yuzu-emu/mbedtls
1
0
Fork 0
mirror of https://github.com/yuzu-emu/mbedtls.git synced 2024-11-30 12:04:16 +01:00
Code Issues Packages Projects Releases Wiki Activity
b8cde4ec03
mbedtls/tests/suites/test_suite_psa_crypto.function

5326 lines
197 KiB
Plaintext
Raw Normal View History

Add PSA crypto module New module psa_crypto.c (MBEDTLS_PSA_CRYPTO_C): Platform Security Architecture compatibility layer on top of libmedcrypto. Implement psa_crypto_init function which sets up a RNG. Add a mbedtls_psa_crypto_free function which deinitializes the library. Define a first batch of error codes.
2018-01-27 23:32:46 +01:00
/* BEGIN_HEADER */
On target testing tests adaptation Updated all psa crypto tests to use the new test format
2018-06-12 16:06:52 +02:00
#include <stdint.h>
Modifications for psa-crypto in order to integrate with SPM Add required includes in tests and psa_crypto.c file in order to be able to compilef for the SPM solution. Some functions needed to be deprecated from psa_crypto.c since they already implemented in the SPM.
2018-07-03 12:16:15 +02:00
Improve and augment export sanity checks Implement sanity checks of exported public keys, using ASN.1 parsing. Rewrite the sanity checks of key pairs using ASN.1 parsing, so as to check more things with simpler code.
2018-08-11 01:22:42 +02:00
#include "mbedtls/asn1.h"
Test that creating RSA keys larger than the maximum fails Test keypair import, public key import and key generation.
2018-06-28 00:16:11 +02:00
#include "mbedtls/asn1write.h"
Improve and augment export sanity checks Implement sanity checks of exported public keys, using ASN.1 parsing. Rewrite the sanity checks of key pairs using ASN.1 parsing, so as to check more things with simpler code.
2018-08-11 01:22:42 +02:00
#include "mbedtls/oid.h"
Add tests to generate more random than MBEDTLS_CTR_DRBG_MAX_REQUEST Add tests that call psa_generate_random() (possibly via psa_generate_key()) with a size that's larger than MBEDTLS_CTR_DRBG_MAX_REQUEST. This causes psa_generate_random() to fail because it calls mbedtls_ctr_drbg_random() without taking the maximum request size of CTR_DRBG into account. Non-regression test for #206
2019-08-07 12:08:04 +02:00
/* For MBEDTLS_CTR_DRBG_MAX_REQUEST, knowing that psa_generate_random()
* uses mbedtls_ctr_drbg internally. */
#include "mbedtls/ctr_drbg.h"
Rename psa_helpers.function to psa_crypto_helpers.h This file isn't like the other .function files: it isn't concatenated by a separate preprocessing script, but included via C preprocessing. Rename this file to .h. This isn't a normal C header, because it defines auxiliary functions. But the functions aren't big and we only have one compilation unit per executable, so this is good enough for what we're doing.
2019-06-20 12:40:56 +02:00
#include "psa_crypto_helpers.h"
On target testing tests adaptation Updated all psa crypto tests to use the new test format
2018-06-12 16:06:52 +02:00
Limit keys to 65528 bits 65528 bits is more than any reasonable key until we start supporting post-quantum cryptography. This limit is chosen to allow bit-sizes to be stored in 16 bits, with 65535 left to indicate an invalid value. It's a whole number of bytes, which facilitates some calculations, in particular allowing a key of exactly PSA_CRYPTO_MAX_STORAGE_SIZE to be created but not one bit more. As a resource usage limit, this is arguably too large, but that's out of scope of the current commit. Test that key import, generation and derivation reject overly large sizes.
2019-07-30 17:26:54 +02:00
/* Tests that require more than 128kB of RAM plus change have this symbol
* as a dependency. Currently we always define this symbol, so the tests
* are always executed. In the future we should make this conditional
* so that tests that require a lot of memory are skipped on constrained
* platforms. */
Avoid a lowercase letter in a macro name
2019-08-07 11:01:30 +02:00
#define HAVE_RAM_AVAILABLE_128K
Limit keys to 65528 bits 65528 bits is more than any reasonable key until we start supporting post-quantum cryptography. This limit is chosen to allow bit-sizes to be stored in 16 bits, with 65535 left to indicate an invalid value. It's a whole number of bytes, which facilitates some calculations, in particular allowing a key of exactly PSA_CRYPTO_MAX_STORAGE_SIZE to be created but not one bit more. As a resource usage limit, this is arguably too large, but that's out of scope of the current commit. Test that key import, generation and derivation reject overly large sizes.
