctx);
return ret;
}
static int aes_gcm_decrypt(const unsigned char *gcm_key, size_t gcm_key_s,
const unsigned char *gcm_iv, size_t gcm_ivlen,
const unsigned char *gcm_pt, size_t gcm_pt_s,
const unsigned char *gcm_aad, size_t gcm_aad_s,
const unsigned char *gcm_ct, size_t gcm_ct_s,
const unsigned char *gcm_tag, size_t gcm_tag_s)
{
int ret = 0;
EVP_CIPHER_CTX *ctx;
EVP_CIPHER *cipher = NULL;
int outlen;
unsigned char outbuf[1024];
OSSL_PARAM params[2] = {
OSSL_PARAM_END, OSSL_PARAM_END
};
if ((ctx = EVP_CIPHER_CTX_new()) == NULL)
goto err;
if ((cipher = EVP_CIPHER_fetch(testctx, "AES-256-GCM", "")) == NULL)
goto err;
params[0] = OSSL_PARAM_construct_size_t(OSSL_CIPHER_PARAM_AEAD_IVLEN,
&gcm_ivlen);
if (!TEST_true(EVP_DecryptInit_ex2(ctx, cipher, gcm_key, gcm_iv, params))
|| (gcm_aad != NULL
&& !TEST_true(EVP_DecryptUpdate(ctx, NULL, &outlen,
gcm_aad, gcm_aad_s)))
|| !TEST_true(EVP_DecryptUpdate(ctx, outbuf, &outlen,
gcm_ct, gcm_ct_s))
|| !TEST_mem_eq(outbuf, outlen, gcm_pt, gcm_pt_s))
goto err;
params[0] = OSSL_PARAM_construct_octet_string(OSSL_CIPHER_PARAM_AEAD_TAG,
(void*)gcm_tag, gcm_tag_s);
if (!TEST_true(EVP_CIPHER_CTX_set_params(ctx, params))
||!TEST_true(EVP_DecryptFinal_ex(ctx, outbuf, &outlen)))
goto err;
ret = 1;
err:
EVP_CIPHER_free(cipher);
EVP_CIPHER_CTX_free(ctx);
return ret;
}
static int test_aes_gcm_ivlen_change_cve_2023_5363(void)
{
/* AES-GCM test data obtained from NIST public test vectors */
static const unsigned char gcm_key[] = {
0xd0, 0xc2, 0x67, 0xc1, 0x9f, 0x30, 0xd8, 0x0b, 0x89, 0x14, 0xbb, 0xbf,
0xb7, 0x2f, 0x73, 0xb8, 0xd3, 0xcd, 0x5f, 0x6a, 0x78, 0x70, 0x15, 0x84,
0x8a, 0x7b, 0x30, 0xe3, 0x8f, 0x16, 0xf1, 0x8b,
};
static const unsigned char gcm_iv[] = {
0xb6, 0xdc, 0xda, 0x95, 0xac, 0x99, 0x77, 0x76, 0x25, 0xae, 0x87, 0xf8,
0xa3, 0xa9, 0xdd, 0x64, 0xd7, 0x9b, 0xbd, 0x5f, 0x4a, 0x0e, 0x54, 0xca,
0x1a, 0x9f, 0xa2, 0xe3, 0xf4, 0x5f, 0x5f, 0xc2, 0xce, 0xa7, 0xb6, 0x14,
0x12, 0x6f, 0xf0, 0xaf, 0xfd, 0x3e, 0x17, 0x35, 0x6e, 0xa0, 0x16, 0x09,
0xdd, 0xa1, 0x3f, 0xd8, 0xdd, 0xf3, 0xdf, 0x4f, 0xcb, 0x18, 0x49, 0xb8,
0xb3, 0x69, 0x2c, 0x5d, 0x4f, 0xad, 0x30, 0x91, 0x08, 0xbc, 0xbe, 0x24,
0x01, 0x0f, 0xbe, 0x9c, 0xfb, 0x4f, 0x5d, 0x19, 0x7f, 0x4c, 0x53, 0xb0,
0x95, 0x90, 0xac, 0x7b, 0x1f, 0x7b, 0xa0, 0x99, 0xe1, 0xf3, 0x48, 0x54,
0xd0, 0xfc, 0xa9, 0xcc, 0x91, 0xf8, 0x1f, 0x9b, 0x6c, 0x9a, 0xe0, 0xdc,
0x63, 0xea, 0x7d, 0x2a, 0x4a, 0x7d, 0xa5, 0xed, 0x68, 0x57, 0x27, 0x6b,
0x68, 0xe0, 0xf2, 0xb8, 0x51, 0x50, 0x8d, 0x3d,
};
static const unsigned char gcm_pt[] = {
0xb8, 0xb6, 0x88, 0x36, 0x44, 0xe2, 0x34, 0xdf, 0x24, 0x32, 0x91, 0x07,
0x4f, 0xe3, 0x6f, 0x81,
};
static const unsigned char gcm_ct[] = {
0xff, 0x4f, 0xb3, 0xf3, 0xf9, 0xa2, 0x51, 0xd4, 0x82, 0xc2, 0xbe, 0xf3,
0xe2, 0xd0, 0xec, 0xed,
};
static const unsigned char gcm_tag[] = {
0xbd, 0x06, 0x38, 0x09, 0xf7, 0xe1, 0xc4, 0x72, 0x0e, 0xf2, 0xea, 0x63,
0xdb, 0x99, 0x6c, 0x21,
};
return aes_gcm_encrypt(gcm_key, sizeof(gcm_key), gcm_iv, sizeof(gcm_iv),
gcm_pt, sizeof(gcm_pt), NULL, 0,
gcm_ct, sizeof(gcm_ct), gcm_tag, sizeof(gcm_tag))
&& aes_gcm_decrypt(gcm_key, sizeof(gcm_key), gcm_iv, sizeof(gcm_iv),
gcm_pt, sizeof(gcm_pt), NULL, 0,
gcm_ct, sizeof(gcm_ct), gcm_tag, sizeof(gcm_tag));
}
#ifndef OPENSSL_NO_RC4
static int rc4_encrypt(const unsigned char *rc4_key, size_t rc4_key_s,
const unsigned char *rc4_pt, size_t rc4_pt_s,
const unsigned char *rc4_ct, size_t rc4_ct_s)
{
int ret = 0;
EVP_CIPHER_CTX *ctx;
EVP_CIPHER *cipher = NULL;
int outlen, tmplen;
unsigned char outbuf[1024];
OSSL_PARAM params[2] = {
OSSL_PARAM_END, OSSL_PARAM_END
};
if (!TEST_ptr(ctx = EVP_CIPHER_CTX_new())
|| !TEST_ptr(cipher = EVP_CIPHER_fetch(testctx, "RC4", "")))
goto err;
params[0] = OSSL_PARAM_construct_size_t(OSSL_CIPHER_PARAM_KEYLEN,
&rc4_key_s);
if (!TEST_true(EVP_EncryptInit_ex2(ctx, cipher, rc4_key, NULL, params))
|| !TEST_true(EVP_EncryptUpdate(ctx, outbuf, &outlen,
rc4_pt, rc4_pt_s))
|| !TEST_true(EVP_EncryptFinal_ex(ctx, outbuf, &tmplen)))
goto err;
if (!TEST_mem_eq(outbuf, outlen, rc4_ct, rc4_ct_s))
goto err;
ret = 1;
err:
EVP_CIPHER_free(cipher);
EVP_CIPHER_CTX_free(ctx);
return ret;
}
static int rc4_decrypt(const unsigned char *rc4_key, size_t rc4_key_s,
const unsigned char *rc4_pt, size_t rc4_pt_s,
const unsigned char *rc4_ct, size_t rc4_ct_s)
{
int ret = 0;
EVP_CIPHER_CTX *ctx;
EVP_CIPHER *cipher = NULL;
int outlen;
unsigned char outbuf[1024];
OSSL_PARAM params[2] = {
OSSL_PARAM_END, OSSL_PARAM_END
};
if ((ctx = EVP_CIPHER_CTX_new()) == NULL)
goto err;
if ((cipher = EVP_CIPHER_fetch(testctx, "RC4", "")) == NULL)
goto err;
params[0] = OSSL_PARAM_construct_size_t(OSSL_CIPHER_PARAM_KEYLEN,
&rc4_key_s);
if (!TEST_true(EVP_DecryptInit_ex2(ctx, cipher, rc4_key, NULL, params))
|| !TEST_true(EVP_DecryptUpdate(ctx, outbuf, &outlen,
rc4_ct, rc4_ct_s))
|| !TEST_mem_eq(outbuf, outlen, rc4_pt, rc4_pt_s))
goto err;
ret = 1;
err:
EVP_CIPHER_free(cipher);
EVP_CIPHER_CTX_free(ctx);
return ret;
}
static int test_aes_rc4_keylen_change_cve_2023_5363(void)
{
/* RC4 test data obtained from RFC 6229 */
static const struct {
unsigned char key[5];
unsigned char padding[11];
} rc4_key = {
{ /* Five bytes of key material */
0x83, 0x32, 0x22, 0x77, 0x2a,
},
{ /* Random padding to 16 bytes */
0x80, 0xad, 0x97, 0xbd, 0xc9, 0x73, 0xdf, 0x8a, 0xaa, 0x32, 0x91
}
};
static const unsigned char rc4_pt[] = {
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00
};
static const unsigned char rc4_ct[] = {
0x80, 0xad, 0x97, 0xbd, 0xc9, 0x73, 0xdf, 0x8a,
0x2e, 0x87, 0x9e, 0x92, 0xa4, 0x97, 0xef, 0xda
};
if (lgcyprov == NULL)
return TEST_skip("Test requires legacy provider to be loaded");
return rc4_encrypt(rc4_key.key, sizeof(rc4_key.key),