2019-07-30 17:26:54 +02:00
Add tests to generate more random than MBEDTLS_CTR_DRBG_MAX_REQUEST Add tests that call psa_generate_random() (possibly via psa_generate_key()) with a size that's larger than MBEDTLS_CTR_DRBG_MAX_REQUEST. This causes psa_generate_random() to fail because it calls mbedtls_ctr_drbg_random() without taking the maximum request size of CTR_DRBG into account. Non-regression test for #206
2019-08-07 12:08:04 +02:00
#include "psa/crypto.h"
psa_export_key: Always set a valid data length Make psa_export_key() always set a valid data_length when exporting, even when there are errors. This makes the API easier to use for buggy programs (like our test code). Our test code previously used exported_length uninitialized when checking to see that the buffer returned was all zero in import_export() in the case where an error was returned from psa_export_key(). Initialize exported_length to an invalid length, and check that it gets set properly by psa_export_key(), to avoid this using export_length uninitialized. Note that the mem_is_zero() check is still valid when psa_export_key() returns an error, e.g. where exported_length is 0, as we want to check that nothing was written to the buffer on error. Out test code also previous passed NULL for the data_length parameter of psa_export_key() when it expected a failure (in key_policy_fail()). However, data_length is not allowed to be NULL, especially now that we write to data_length from psa_export_key() even when there are errors. Update the test code to not pass in a NULL data_length.
2018-06-27 18:20:43 +02:00
/** An invalid export length that will never be set by psa_export_key(). */
static const size_t INVALID_EXPORT_LENGTH = ~0U;
Smoke-test operation contexts after setup+abort After a successful setup followed by abort, or after a failed setup from an inactive state, a context must be usable. Test this for hash, MAC and cipher contexts.
2019-02-25 17:42:03 +01:00
/* A hash algorithm that is known to be supported.
*
* This is used in some smoke tests.
*/
#if defined(MBEDTLS_MD2_C)
#define KNOWN_SUPPORTED_HASH_ALG PSA_ALG_MD2
#elif defined(MBEDTLS_MD4_C)
#define KNOWN_SUPPORTED_HASH_ALG PSA_ALG_MD4
#elif defined(MBEDTLS_MD5_C)
#define KNOWN_SUPPORTED_HASH_ALG PSA_ALG_MD5
/* MBEDTLS_RIPEMD160_C omitted. This is necessary for the sake of
* exercise_signature_key() because Mbed TLS doesn't support RIPEMD160
* in RSA PKCS#1v1.5 signatures. A RIPEMD160-only configuration would be
* implausible anyway. */
#elif defined(MBEDTLS_SHA1_C)
#define KNOWN_SUPPORTED_HASH_ALG PSA_ALG_SHA_1
#elif defined(MBEDTLS_SHA256_C)
#define KNOWN_SUPPORTED_HASH_ALG PSA_ALG_SHA_256
#elif defined(MBEDTLS_SHA512_C)
#define KNOWN_SUPPORTED_HASH_ALG PSA_ALG_SHA_384
#elif defined(MBEDTLS_SHA3_C)
#define KNOWN_SUPPORTED_HASH_ALG PSA_ALG_SHA3_256
#else
#undef KNOWN_SUPPORTED_HASH_ALG
#endif
/* A block cipher that is known to be supported.
*
* For simplicity's sake, stick to block ciphers with 16-byte blocks.
*/
#if defined(MBEDTLS_AES_C)
#define KNOWN_SUPPORTED_BLOCK_CIPHER PSA_KEY_TYPE_AES
#elif defined(MBEDTLS_ARIA_C)
#define KNOWN_SUPPORTED_BLOCK_CIPHER PSA_KEY_TYPE_ARIA
#elif defined(MBEDTLS_CAMELLIA_C)
#define KNOWN_SUPPORTED_BLOCK_CIPHER PSA_KEY_TYPE_CAMELLIA
#undef KNOWN_SUPPORTED_BLOCK_CIPHER
#endif
/* A MAC mode that is known to be supported.
*
* It must either be HMAC with #KNOWN_SUPPORTED_HASH_ALG or
* a block cipher-based MAC with #KNOWN_SUPPORTED_BLOCK_CIPHER.