rc4_pt, sizeof(rc4_pt), rc4_ct, sizeof(rc4_ct))
&& rc4_decrypt(rc4_key.key, sizeof(rc4_key.key),
rc4_pt, sizeof(rc4_pt), rc4_ct, sizeof(rc4_ct));
}
#endif
EVP_DigestUpdate(): Check if ctx->update is set The issue has been discovered by libFuzzer running on provider target. There are currently three distinct reports which are addressed by code change here. https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=69236#c1 https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=69243#c1 https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=69261#c1 the issue has been introduced with openssl 3.0. Reviewed-by: Neil Horman <nhorman@openssl.org> Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/24753)
2024-06-27 22:31:41 +08:00
static int test_invalid_ctx_for_digest(void)
{
int ret;
EVP_MD_CTX *mdctx;
mdctx = EVP_MD_CTX_new();
if (!TEST_ptr(mdctx))
return 0;
if (!TEST_int_eq(EVP_DigestUpdate(mdctx, "test", sizeof("test") - 1), 0))
ret = 0;
else
ret = 1;
EVP_MD_CTX_free(mdctx);
return ret;
}
test: add evp_extra_test case for cipher pipeline API with fake pipeline provider Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/24636)
2024-08-31 20:03:49 +08:00
static int test_evp_cipher_pipeline(void)
{
OSSL_PROVIDER *fake_pipeline = NULL;
int testresult = 0;
EVP_CIPHER *cipher = NULL;
EVP_CIPHER *pipeline_cipher = NULL;
EVP_CIPHER_CTX *ctx = NULL;
unsigned char key[32];
size_t keylen = 32;
size_t ivlen = EVP_GCM_TLS_EXPLICIT_IV_LEN + EVP_GCM_TLS_FIXED_IV_LEN;
size_t taglen = EVP_GCM_TLS_TAG_LEN;
unsigned char *iv_array[EVP_MAX_PIPES], *tag_array[EVP_MAX_PIPES];
unsigned char *plaintext_array[EVP_MAX_PIPES];
unsigned char *ciphertext_array_p[EVP_MAX_PIPES];
void **aead_tags = (void **)&tag_array;
unsigned char *temp[EVP_MAX_PIPES];
size_t outsize_array[EVP_MAX_PIPES], outlen_array[EVP_MAX_PIPES];
size_t ciphertextlen_array[EVP_MAX_PIPES];
size_t inlen_array[EVP_MAX_PIPES];
OSSL_PARAM params[2] = { OSSL_PARAM_END, OSSL_PARAM_END };
test_evp_cipher_pipeline(): Fix memory leaks on errors Fixes Coverity 1636844, 1636845 Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Saša Nedvědický <sashan@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26206)
2024-12-18 16:21:10 +08:00
unsigned char *ciphertext, *exp_plaintext, *tag;
test: add evp_extra_test case for cipher pipeline API with fake pipeline provider Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/24636)
2024-08-31 20:03:49 +08:00
size_t numpipes, plaintextlen, i;
if (!TEST_ptr(fake_pipeline = fake_pipeline_start(testctx)))
return 0;
if (!TEST_ptr(pipeline_cipher = EVP_CIPHER_fetch(testctx, "AES-256-GCM",
"provider=fake-pipeline"))
Fix the sporadic test failure in 30-test_evp_extra.t Fixes #26276 Reviewed-by: Saša Nedvědický <sashan@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26269)
2024-12-29 23:43:34 +08:00
|| !TEST_ptr(cipher = EVP_CIPHER_fetch(testctx, "AES-256-GCM",
"provider!=fake-pipeline"))
test: add evp_extra_test case for cipher pipeline API with fake pipeline provider Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/24636)
2024-08-31 20:03:49 +08:00
|| !TEST_ptr(ctx = EVP_CIPHER_CTX_new()))
goto end;
memset(key, 0x01, sizeof(key));
/* Negative tests */
if (!TEST_false(EVP_CIPHER_can_pipeline(cipher, 1)))
goto end;
if (!TEST_false(EVP_CIPHER_can_pipeline(EVP_aes_256_gcm(), 1)))
goto end;
if (!TEST_false(EVP_CipherPipelineEncryptInit(ctx, pipeline_cipher,
key, keylen,
EVP_MAX_PIPES + 1, NULL, 0)))
goto end;
/* Positive tests */
for (numpipes = 1; numpipes <= EVP_MAX_PIPES; numpipes++) {
for (plaintextlen = 1; plaintextlen <= 256; plaintextlen++) {
size_t ciphertextlen = 0;
int outlen = 0;
test_evp_cipher_pipeline(): Fix memory leaks on errors Fixes Coverity 1636844, 1636845 Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Saša Nedvědický <sashan@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26206)
2024-12-18 16:21:10 +08:00
/* Cleanup to be able to error out */
memset(iv_array, 0, sizeof(iv_array));
memset(plaintext_array, 0, sizeof(plaintext_array));
memset(ciphertext_array_p, 0, sizeof(ciphertext_array_p));
memset(tag_array, 0, sizeof(tag_array));
ciphertext = NULL;
exp_plaintext = NULL;
tag = NULL;
test: add evp_extra_test case for cipher pipeline API with fake pipeline provider Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/24636)
2024-08-31 20:03:49 +08:00
/* Allocate fresh buffers with exact size to catch buffer overwrites */
for (i = 0; i < numpipes; i++) {
if (!TEST_ptr(iv_array[i] = OPENSSL_malloc(ivlen))
|| !TEST_ptr(plaintext_array[i] = OPENSSL_malloc(plaintextlen))
|| !TEST_ptr(ciphertext_array_p[i] =
OPENSSL_malloc(plaintextlen + EVP_MAX_BLOCK_LENGTH))
|| !TEST_ptr(tag_array[i] = OPENSSL_malloc(taglen)))
test_evp_cipher_pipeline(): Fix memory leaks on errors Fixes Coverity 1636844, 1636845 Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Saša Nedvědický <sashan@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26206)
2024-12-18 16:21:10 +08:00
goto err;
test: add evp_extra_test case for cipher pipeline API with fake pipeline provider Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/24636)
2024-08-31 20:03:49 +08:00
memset(iv_array[i], i + 33, ivlen);
memset(plaintext_array[i], i + 1, plaintextlen);
inlen_array[i] = plaintextlen;
outlen_array[i] = 0;
ciphertextlen_array[i] = 0;
outsize_array[i] = plaintextlen + EVP_MAX_BLOCK_LENGTH;
}
if (!TEST_ptr(ciphertext =
OPENSSL_malloc(plaintextlen + EVP_MAX_BLOCK_LENGTH))
|| !TEST_ptr(tag = OPENSSL_malloc(taglen))
|| !TEST_ptr(exp_plaintext = OPENSSL_malloc(plaintextlen)))
test_evp_cipher_pipeline(): Fix memory leaks on errors Fixes Coverity 1636844, 1636845 Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Saša Nedvědický <sashan@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26206)