*
* This is used in some smoke tests.
*/
#if defined(KNOWN_SUPPORTED_HASH_ALG)
#define KNOWN_SUPPORTED_MAC_ALG ( PSA_ALG_HMAC( KNOWN_SUPPORTED_HASH_ALG ) )
#define KNOWN_SUPPORTED_MAC_KEY_TYPE PSA_KEY_TYPE_HMAC
#elif defined(KNOWN_SUPPORTED_BLOCK_CIPHER) && defined(MBEDTLS_CMAC_C)
#define KNOWN_SUPPORTED_MAC_ALG PSA_ALG_CMAC
#define KNOWN_SUPPORTED_MAC_KEY_TYPE KNOWN_SUPPORTED_BLOCK_CIPHER
#else
#undef KNOWN_SUPPORTED_MAC_ALG
#undef KNOWN_SUPPORTED_MAC_KEY_TYPE
#endif
/* A cipher algorithm and key type that are known to be supported.
*
* This is used in some smoke tests.
*/
#if defined(KNOWN_SUPPORTED_BLOCK_CIPHER) && defined(MBEDTLS_CIPHER_MODE_CTR)
#define KNOWN_SUPPORTED_BLOCK_CIPHER_ALG PSA_ALG_CTR
#elif defined(KNOWN_SUPPORTED_BLOCK_CIPHER) && defined(MBEDTLS_CIPHER_MODE_CBC)
#define KNOWN_SUPPORTED_BLOCK_CIPHER_ALG PSA_ALG_CBC_NO_PADDING
#elif defined(KNOWN_SUPPORTED_BLOCK_CIPHER) && defined(MBEDTLS_CIPHER_MODE_CFB)
#define KNOWN_SUPPORTED_BLOCK_CIPHER_ALG PSA_ALG_CFB
#elif defined(KNOWN_SUPPORTED_BLOCK_CIPHER) && defined(MBEDTLS_CIPHER_MODE_OFB)
#define KNOWN_SUPPORTED_BLOCK_CIPHER_ALG PSA_ALG_OFB
#else
#undef KNOWN_SUPPORTED_BLOCK_CIPHER_ALG
#endif
#if defined(KNOWN_SUPPORTED_BLOCK_CIPHER_ALG)
#define KNOWN_SUPPORTED_CIPHER_ALG KNOWN_SUPPORTED_BLOCK_CIPHER_ALG
#define KNOWN_SUPPORTED_CIPHER_KEY_TYPE KNOWN_SUPPORTED_BLOCK_CIPHER
#elif defined(MBEDTLS_RC4_C)
#define KNOWN_SUPPORTED_CIPHER_ALG PSA_ALG_RC4
#define KNOWN_SUPPORTED_CIPHER_KEY_TYPE PSA_KEY_TYPE_RC4
#else
#undef KNOWN_SUPPORTED_CIPHER_ALG
#undef KNOWN_SUPPORTED_CIPHER_KEY_TYPE
#endif
Add tests for mac_sign
2018-08-14 15:17:54 +02:00
/** Test if a buffer contains a constant byte value.
*
* `mem_is_char(buffer, c, size)` is true after `memset(buffer, c, size)`.
psa_export_key: zero out potential garbage in the output buffer In psa_export_key, ensure that each byte of the output buffer either contains its original value, is zero, or is part of the actual output. Specifically, don't risk having partial output on error, and don't leave extra data at the end of the buffer when exporting an asymmetric key. Test that exporting to a previously zeroed buffer leaves the buffer zeroed outside the actual output if any.
2018-06-20 00:11:45 +02:00
*
* \param buffer Pointer to the beginning of the buffer.
Add tests for mac_sign
2018-08-14 15:17:54 +02:00
* \param c Expected value of every byte.
psa_export_key: zero out potential garbage in the output buffer In psa_export_key, ensure that each byte of the output buffer either contains its original value, is zero, or is part of the actual output. Specifically, don't risk having partial output on error, and don't leave extra data at the end of the buffer when exporting an asymmetric key. Test that exporting to a previously zeroed buffer leaves the buffer zeroed outside the actual output if any.
2018-06-20 00:11:45 +02:00
* \param size Size of the buffer in bytes.
*
Simplify mem_is_nonzero to mem_is_zero This also fixes a bug that the value that mem_is_nonzero tried to return could overflow int.