2024-12-18 16:21:10 +08:00
goto err;
test: add evp_extra_test case for cipher pipeline API with fake pipeline provider Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/24636)
2024-08-31 20:03:49 +08:00
/* Encrypt using pipeline API */
if (!TEST_true(EVP_CIPHER_CTX_reset(ctx))
|| !TEST_true(EVP_CIPHER_can_pipeline(pipeline_cipher, 1))
|| !TEST_true(EVP_CipherPipelineEncryptInit(ctx, pipeline_cipher,
key, keylen, numpipes,
(const unsigned char **)iv_array,
ivlen))
/* reuse plaintext for AAD as it won't affect test */
|| !TEST_true(EVP_CipherPipelineUpdate(ctx, NULL, outlen_array, NULL,
(const unsigned char **)plaintext_array,
inlen_array))
|| !TEST_true(EVP_CipherPipelineUpdate(ctx, ciphertext_array_p,
outlen_array, outsize_array,
(const unsigned char **)plaintext_array,
inlen_array)))
goto err;
for (i = 0; i < numpipes; i++) {
ciphertextlen_array[i] = outlen_array[i];
temp[i] = ciphertext_array_p[i] + ciphertextlen_array[i];
outsize_array[i] = outsize_array[i] - ciphertextlen_array[i];
}
if (!TEST_true(EVP_CipherPipelineFinal(ctx, temp, outlen_array, outsize_array)))
goto err;
for (i = 0; i < numpipes; i++)
ciphertextlen_array[i] += outlen_array[i];
params[0] = OSSL_PARAM_construct_octet_ptr(OSSL_CIPHER_PARAM_PIPELINE_AEAD_TAG,
(void **)&aead_tags, taglen);
if (!TEST_true(EVP_CIPHER_CTX_get_params(ctx, params)))
goto err;
/* Encrypt using non-pipeline API and compare */
if (!TEST_true(EVP_CIPHER_CTX_reset(ctx)))
goto err;
for (i = 0; i < numpipes; i++) {
if (!TEST_true(EVP_EncryptInit(ctx, cipher, key, iv_array[i]))
|| !TEST_true(EVP_EncryptUpdate(ctx, NULL, &outlen,
plaintext_array[i],
plaintextlen))
|| !TEST_true(EVP_EncryptUpdate(ctx, ciphertext, &outlen,
plaintext_array[i],
plaintextlen)))
goto err;
ciphertextlen = outlen;
if (!TEST_true(EVP_EncryptFinal_ex(ctx, ciphertext + outlen, &outlen)))
goto err;
ciphertextlen += outlen;
params[0] = OSSL_PARAM_construct_octet_string(OSSL_CIPHER_PARAM_AEAD_TAG,
(void *)tag, taglen);
if (!TEST_true(EVP_CIPHER_CTX_get_params(ctx, params)))
goto err;
if (!TEST_mem_eq(ciphertext_array_p[i], ciphertextlen_array[i],
ciphertext, ciphertextlen)
|| !TEST_mem_eq(tag_array[i], taglen, tag, taglen))
goto err;
}
for (i = 0; i < numpipes; i++)
outsize_array[i] = plaintextlen;
/* Decrypt using pipeline API and compare */
params[0] = OSSL_PARAM_construct_octet_ptr(OSSL_CIPHER_PARAM_PIPELINE_AEAD_TAG,
(void **)&aead_tags, taglen);
if (!TEST_true(EVP_CIPHER_CTX_reset(ctx))
|| !TEST_true(EVP_CIPHER_can_pipeline(pipeline_cipher, 0))
|| !TEST_true(EVP_CipherPipelineDecryptInit(ctx, pipeline_cipher,
key, keylen, numpipes,
(const unsigned char **)iv_array,
ivlen))
|| !TEST_true(EVP_CIPHER_CTX_set_params(ctx, params))
|| !TEST_true(EVP_CipherPipelineUpdate(ctx, NULL, outlen_array, NULL,
(const unsigned char **)plaintext_array,
inlen_array))
|| !TEST_true(EVP_CipherPipelineUpdate(ctx, plaintext_array,
outlen_array, outsize_array,
(const unsigned char **)ciphertext_array_p,
ciphertextlen_array)))
goto err;
for (i = 0; i < numpipes; i++) {
temp[i] = plaintext_array[i] + outlen_array[i];
outsize_array[i] = outsize_array[i] - outlen_array[i];
}
if (!TEST_true(EVP_CipherPipelineFinal(ctx, temp, outlen_array, outsize_array)))
goto err;
for (i = 0; i < numpipes; i++) {
memset(exp_plaintext, i + 1, plaintextlen);
if (!TEST_mem_eq(plaintext_array[i], plaintextlen,
exp_plaintext, plaintextlen))
goto err;
}
for (i = 0; i < numpipes; i++) {
OPENSSL_free(iv_array[i]);
OPENSSL_free(plaintext_array[i]);
OPENSSL_free(ciphertext_array_p[i]);
OPENSSL_free(tag_array[i]);
}
OPENSSL_free(exp_plaintext);
OPENSSL_free(ciphertext);
OPENSSL_free(tag);
}
}
testresult = 1;
goto end;
err:
for (i = 0; i < numpipes; i++) {
OPENSSL_free(iv_array[i]);
OPENSSL_free(plaintext_array[i]);
OPENSSL_free(ciphertext_array_p[i]);
OPENSSL_free(tag_array[i]);
}
OPENSSL_free(exp_plaintext);
OPENSSL_free(ciphertext);
OPENSSL_free(tag);
end:
EVP_CIPHER_CTX_free(ctx);
EVP_CIPHER_free(cipher);
EVP_CIPHER_free(pipeline_cipher);
fake_pipeline_finish(fake_pipeline);
return testresult;
}
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
int setup_tests(void)
{
EVP_get_default_properties - tests Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/25434)
2024-09-11 22:48:44 +08:00
char *config_file = NULL;
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
OPTION_CHOICE o;
while ((o = opt_next()) != OPT_EOF) {
switch (o) {
case OPT_CONTEXT:
/* Set up an alternate library context */
testctx = OSSL_LIB_CTX_new();
if (!TEST_ptr(testctx))
return 0;
Statically link legacy provider to evp_extra_test Like in #17345, evp_extra_test links libcrypto statically, but also has a dynamic/shared load via the legacy provider, which leads to ambiguous behavior in evp_extra_test on some platforms, usually a crash (SIGSEGV) on exit via the atexit handlers. Statically link the legacy provider to avoid this. Fixes #22819 Helped-by: Neil Horman <nhorman@openssl.org> Helped-by: Tomas Mraz <tomas@openssl.org> Signed-off-by: Randall S. Becker <randall.becker@nexbridge.ca> Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/22905)
2023-12-01 00:20:34 +08:00
#ifdef STATIC_LEGACY
test_evp_cipher_pipeline(): Fix memory leaks on errors Fixes Coverity 1636844, 1636845 Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Saša Nedvědický <sashan@openssl.org> (Merged from https://github.com/openssl/openssl/pull/26206)
2024-12-18 16:21:10 +08:00
/*
* This test is always statically linked against libcrypto. We must not
* attempt to load legacy.so that might be dynamically linked against
* libcrypto. Instead we use a built-in version of the legacy provider.
*/
if (!OSSL_PROVIDER_add_builtin(testctx, "legacy", ossl_legacy_provider_init))
return 0;
Statically link legacy provider to evp_extra_test Like in #17345, evp_extra_test links libcrypto statically, but also has a dynamic/shared load via the legacy provider, which leads to ambiguous behavior in evp_extra_test on some platforms, usually a crash (SIGSEGV) on exit via the atexit handlers. Statically link the legacy provider to avoid this. Fixes #22819 Helped-by: Neil Horman <nhorman@openssl.org> Helped-by: Tomas Mraz <tomas@openssl.org> Signed-off-by: Randall S. Becker <randall.becker@nexbridge.ca> Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/22905)
2023-12-01 00:20:34 +08:00
#endif
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
/* Swap the libctx to test non-default context only */
nullprov = OSSL_PROVIDER_load(NULL, "null");
deflprov = OSSL_PROVIDER_load(testctx, "default");
Workaround crash in atexit on NonStop platforms We cannot dynamically load the legacy provider into an application that is linked statically to libcrypto as this causes a double loading of libcrypto (one static and one dynamic) and on NonStop this leads to a segfault in atexit(). Fixes #17537 Reviewed-by: Hugo Landau <hlandau@openssl.org> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/19844)
2022-12-06 17:52:47 +08:00
#ifndef OPENSSL_SYS_TANDEM
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
lgcyprov = OSSL_PROVIDER_load(testctx, "legacy");
Workaround crash in atexit on NonStop platforms We cannot dynamically load the legacy provider into an application that is linked statically to libcrypto as this causes a double loading of libcrypto (one static and one dynamic) and on NonStop this leads to a segfault in atexit(). Fixes #17537 Reviewed-by: Hugo Landau <hlandau@openssl.org> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/19844)
2022-12-06 17:52:47 +08:00
#endif
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
break;
EVP_get_default_properties - tests Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/25434)
2024-09-11 22:48:44 +08:00
case OPT_CONFIG_FILE:
config_file = opt_arg();
if (!test_get_libctx(&testctx, &nullprov, config_file, NULL, NULL))
return 0;
break;
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
case OPT_TEST_CASES:
break;
default:
return 0;
}
}
Add some tests for the newly added raw private/public key functions We already had soem tests for the older raw private/public key functions so we expand those to call the new versions as well and pass in a libctx. Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/11635)
2020-04-28 00:17:05 +08:00
EVP_get_default_properties - tests Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/25434)
2024-09-11 22:48:44 +08:00
if (config_file != NULL) {
ADD_TEST(test_EVP_set_config_properties);
return 1;
}
property: Move global default properties to the library context. Fixes a problem where global properties don't work with a NULL query. Specifying an algorithm with a NULL query ignores the default properties. Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/12123)
2020-06-12 08:34:46 +08:00
ADD_TEST(test_EVP_set_default_properties);
test: Add testing of reinitialization via EVP_DigestSignInit() Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16964)
2021-11-04 22:35:40 +08:00
ADD_ALL_TESTS(test_EVP_DigestSignInit, 30);
Convert more tests ct_test,evp_extra_test,wpackettest,packettest Add strncmp TEST wrappers And make some style/consistency fixes to ct_test Silence travis; gcc bug? Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/3234)
2017-04-19 04:33:15 +08:00
ADD_TEST(test_EVP_DigestVerifyInit);
evp_extra_test: Add SIPHASH MAC digestsign test with reinitialization Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16964)
2021-11-05 21:04:25 +08:00
#ifndef OPENSSL_NO_SIPHASH
ADD_TEST(test_siphash_digestsign);
#endif
Add test for no reset after DigestFinal_ex and DigestFinalXOF Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com> Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/13402)
2020-11-13 21:16:35 +08:00
ADD_TEST(test_EVP_Digest);
Add null digest implementation to the default provider This is necessary to keep compatibility with 1.1.1. Fixes #16660 Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/17016)
2021-11-12 23:31:35 +08:00
ADD_TEST(test_EVP_md_null);
Test short buffers Test that calling EVP_DigestSign(), EVP_DigestSignFinal(), EVP_PKEY_sign(), EVP_PKEY_get_raw_private_key(), or EVP_PKEY_get_raw_public_key() with a short output buffer results in a failure. Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16789)
2021-10-07 21:15:47 +08:00
ADD_ALL_TESTS(test_EVP_PKEY_sign, 3);
test: evp_extra: test signing with legacy app method based keys This commit adds `test_EVP_PKEY_sign_with_app_method`, a regression test for the bug fix in commit 1acc3e8cc3c6 (pull request #22163). It is analogous to `test_EVP_PKEY_sign`, only with a fake app method based key. (The EC key test case was omitted, because there is no `EC_KEY_METHOD_dup` method.) Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/22185)
2023-09-24 07:13:20 +08:00
#ifndef OPENSSL_NO_DEPRECATED_3_0
ADD_ALL_TESTS(test_EVP_PKEY_sign_with_app_method, 2);
#endif
seal: make EVP_SealInit() library context aware Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/15300)
2021-05-17 10:18:53 +08:00
ADD_ALL_TESTS(test_EVP_Enveloped, 2);
Update the test framework so that the need for test_main is removed. Everything that needed test_main now works using the same infrastructure as tests that used register_tests. This meant: * renaming register_tests to setup_tests and giving it a success/failure return. * renaming the init_test function to setup_test_framework. * renaming the finish_test function to pulldown_test_framework. * adding a user provided global_init function that runs before the test frame work is initialised. It returns a failure indication that stops the stest. * adding helper functions that permit tests to access their command line args. * spliting the BIO initialisation and finalisation out from the test setup and teardown. * hiding some of the now test internal functions. * fix the comments in testutil.h Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/3953)
2017-07-18 09:48:27 +08:00
ADD_ALL_TESTS(test_d2i_AutoPrivateKey, OSSL_NELEM(keydata));
TEST: Test i2d_PKCS8PrivateKey_bio() and PEM_write_bio_PKCS8PrivateKey() Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/11855)
2020-05-18 14:35:29 +08:00
ADD_TEST(test_privatekey_to_pkcs8);
Mark pop/clear error stack in der2key_decode_p8 This commit sets the error mark before calling d2i_X509_SIG and clear it if that function call is successful. The motivation for this is that if d2i_X509_SIG returns NULL then the else clause will be entered and d2i_PKCS8_PRIV_KEY_INFO will be called. If d2i_X509_SIG raised any errors those error will be on the error stack when d2i_PKCS8_PRIV_KEY_INFO gets called, and even if it returns successfully those errors will still be on the error stack. We ran into this issue when upgrading Node.js to 3.0.0-alpha15. More details can be found in the ref links below. Refs: https://github.com/nodejs/node/issues/38373 Refs: https://github.com/danbev/learning-libcrypto/blob/master/notes/wrong-tag-issue2.md Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Ben Kaduk <kaduk@mit.edu> (Merged from https://github.com/openssl/openssl/pull/15067)