2018-06-21 09:21:51 +02:00
* \return 1 if the buffer is all-bits-zero.
* \return 0 if there is at least one nonzero byte.
psa_export_key: zero out potential garbage in the output buffer In psa_export_key, ensure that each byte of the output buffer either contains its original value, is zero, or is part of the actual output. Specifically, don't risk having partial output on error, and don't leave extra data at the end of the buffer when exporting an asymmetric key. Test that exporting to a previously zeroed buffer leaves the buffer zeroed outside the actual output if any.
2018-06-20 00:11:45 +02:00
*/
Add tests for mac_sign
2018-08-14 15:17:54 +02:00
static int mem_is_char( void *buffer, unsigned char c, size_t size )
psa_export_key: zero out potential garbage in the output buffer In psa_export_key, ensure that each byte of the output buffer either contains its original value, is zero, or is part of the actual output. Specifically, don't risk having partial output on error, and don't leave extra data at the end of the buffer when exporting an asymmetric key. Test that exporting to a previously zeroed buffer leaves the buffer zeroed outside the actual output if any.
2018-06-20 00:11:45 +02:00
{
size_t i;
for( i = 0; i < size; i++ )
{
Add tests for mac_sign
2018-08-14 15:17:54 +02:00
if( ( (unsigned char *) buffer )[i] != c )
Simplify mem_is_nonzero to mem_is_zero This also fixes a bug that the value that mem_is_nonzero tried to return could overflow int.
2018-06-21 09:21:51 +02:00
return( 0 );
psa_export_key: zero out potential garbage in the output buffer In psa_export_key, ensure that each byte of the output buffer either contains its original value, is zero, or is part of the actual output. Specifically, don't risk having partial output on error, and don't leave extra data at the end of the buffer when exporting an asymmetric key. Test that exporting to a previously zeroed buffer leaves the buffer zeroed outside the actual output if any.
2018-06-20 00:11:45 +02:00
}
Simplify mem_is_nonzero to mem_is_zero This also fixes a bug that the value that mem_is_nonzero tried to return could overflow int.
2018-06-21 09:21:51 +02:00
return( 1 );
psa_export_key: zero out potential garbage in the output buffer In psa_export_key, ensure that each byte of the output buffer either contains its original value, is zero, or is part of the actual output. Specifically, don't risk having partial output on error, and don't leave extra data at the end of the buffer when exporting an asymmetric key. Test that exporting to a previously zeroed buffer leaves the buffer zeroed outside the actual output if any.
2018-06-20 00:11:45 +02:00
}
generate_key tests: exercise the key After generating a key, perform a smoke test: run one operation with it and check that the operation has the expected status.
2018-06-20 18:16:48 +02:00
Test that creating RSA keys larger than the maximum fails Test keypair import, public key import and key generation.
2018-06-28 00:16:11 +02:00
/* Write the ASN.1 INTEGER with the value 2^(bits-1)+x backwards from *p. */
static int asn1_write_10x( unsigned char **p,
unsigned char *start,
size_t bits,
unsigned char x )
{
int ret;
int len = bits / 8 + 1;
Fix argument validation in asn1_write_10x 1 << bits doesn't work when bits is too large. Found by ASan.
2018-06-28 19:04:07 +02:00
if( bits == 0 )
return( MBEDTLS_ERR_ASN1_INVALID_DATA );
if( bits <= 8 && x >= 1 << ( bits - 1 ) )
Test that creating RSA keys larger than the maximum fails Test keypair import, public key import and key generation.
2018-06-28 00:16:11 +02:00
return( MBEDTLS_ERR_ASN1_INVALID_DATA );
Replace ssize_t by ptrdiff_t ssize_t is a POSIX thing, not standard C
2018-07-17 16:36:59 +02:00
if( *p < start || *p - start < (ptrdiff_t) len )
Test that creating RSA keys larger than the maximum fails Test keypair import, public key import and key generation.
2018-06-28 00:16:11 +02:00
return( MBEDTLS_ERR_ASN1_BUF_TOO_SMALL );
*p -= len;
( *p )[len-1] = x;
if( bits % 8 == 0 )
( *p )[1] |= 1;
else
( *p )[0] |= 1 << ( bits % 8 );
MBEDTLS_ASN1_CHK_ADD( len, mbedtls_asn1_write_len( p, start, len ) );
MBEDTLS_ASN1_CHK_ADD( len, mbedtls_asn1_write_tag( p, start,
MBEDTLS_ASN1_INTEGER ) );
return( len );
}
static int construct_fake_rsa_key( unsigned char *buffer,
size_t buffer_size,
unsigned char **p,
size_t bits,
int keypair )
{
size_t half_bits = ( bits + 1 ) / 2;
int ret;
int len = 0;
/* Construct something that looks like a DER encoding of
* as defined by PKCS#1 v2.2 (RFC 8017) section A.1.2:
* RSAPrivateKey ::= SEQUENCE {
* version Version,
* modulus INTEGER, -- n
* publicExponent INTEGER, -- e
* privateExponent INTEGER, -- d
* prime1 INTEGER, -- p
* prime2 INTEGER, -- q
* exponent1 INTEGER, -- d mod (p-1)
* exponent2 INTEGER, -- d mod (q-1)
* coefficient INTEGER, -- (inverse of q) mod p
* otherPrimeInfos OtherPrimeInfos OPTIONAL
* }
* Or, for a public key, the same structure with only
* version, modulus and publicExponent.
*/
*p = buffer + buffer_size;
if( keypair )
{
MBEDTLS_ASN1_CHK_ADD( len, /* pq */
asn1_write_10x( p, buffer, half_bits, 1 ) );
MBEDTLS_ASN1_CHK_ADD( len, /* dq */
asn1_write_10x( p, buffer, half_bits, 1 ) );
MBEDTLS_ASN1_CHK_ADD( len, /* dp */
asn1_write_10x( p, buffer, half_bits, 1 ) );
MBEDTLS_ASN1_CHK_ADD( len, /* q */
asn1_write_10x( p, buffer, half_bits, 1 ) );
MBEDTLS_ASN1_CHK_ADD( len, /* p != q to pass mbedtls sanity checks */
asn1_write_10x( p, buffer, half_bits, 3 ) );
MBEDTLS_ASN1_CHK_ADD( len, /* d */
asn1_write_10x( p, buffer, bits, 1 ) );
}
MBEDTLS_ASN1_CHK_ADD( len, /* e = 65537 */
asn1_write_10x( p, buffer, 17, 1 ) );
MBEDTLS_ASN1_CHK_ADD( len, /* n */
asn1_write_10x( p, buffer, bits, 1 ) );
if( keypair )
MBEDTLS_ASN1_CHK_ADD( len, /* version = 0 */
mbedtls_asn1_write_int( p, buffer, 0 ) );
MBEDTLS_ASN1_CHK_ADD( len, mbedtls_asn1_write_len( p, buffer, len ) );
{
const unsigned char tag =
MBEDTLS_ASN1_CONSTRUCTED | MBEDTLS_ASN1_SEQUENCE;
MBEDTLS_ASN1_CHK_ADD( len, mbedtls_asn1_write_tag( p, buffer, tag ) );
}
return( len );
}
Smoke-test operation contexts after setup+abort After a successful setup followed by abort, or after a failed setup from an inactive state, a context must be usable. Test this for hash, MAC and cipher contexts.
2019-02-25 17:42:03 +01:00
int exercise_mac_setup( psa_key_type_t key_type,
const unsigned char *key_bytes,
size_t key_length,
psa_algorithm_t alg,
psa_mac_operation_t *operation,
psa_status_t *status )
{
psa_key_handle_t handle = 0;
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
Smoke-test operation contexts after setup+abort After a successful setup followed by abort, or after a failed setup from an inactive state, a context must be usable. Test this for hash, MAC and cipher contexts.
2019-02-25 17:42:03 +01:00
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_set_key_usage_flags( &attributes, PSA_KEY_USAGE_SIGN );
psa_set_key_algorithm( &attributes, alg );
psa_set_key_type( &attributes, key_type );
Split long lines after psa_import_key refactoring
2019-05-15 20:22:09 +02:00
PSA_ASSERT( psa_import_key( &attributes, key_bytes, key_length,
&handle ) );
Smoke-test operation contexts after setup+abort After a successful setup followed by abort, or after a failed setup from an inactive state, a context must be usable. Test this for hash, MAC and cipher contexts.
2019-02-25 17:42:03 +01:00
*status = psa_mac_sign_setup( operation, handle, alg );
Test abort after failed setup Commit "Smoke-test operation contexts after setup+abort" replaced {failed-setup; abort} sequences by {failed-setup; successful-setup}. We want to test that, but we also want to test {failed-setup; abort}. So test {failed-setup; abort; failed-setup; successful-setup}.