2021-04-28 16:30:13 +08:00
ADD_TEST(test_EVP_PKCS82PKEY_wrong_tag);
Fix evp_extra_test.c with no-ec When OpenSSL is configured with no-ec, then the new evp_extra_test fails to pass. This change adds appropriate OPENSSL_NO_EC guards around the code. Reviewed-by: Tim Hudson <tjh@openssl.org>
2015-02-26 18:35:50 +08:00
#ifndef OPENSSL_NO_EC
Convert more tests ct_test,evp_extra_test,wpackettest,packettest Add strncmp TEST wrappers And make some style/consistency fixes to ct_test Silence travis; gcc bug? Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/3234)
2017-04-19 04:33:15 +08:00
ADD_TEST(test_EVP_PKCS82PKEY);
Move SM2 algos to SM2 specific PKEY method Use EVP_PKEY_set_alias_type to access Reviewed-by: Andy Polyakov <appro@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6443)
2018-06-19 03:51:56 +08:00
#endif
TEST: Add a test of EC key generation with encoding spec The primary function of this test is to exercise EVP_PKEY_CTX_set_ec_param_enc(). Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com> Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/12853)
2020-09-11 14:36:54 +08:00
#ifndef OPENSSL_NO_EC
ADD_ALL_TESTS(test_EC_keygen_with_enc, OSSL_NELEM(ec_encodings));
#endif
Remove FIPS condition on SM2 test. Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/19237)
2022-09-19 08:28:25 +08:00
#if !defined(OPENSSL_NO_SM2)
Move SM2 algos to SM2 specific PKEY method Use EVP_PKEY_set_alias_type to access Reviewed-by: Andy Polyakov <appro@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6443)
2018-06-19 03:51:56 +08:00
ADD_TEST(test_EVP_SM2);
Add test case for SM2 evp verification This test case is originally submitted in #6757, by Jack Lloyd. The test case has been modified to use the a different method to set the ID when computing the Z hash of SM2 signature. Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Yang <yang.yang@baishancloud.com> (Merged from https://github.com/openssl/openssl/pull/7113)
2018-09-04 23:25:29 +08:00
ADD_TEST(test_EVP_SM2_verify);
Fix evp_extra_test.c with no-ec When OpenSSL is configured with no-ec, then the new evp_extra_test fails to pass. This change adds appropriate OPENSSL_NO_EC guards around the code. Reviewed-by: Tim Hudson <tjh@openssl.org>
2015-02-26 18:35:50 +08:00
#endif
Add a test for the raw private/public key getters Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/6394)
2018-06-01 22:06:52 +08:00
ADD_ALL_TESTS(test_set_get_raw_keys, OSSL_NELEM(keys));
EVP: deprecate the EVP_X_meth_ functions. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/11082)
2020-02-13 09:00:57 +08:00
#ifndef OPENSSL_NO_DEPRECATED_3_0
Support key check in EVP interface A new method is added to EVP_PKEY_METH as: int (*check) (EVP_PKEY_CTX *ctx); and to EVP_PKEY_ASN1_METHOD as: int (*pkey_check) (EVP_PKEY_CTX *ctx); This is used to check the validity of a specific key. The order of calls is: EVP_PKEY_check -> pmeth.check -> ameth.pkey_check. Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/4337)
2017-09-04 22:02:59 +08:00
custom_pmeth = EVP_PKEY_meth_new(0xdefaced, 0);
if (!TEST_ptr(custom_pmeth))
return 0;
EVP_PKEY_meth_set_check(custom_pmeth, pkey_custom_check);
Support public key and param check in EVP interface EVP_PKEY_public_check() and EVP_PKEY_param_check() Doc and test cases are added Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/4647)
2017-11-01 00:45:24 +08:00
EVP_PKEY_meth_set_public_check(custom_pmeth, pkey_custom_pub_check);
EVP_PKEY_meth_set_param_check(custom_pmeth, pkey_custom_param_check);
Support key check in EVP interface A new method is added to EVP_PKEY_METH as: int (*check) (EVP_PKEY_CTX *ctx); and to EVP_PKEY_ASN1_METHOD as: int (*pkey_check) (EVP_PKEY_CTX *ctx); This is used to check the validity of a specific key. The order of calls is: EVP_PKEY_check -> pmeth.check -> ameth.pkey_check. Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/4337)
2017-09-04 22:02:59 +08:00
if (!TEST_int_eq(EVP_PKEY_meth_add0(custom_pmeth), 1))
return 0;
EVP: deprecate the EVP_X_meth_ functions. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/11082)
2020-02-13 09:00:57 +08:00
#endif
Support key check in EVP interface A new method is added to EVP_PKEY_METH as: int (*check) (EVP_PKEY_CTX *ctx); and to EVP_PKEY_ASN1_METHOD as: int (*pkey_check) (EVP_PKEY_CTX *ctx); This is used to check the validity of a specific key. The order of calls is: EVP_PKEY_check -> pmeth.check -> ameth.pkey_check. Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/4337)
2017-09-04 22:02:59 +08:00
ADD_ALL_TESTS(test_EVP_PKEY_check, OSSL_NELEM(keycheckdata));
test/evp_extra_test.c: Add OPENSSL_NO_CMAC around CMAC test Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/11833)
2020-05-14 23:15:05 +08:00
#ifndef OPENSSL_NO_CMAC
test/evp_extra_test.c: Add test for CMAC keygen with a NULL engine Verifies that #11671 is fixed Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/11674)
2020-05-12 21:27:32 +08:00
ADD_TEST(test_CMAC_keygen);
test/evp_extra_test.c: Add OPENSSL_NO_CMAC around CMAC test Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/11833)
2020-05-14 23:15:05 +08:00
#endif
Add a test where we reuse the EVP_PKEY_CTX for two HKDF test runs Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7501)
2018-10-26 19:45:27 +08:00
ADD_TEST(test_HKDF);
test HKDF with empty IKM Add an extra EVP test that provides empty input key material. It currently fails, since we lose the information about "key present but zero length" as we deserialize parameters in the provider. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/11920)
2020-05-22 03:53:59 +08:00
ADD_TEST(test_emptyikm_HKDF);
Handle empty param in EVP_PKEY_CTX_add1_hkdf_info Fixes #24130 The regression was introduced in PR #23456. Reviewed-by: Paul Dale <ppzgs1@gmail.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/24141)
2024-04-15 17:13:14 +08:00
ADD_TEST(test_empty_salt_info_HKDF);
Fix a crash in reuse of i2d_X509_PUBKEY If the second PUBKEY is malformed there is use after free. Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/8122)
2019-01-30 23:20:31 +08:00
#ifndef OPENSSL_NO_EC
ADD_TEST(test_X509_PUBKEY_inplace);
test: add test case for X508_PUBKEY_dup() function Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14642)
2021-03-25 10:05:54 +08:00
ADD_TEST(test_X509_PUBKEY_dup);
fixed public range check in ec_GF2m_simple_oct2point Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de> (Merged from https://github.com/openssl/openssl/pull/8607)
2019-03-27 15:38:28 +08:00
ADD_ALL_TESTS(test_invalide_ec_char2_pub_range_decode,
OSSL_NELEM(ec_der_pub_keys));
Fix a crash in reuse of i2d_X509_PUBKEY If the second PUBKEY is malformed there is use after free. Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/8122)
2019-01-30 23:20:31 +08:00
#endif
Fix no-dsa Also fixes no-dh Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/9861)
2019-09-11 17:04:23 +08:00
#ifndef OPENSSL_NO_DSA
test/evp_extra_test.c: adapt for RSA signature tests Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/10557)
2019-12-02 18:26:15 +08:00
ADD_TEST(test_DSA_get_set_params);
Allow EVP_PKEY private key objects to be created without a public component Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/13855)
2021-01-13 12:10:51 +08:00
ADD_TEST(test_DSA_priv_pub);
Add a test for NULL chunks in encrypt/decrypt Issue #8675 describes a problem where calling EVP_DecryptUpdate() with an empty chunk causes the result to be different compared to if you do not use an empty chunk. This adds a test for that case. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10530)
2019-05-31 21:32:55 +08:00
#endif
test/evp_extra_test.c: adapt for RSA signature tests Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/10557)
2019-12-02 18:26:15 +08:00
ADD_TEST(test_RSA_get_set_params);
set MGF1 digest correctly Fixes #19290 update rsa_set_ctx_params() so that the digest function used in the MGF1 construction is set correctly. Add a test for this to evp_extra_test.c based on the code scaro-axway provided in #19290. Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com> Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/19342)
2022-10-04 22:48:43 +08:00
ADD_TEST(test_RSA_OAEP_set_get_params);
rsa: Accept NULL OAEP label for backward compatibility According to the manual page, EVP_PKEY_CTX_set0_rsa_oaep_label() should accept NULL as the label argument, though the function currently rejects it while setting the corresponding octet string parameter with OSSL_PARAM_construct_octet_string, which expects non-NULL input. This adds a workaround to the caller for backward compatibility. Signed-off-by: Daiki Ueno <dueno@redhat.com> Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/22397)
2023-10-16 13:42:12 +08:00
ADD_TEST(test_RSA_OAEP_set_null_label);
Augment RSA provider to generate CRT coefficients on EVP_PKEY_fromdata() It would be helpful to be able to generate RSA's dmp1/dmq1/iqmp values when not provided in the param list to EVP_PKEY_fromdata. Augment the provider in ossl_rsa_fromdata to preform this generation iff: a) At least p q n e and e are provided b) the new parameter OSSL_PARAM_RSA_DERIVE_PQ is set to 1 Fixes #21826 Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/21875)
2023-08-28 20:48:34 +08:00
#ifndef OPENSSL_NO_DEPRECATED_3_0
ADD_TEST(test_RSA_legacy);
#endif
Add a test for NULL chunks in encrypt/decrypt Issue #8675 describes a problem where calling EVP_DecryptUpdate() with an empty chunk causes the result to be different compared to if you do not use an empty chunk. This adds a test for that case. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10530)
2019-05-31 21:32:55 +08:00
#if !defined(OPENSSL_NO_CHACHA) && !defined(OPENSSL_NO_POLY1305)
ADD_TEST(test_decrypt_null_chunks);
Fix no-dsa Also fixes no-dh Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/9861)
2019-09-11 17:04:23 +08:00
#endif
Allow EVP_PKEY private key objects to be created without a public component Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/13855)
2021-01-13 12:10:51 +08:00
#ifndef OPENSSL_NO_DH
ADD_TEST(test_DH_priv_pub);
# ifndef OPENSSL_NO_DEPRECATED_3_0
Test that EVP_PKEY_set1_DH() correctly identifies the DH type Provide a test to check tat when we assign a DH object we know whether we are dealing with PKCS#3 or X9.42 DH keys. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10593)
2019-12-09 20:03:02 +08:00
ADD_TEST(test_EVP_PKEY_set1_DH);
Allow EVP_PKEY private key objects to be created without a public component Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from https://github.com/openssl/openssl/pull/13855)
2021-01-13 12:10:51 +08:00
# endif
Test that EC keys without a public key in them work as expected We create EC keys via both the "fromdata" and legacy key routes to make sure that they can be used without a public key. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13922)
2021-01-21 23:14:15 +08:00
#endif
#ifndef OPENSSL_NO_EC
ADD_TEST(test_EC_priv_pub);
test: evp_extra: EC, read affine coordinates Add a test to read the EC X,Y coordinates. Support legacy keys. Signed-off-by: Jorge Ramirez-Ortiz <jorge@foundries.io> Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/20535)
2023-03-20 19:51:55 +08:00
ADD_TEST(test_evp_get_ec_pub);
Test that EC keys without a public key in them work as expected We create EC keys via both the "fromdata" and legacy key routes to make sure that they can be used without a public key. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13922)
2021-01-21 23:14:15 +08:00
# ifndef OPENSSL_NO_DEPRECATED_3_0
ADD_TEST(test_EC_priv_only_legacy);
test: evp_extra: EC, read affine coordinates Add a test to read the EC X,Y coordinates. Support legacy keys. Signed-off-by: Jorge Ramirez-Ortiz <jorge@foundries.io> Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/20535)
2023-03-20 19:51:55 +08:00
ADD_TEST(test_evp_get_ec_pub_legacy);
Test that EC keys without a public key in them work as expected We create EC keys via both the "fromdata" and legacy key routes to make sure that they can be used without a public key. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13922)
2021-01-21 23:14:15 +08:00
# endif
Fix evp_extra_test with no-dh The new DH test in evp_extra_test.c broke the no-dh build so we add some guards to fix it. Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de> (Merged from https://github.com/openssl/openssl/pull/10644)
2019-12-18 19:00:42 +08:00
#endif
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)
2020-04-15 19:36:19 +08:00
ADD_ALL_TESTS(test_keygen_with_empty_template, 2);
Add a test for fetching EVP_PKEY style algs without a provider Following on from the previous commit, add a test to check that we fail to create an EVP_PKEY_CTX if an algorithm is not available in any provider, *unless* it is an algorithm that has no provider support. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/11826)
2020-05-14 18:33:01 +08:00
ADD_ALL_TESTS(test_pkey_ctx_fail_without_provider, 2);
Test that EVP_PKEY_set1_DH() correctly identifies the DH type Provide a test to check tat when we assign a DH object we know whether we are dealing with PKCS#3 or X9.42 DH keys. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/10593)
2019-12-09 20:03:02 +08:00
test: add test for generation of random data in chunks. THe EVP_RAND wrapper works with the underlying RNG to produce the amount of random data requested even if it is larger than the largest single generation the source allows. This test verified that this works. Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> (Merged from https://github.com/openssl/openssl/pull/11682)
2020-06-11 09:07:13 +08:00
ADD_TEST(test_rand_agglomeration);
Add test for the IV handling of DES based ciphers Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/14882)
2021-04-15 17:53:42 +08:00
ADD_ALL_TESTS(test_evp_iv_aes, 12);
#ifndef OPENSSL_NO_DES
ADD_ALL_TESTS(test_evp_iv_des, 6);
Test the default key length of the Blowfish ciphers Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/18362)
2022-05-20 22:31:33 +08:00
#endif
#ifndef OPENSSL_NO_BF
ADD_ALL_TESTS(test_evp_bf_default_keylen, 4);
Add test for the IV handling of DES based ciphers Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/14882)
2021-04-15 17:53:42 +08:00
#endif