2019-02-25 22:11:18 +01:00
/* Whether setup succeeded or failed, abort must succeed. */
PSA_ASSERT( psa_mac_abort( operation ) );
/* If setup failed, reproduce the failure, so that the caller can
* test the resulting state of the operation object. */
if( *status != PSA_SUCCESS )
Smoke-test operation contexts after setup+abort After a successful setup followed by abort, or after a failed setup from an inactive state, a context must be usable. Test this for hash, MAC and cipher contexts.
2019-02-25 17:42:03 +01:00
{
Test abort after failed setup Commit "Smoke-test operation contexts after setup+abort" replaced {failed-setup; abort} sequences by {failed-setup; successful-setup}. We want to test that, but we also want to test {failed-setup; abort}. So test {failed-setup; abort; failed-setup; successful-setup}.
2019-02-25 22:11:18 +01:00
TEST_EQUAL( psa_mac_sign_setup( operation, handle, alg ),
*status );
Smoke-test operation contexts after setup+abort After a successful setup followed by abort, or after a failed setup from an inactive state, a context must be usable. Test this for hash, MAC and cipher contexts.
2019-02-25 17:42:03 +01:00
}
psa_destroy_key( handle );
return( 1 );
exit:
psa_destroy_key( handle );
return( 0 );
}
int exercise_cipher_setup( psa_key_type_t key_type,
const unsigned char *key_bytes,
size_t key_length,
psa_algorithm_t alg,
psa_cipher_operation_t *operation,
psa_status_t *status )
{
psa_key_handle_t handle = 0;
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
Smoke-test operation contexts after setup+abort After a successful setup followed by abort, or after a failed setup from an inactive state, a context must be usable. Test this for hash, MAC and cipher contexts.
2019-02-25 17:42:03 +01:00
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_set_key_usage_flags( &attributes, PSA_KEY_USAGE_ENCRYPT );
psa_set_key_algorithm( &attributes, alg );
psa_set_key_type( &attributes, key_type );
Split long lines after psa_import_key refactoring
2019-05-15 20:22:09 +02:00
PSA_ASSERT( psa_import_key( &attributes, key_bytes, key_length,
&handle ) );
Smoke-test operation contexts after setup+abort After a successful setup followed by abort, or after a failed setup from an inactive state, a context must be usable. Test this for hash, MAC and cipher contexts.
2019-02-25 17:42:03 +01:00
*status = psa_cipher_encrypt_setup( operation, handle, alg );
Test abort after failed setup Commit "Smoke-test operation contexts after setup+abort" replaced {failed-setup; abort} sequences by {failed-setup; successful-setup}. We want to test that, but we also want to test {failed-setup; abort}. So test {failed-setup; abort; failed-setup; successful-setup}.
2019-02-25 22:11:18 +01:00
/* Whether setup succeeded or failed, abort must succeed. */
PSA_ASSERT( psa_cipher_abort( operation ) );
/* If setup failed, reproduce the failure, so that the caller can
* test the resulting state of the operation object. */
if( *status != PSA_SUCCESS )
Smoke-test operation contexts after setup+abort After a successful setup followed by abort, or after a failed setup from an inactive state, a context must be usable. Test this for hash, MAC and cipher contexts.
2019-02-25 17:42:03 +01:00
{
Test abort after failed setup Commit "Smoke-test operation contexts after setup+abort" replaced {failed-setup; abort} sequences by {failed-setup; successful-setup}. We want to test that, but we also want to test {failed-setup; abort}. So test {failed-setup; abort; failed-setup; successful-setup}.
2019-02-25 22:11:18 +01:00
TEST_EQUAL( psa_cipher_encrypt_setup( operation, handle, alg ),
*status );
Smoke-test operation contexts after setup+abort After a successful setup followed by abort, or after a failed setup from an inactive state, a context must be usable. Test this for hash, MAC and cipher contexts.
2019-02-25 17:42:03 +01:00
}
psa_destroy_key( handle );
return( 1 );
exit:
psa_destroy_key( handle );
return( 0 );
}
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
static int exercise_mac_key( psa_key_handle_t handle,
generate_key tests: exercise the key After generating a key, perform a smoke test: run one operation with it and check that the operation has the expected status.