EVP: fix keygen for EVP_PKEY_RSA_PSS This commit attempts to fix the an issue when generating a key of type EVP_PKEY_RSA_PSS. Currently, EVP_PKEY_CTX_set_rsa_keygen_bits will return -1 if the key id is not of type EVP_PKEY_RSA. This commit adds EVP_PKEY_RSA_PSS to also be accepted. The macro EVP_PKEY_CTX_set_rsa_pss_keygen_md si converted into a function and it is now called in legacy_ctrl_to_param. Fixes #12384 Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/13099)
2020-10-09 12:07:43 +08:00
ADD_TEST(test_EVP_rsa_pss_with_keygen_bits);
Fix EVP_PKEY_CTX_get_rsa_pss_saltlen() not returning a value When an integer value was specified, it was not being passed back via the orig_p2 weirdness. Regression test included. Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/17136)
2021-11-25 23:49:26 +08:00
ADD_TEST(test_EVP_rsa_pss_set_saltlen);
RSA decoder should check also sanity of p, q, e, d ... with respect to n This issue has been discovered by osss-fuzzer [1]. The test function decodes RSA key created by fuzzer and calls EVP_PKEY_pairwise_check() which proceeds to ossl_bn_miller_rabin_is_prime() check which takes too long exceeding timeout (45secs). The idea is to fix OSSL_DECODER_from_data() code path so invalid RSA keys will be refused. [1] https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=69134 Test case generated by the fuzzer is added. Reviewed-by: Neil Horman <nhorman@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/25190)
2024-08-15 02:07:29 +08:00
ADD_TEST(test_EVP_rsa_invalid_key);
Add test for EC pubkey export/import There seems to be an issue with i2d_provided() in i2d_evp.c that causes us to fail to construct a valid chain of encoders for the "type-specific" output when it's an EC pubkey. This test is designed to exercise that codepath for a variety of curves. Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14291)
2021-01-26 04:19:16 +08:00
#ifndef OPENSSL_NO_EC
ADD_ALL_TESTS(test_ecpub, OSSL_NELEM(ecpub_nids));
#endif
test: add test for generation of random data in chunks. THe EVP_RAND wrapper works with the underlying RNG to produce the amount of random data requested even if it is larger than the largest single generation the source allows. This test verified that this works. Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> (Merged from https://github.com/openssl/openssl/pull/11682)
2020-06-11 09:07:13 +08:00
Add a test for a names_do_all function Make sure that if we change the namemap part way through calling a names_do_all function it still works. Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14250)
2021-02-20 01:47:21 +08:00
ADD_TEST(test_names_do_all);
Test EVP_CipherInit sequences and resets Various EVP_CipherInit sequences including partial inits and initializations with different "enc" flags caused problems on s390x. Similarly, cipher reinitialization and especially GCM reinitialization with different tag length led to wrong results. Add some unit tests to cover these rather exotic use cases. Signed-off-by: Juergen Christ <jchrist@linux.ibm.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/15521)
2021-06-03 01:33:50 +08:00
ADD_ALL_TESTS(test_evp_init_seq, OSSL_NELEM(evp_init_tests));
ADD_ALL_TESTS(test_evp_reset, OSSL_NELEM(evp_reset_tests));
Add tests for re-using cipher contexts Add test case for re-using a cipher context with the same key, iv and cipher. It detects, if the hardware-specific cipher context is reset correctly, like reported in issue #23175. This test has encrypt and decrypt iterations for cfb128 and ofb128. All iteations use the same key, iv and plaintext. Signed-off-by: Holger Dengler <dengler@linux.ibm.com> Reviewed-by: Shane Lontis <shane.lontis@oracle.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/23201)
2024-01-05 02:25:08 +08:00
ADD_ALL_TESTS(test_evp_reinit_seq, OSSL_NELEM(evp_reinit_tests));
Test EVP_CipherInit sequences and resets Various EVP_CipherInit sequences including partial inits and initializations with different "enc" flags caused problems on s390x. Similarly, cipher reinitialization and especially GCM reinitialization with different tag length led to wrong results. Add some unit tests to cover these rather exotic use cases. Signed-off-by: Juergen Christ <jchrist@linux.ibm.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/15521)
2021-06-03 01:33:50 +08:00
ADD_ALL_TESTS(test_gcm_reinit, OSSL_NELEM(gcm_reinit_tests));
Test EVP Cipher updating the context's IV Ensure that an EVP_CipherUpdate operation updates the context's IV for AES CBC, CFB, OFB, and CTR. An application can get the updated IV via EVP_CIPHER_CTX_iv(). The s390x implementation of the CFB and OFB ciphers in e_aes.c did not update the IV in the context, but only within its s390x specific context data. Signed-off-by: Ingo Franzki <ifranzki@linux.ibm.com> Reviewed-by: Patrick Steuer <patrick.steuer@de.ibm.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16291)
2021-08-11 19:04:52 +08:00
ADD_ALL_TESTS(test_evp_updated_iv, OSSL_NELEM(evp_updated_iv_tests));
Add negative test for iv length change Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/22590)
2023-11-02 01:14:09 +08:00
ADD_ALL_TESTS(test_ivlen_change, OSSL_NELEM(ivlen_change_ciphers));
Add negative test for key length change Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/22590)
2023-11-02 01:39:32 +08:00
if (OSSL_NELEM(keylen_change_ciphers) - 1 > 0)
ADD_ALL_TESTS(test_keylen_change, OSSL_NELEM(keylen_change_ciphers) - 1);
Test EVP_CipherInit sequences and resets Various EVP_CipherInit sequences including partial inits and initializations with different "enc" flags caused problems on s390x. Similarly, cipher reinitialization and especially GCM reinitialization with different tag length led to wrong results. Add some unit tests to cover these rather exotic use cases. Signed-off-by: Juergen Christ <jchrist@linux.ibm.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/15521)
2021-06-03 01:33:50 +08:00
Add a test for custom EVP_PKEY_METHODs Adds a test for using custom EVP_PKEY_METHODs without an ENGINE. As part of this we also test having a custom EVP_PKEY_METHOD that wraps a built-in EVP_PKEY_METHOD. We do this for both legacy and provided keys. Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16118)
2021-07-20 23:18:58 +08:00
#ifndef OPENSSL_NO_DEPRECATED_3_0
ADD_ALL_TESTS(test_custom_pmeth, 12);
Add a test case for EVP_MD_meth_dup() and EVP_CIPHER_meth_dup() Check that EVP_MD_meth_free() and EVP_CIPHER_meth_free() does actually free the data. Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16159)
2021-07-27 17:32:49 +08:00
ADD_TEST(test_evp_md_cipher_meth);
Add a test for a custom digest created via EVP_MD_meth_new() We check that the init and cleanup functions for the custom method are called as expected. Based on an original reproducer by Dmitry Belyavsky from issue #17149. Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com> (Merged from https://github.com/openssl/openssl/pull/17255)
2021-12-30 00:39:11 +08:00
ADD_TEST(test_custom_md_meth);
Test usage of a custom EVP_CIPHER Test that a custom EVP_CIPHER gets used in EVP_CipherInit_ex() calls. Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/19300)
2022-09-29 19:10:30 +08:00
ADD_TEST(test_custom_ciph_meth);
Add tests for ENGINE problems Add some tests which would have caught the issues fixed in the previous 3 commits related to engine handling. Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16846)
2021-10-15 23:30:45 +08:00
# ifndef OPENSSL_NO_DYNAMIC_ENGINE
/* Tests only support the default libctx */
if (testctx == NULL) {
# ifndef OPENSSL_NO_EC
ADD_ALL_TESTS(test_signatures_with_engine, 3);
# else
ADD_ALL_TESTS(test_signatures_with_engine, 2);
# endif
ADD_TEST(test_cipher_with_engine);
}
# endif
Add a test for custom EVP_PKEY_METHODs Adds a test for using custom EVP_PKEY_METHODs without an ENGINE. As part of this we also test having a custom EVP_PKEY_METHOD that wraps a built-in EVP_PKEY_METHOD. We do this for both legacy and provided keys. Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/16118)
2021-07-20 23:18:58 +08:00
#endif
configure: introduce no-ecx to remove ECX related feature This can effectively reduce the binary size for platforms that don't need ECX feature(~100KB). Signed-off-by: Yi Li <yi1.li@intel.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/20781)
2023-04-17 16:20:31 +08:00
#ifndef OPENSSL_NO_ECX
Add a test for creating ECX private keys that are too short We expect attempting to create such short keys to fail Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/17041)
2021-11-15 20:24:05 +08:00
ADD_ALL_TESTS(test_ecx_short_keys, OSSL_NELEM(ecxnids));
Check for private key existence before calling eddsa sign functions Fixes #19524 Reviewed-by: Paul Dale <pauli@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/19525)
2022-10-28 11:47:50 +08:00
ADD_ALL_TESTS(test_ecx_not_private_key, OSSL_NELEM(keys));
#endif
Add test for context duplication failure This checks that the first operation successfully completes even if context duplication fails. But follwing operations get errors as if the context was finlised. Signed-off-by: Simo Sorce <simo@redhat.com> Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com> (Merged from https://github.com/openssl/openssl/pull/20375)
2023-03-07 07:38:00 +08:00
ADD_TEST(test_sign_continuation);
test: add unit test for CVE-2023-5363 Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Hugo Landau <hlandau@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org>
2023-10-06 07:32:20 +08:00
/* Test cases for CVE-2023-5363 */
ADD_TEST(test_aes_gcm_ivlen_change_cve_2023_5363);
#ifndef OPENSSL_NO_RC4
ADD_TEST(test_aes_rc4_keylen_change_cve_2023_5363);
#endif
EVP_DigestUpdate(): Check if ctx->update is set The issue has been discovered by libFuzzer running on provider target. There are currently three distinct reports which are addressed by code change here. https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=69236#c1 https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=69243#c1 https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=69261#c1 the issue has been introduced with openssl 3.0. Reviewed-by: Neil Horman <nhorman@openssl.org> Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/24753)
2024-06-27 22:31:41 +08:00
ADD_TEST(test_invalid_ctx_for_digest);
Add ML-KEM-768 KATs from BoringSSL Add KATs for ML-KEM-768 under CCLA from https://boringssl.googlesource.com/boringssl/ These KATs test key generation, encapsulation, and decapsulation for the ML-KEM-768 algorithm. Relevant notes: - Added functionality to the ML-KEM key management to export/import. These may not be fully implemented yet (see openssl/openssl#25885) - Exposed some more low-level ML-KEM API's to the provider implementation to allow for deterministic encapsulation/key generation - Actually run 'mlkem_internal_test' with `make test` Reviewed-by: Neil Horman <nhorman@openssl.org> Reviewed-by: Tomas Mraz <tomas@openssl.org> (Merged from https://github.com/openssl/openssl/pull/25938)
2024-11-13 21:18:13 +08:00
test: add evp_extra_test case for cipher pipeline API with fake pipeline provider Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/24636)
2024-08-31 20:03:49 +08:00
ADD_TEST(test_evp_cipher_pipeline);
Update the test framework so that the need for test_main is removed. Everything that needed test_main now works using the same infrastructure as tests that used register_tests. This meant: * renaming register_tests to setup_tests and giving it a success/failure return. * renaming the init_test function to setup_test_framework. * renaming the finish_test function to pulldown_test_framework. * adding a user provided global_init function that runs before the test frame work is initialised. It returns a failure indication that stops the stest. * adding helper functions that permit tests to access their command line args. * spliting the BIO initialisation and finalisation out from the test setup and teardown. * hiding some of the now test internal functions. * fix the comments in testutil.h Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/3953)
2017-07-18 09:48:27 +08:00
return 1;
Import evp_test.c from BoringSSL. Unfortunately we already have a file called evp_test.c, so I have called this one evp_extra_test.c Reviewed-by: Emilia Käsper <emilia@openssl.org>
2015-02-09 17:45:35 +08:00
}
Add some tests for the newly added raw private/public key functions We already had soem tests for the older raw private/public key functions so we expand those to call the new versions as well and pass in a libctx. Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/11635)
2020-04-28 00:17:05 +08:00
void cleanup_tests(void)
{
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
OSSL_PROVIDER_unload(nullprov);
OSSL_PROVIDER_unload(deflprov);
Workaround crash in atexit on NonStop platforms We cannot dynamically load the legacy provider into an application that is linked statically to libcrypto as this causes a double loading of libcrypto (one static and one dynamic) and on NonStop this leads to a segfault in atexit(). Fixes #17537 Reviewed-by: Hugo Landau <hlandau@openssl.org> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/19844)
2022-12-06 17:52:47 +08:00
#ifndef OPENSSL_SYS_TANDEM
Add testing for non-default library context into evp_extra_test Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/14478)
2021-03-09 18:42:57 +08:00
OSSL_PROVIDER_unload(lgcyprov);
Workaround crash in atexit on NonStop platforms We cannot dynamically load the legacy provider into an application that is linked statically to libcrypto as this causes a double loading of libcrypto (one static and one dynamic) and on NonStop this leads to a segfault in atexit(). Fixes #17537 Reviewed-by: Hugo Landau <hlandau@openssl.org> Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Paul Dale <pauli@openssl.org> (Merged from https://github.com/openssl/openssl/pull/19844)
2022-12-06 17:52:47 +08:00
#endif
Rename OPENSSL_CTX prefix to OSSL_LIB_CTX Many of the new types introduced by OpenSSL 3.0 have an OSSL_ prefix, e.g., OSSL_CALLBACK, OSSL_PARAM, OSSL_ALGORITHM, OSSL_SERIALIZER. The OPENSSL_CTX type stands out a little by using a different prefix. For consistency reasons, this type is renamed to OSSL_LIB_CTX. Reviewed-by: Paul Dale <paul.dale@oracle.com> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/12621)
2020-10-15 17:55:50 +08:00
OSSL_LIB_CTX_free(testctx);
Add some tests for the newly added raw private/public key functions We already had soem tests for the older raw private/public key functions so we expand those to call the new versions as well and pass in a libctx. Reviewed-by: Shane Lontis <shane.lontis@oracle.com> (Merged from https://github.com/openssl/openssl/pull/11635)
2020-04-28 00:17:05 +08:00
}