2018-06-20 18:16:48 +02:00
psa_key_usage_t usage,
psa_algorithm_t alg )
{
psa: Add initializers for MAC operation objects Add new initializers for MAC operation objects and use them in our tests and library code. Prefer using the macro initializers due to their straightforwardness.
2019-01-04 12:48:03 +01:00
psa_mac_operation_t operation = PSA_MAC_OPERATION_INIT;
generate_key tests: exercise the key After generating a key, perform a smoke test: run one operation with it and check that the operation has the expected status.
2018-06-20 18:16:48 +02:00
const unsigned char input[] = "foo";
Use PSA_xxx_MAX_SIZE for hash/MAC/signature size in tests In tests that had a hard-coded buffer size, use PSA_MAC_MAX_SIZE or PSA_ASYMMETRIC_SIGNATURE_MAX_SIZE as appropriate. Test that PSA_xxx_MAX_SIZE is larger than the size used in tests that expect a specific output.
2018-06-28 00:07:19 +02:00
unsigned char mac[PSA_MAC_MAX_SIZE] = {0};
generate_key tests: exercise the key After generating a key, perform a smoke test: run one operation with it and check that the operation has the expected status.
2018-06-20 18:16:48 +02:00
size_t mac_length = sizeof( mac );
if( usage & PSA_KEY_USAGE_SIGN )
{
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_mac_sign_setup( &operation,
handle, alg ) );
PSA_ASSERT( psa_mac_update( &operation,
input, sizeof( input ) ) );
PSA_ASSERT( psa_mac_sign_finish( &operation,
mac, sizeof( mac ),
&mac_length ) );
generate_key tests: exercise the key After generating a key, perform a smoke test: run one operation with it and check that the operation has the expected status.
2018-06-20 18:16:48 +02:00
}
if( usage & PSA_KEY_USAGE_VERIFY )
{
psa_status_t verify_status =
( usage & PSA_KEY_USAGE_SIGN ?
PSA_SUCCESS :
PSA_ERROR_INVALID_SIGNATURE );
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_mac_verify_setup( &operation,
handle, alg ) );
PSA_ASSERT( psa_mac_update( &operation,
input, sizeof( input ) ) );
Rewrap some lines after the macro changes Change the way some lines are wrapped to cut at a more logical place. This commit mainly rewrites multi-line calls to TEST_EQUAL, and also a few calls to PSA_ASSERT.
2018-12-18 00:33:25 +01:00
TEST_EQUAL( psa_mac_verify_finish( &operation, mac, mac_length ),
verify_status );
generate_key tests: exercise the key After generating a key, perform a smoke test: run one operation with it and check that the operation has the expected status.
2018-06-20 18:16:48 +02:00
}
return( 1 );
exit:
psa_mac_abort( &operation );
return( 0 );
}
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
static int exercise_cipher_key( psa_key_handle_t handle,
generate_key tests: exercise the key After generating a key, perform a smoke test: run one operation with it and check that the operation has the expected status.
2018-06-20 18:16:48 +02:00
psa_key_usage_t usage,
psa_algorithm_t alg )
{
psa: Add initializers for cipher operation objects Add new initializers for cipher operation objects and use them in our tests and library code. Prefer using the macro initializers due to their straightforwardness.
2019-01-04 12:48:27 +01:00
psa_cipher_operation_t operation = PSA_CIPHER_OPERATION_INIT;
generate_key tests: exercise the key After generating a key, perform a smoke test: run one operation with it and check that the operation has the expected status.
2018-06-20 18:16:48 +02:00
unsigned char iv[16] = {0};
size_t iv_length = sizeof( iv );
const unsigned char plaintext[16] = "Hello, world...";
unsigned char ciphertext[32] = "(wabblewebblewibblewobblewubble)";
size_t ciphertext_length = sizeof( ciphertext );
unsigned char decrypted[sizeof( ciphertext )];
size_t part_length;
if( usage & PSA_KEY_USAGE_ENCRYPT )
{
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_cipher_encrypt_setup( &operation,
handle, alg ) );
PSA_ASSERT( psa_cipher_generate_iv( &operation,
iv, sizeof( iv ),
&iv_length ) );
PSA_ASSERT( psa_cipher_update( &operation,
plaintext, sizeof( plaintext ),
ciphertext, sizeof( ciphertext ),
&ciphertext_length ) );
PSA_ASSERT( psa_cipher_finish( &operation,
ciphertext + ciphertext_length,
sizeof( ciphertext ) - ciphertext_length,
&part_length ) );
generate_key tests: exercise the key After generating a key, perform a smoke test: run one operation with it and check that the operation has the expected status.
2018-06-20 18:16:48 +02:00
ciphertext_length += part_length;
}
if( usage & PSA_KEY_USAGE_DECRYPT )
{
psa_status_t status;
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
int maybe_invalid_padding = 0;
generate_key tests: exercise the key After generating a key, perform a smoke test: run one operation with it and check that the operation has the expected status.
2018-06-20 18:16:48 +02:00
if( ! ( usage & PSA_KEY_USAGE_ENCRYPT ) )
{
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
PSA_ASSERT( psa_get_key_attributes( handle, &attributes ) );
/* This should be PSA_CIPHER_GET_IV_SIZE but the API doesn't
* have this macro yet. */
iv_length = PSA_BLOCK_CIPHER_BLOCK_SIZE(
psa_get_key_type( &attributes ) );
maybe_invalid_padding = ! PSA_ALG_IS_STREAM_CIPHER( alg );
generate_key tests: exercise the key After generating a key, perform a smoke test: run one operation with it and check that the operation has the expected status.
2018-06-20 18:16:48 +02:00
}
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_cipher_decrypt_setup( &operation,
handle, alg ) );
PSA_ASSERT( psa_cipher_set_iv( &operation,
iv, iv_length ) );
PSA_ASSERT( psa_cipher_update( &operation,
ciphertext, ciphertext_length,
decrypted, sizeof( decrypted ),
&part_length ) );
generate_key tests: exercise the key After generating a key, perform a smoke test: run one operation with it and check that the operation has the expected status.
2018-06-20 18:16:48 +02:00
status = psa_cipher_finish( &operation,
decrypted + part_length,
sizeof( decrypted ) - part_length,
&part_length );
/* For a stream cipher, all inputs are valid. For a block cipher,
* if the input is some aribtrary data rather than an actual
ciphertext, a padding error is likely. */
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
if( maybe_invalid_padding )
generate_key tests: exercise the key After generating a key, perform a smoke test: run one operation with it and check that the operation has the expected status.
2018-06-20 18:16:48 +02:00
TEST_ASSERT( status == PSA_SUCCESS ||
status == PSA_ERROR_INVALID_PADDING );
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
else
PSA_ASSERT( status );
generate_key tests: exercise the key After generating a key, perform a smoke test: run one operation with it and check that the operation has the expected status.
2018-06-20 18:16:48 +02:00
}
return( 1 );
exit:
psa_cipher_abort( &operation );
return( 0 );
}
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
static int exercise_aead_key( psa_key_handle_t handle,
generate_key tests: exercise the key After generating a key, perform a smoke test: run one operation with it and check that the operation has the expected status.
2018-06-20 18:16:48 +02:00
psa_key_usage_t usage,
psa_algorithm_t alg )
{
unsigned char nonce[16] = {0};
size_t nonce_length = sizeof( nonce );
unsigned char plaintext[16] = "Hello, world...";
unsigned char ciphertext[48] = "(wabblewebblewibblewobblewubble)";
size_t ciphertext_length = sizeof( ciphertext );
size_t plaintext_length = sizeof( ciphertext );
if( usage & PSA_KEY_USAGE_ENCRYPT )
{
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_aead_encrypt( handle, alg,
nonce, nonce_length,
NULL, 0,
plaintext, sizeof( plaintext ),
ciphertext, sizeof( ciphertext ),
&ciphertext_length ) );
generate_key tests: exercise the key After generating a key, perform a smoke test: run one operation with it and check that the operation has the expected status.
2018-06-20 18:16:48 +02:00
}
if( usage & PSA_KEY_USAGE_DECRYPT )
{
psa_status_t verify_status =
( usage & PSA_KEY_USAGE_ENCRYPT ?
PSA_SUCCESS :
PSA_ERROR_INVALID_SIGNATURE );
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( psa_aead_decrypt( handle, alg,
nonce, nonce_length,
NULL, 0,
ciphertext, ciphertext_length,
plaintext, sizeof( plaintext ),
Rewrap some lines after the macro changes Change the way some lines are wrapped to cut at a more logical place. This commit mainly rewrites multi-line calls to TEST_EQUAL, and also a few calls to PSA_ASSERT.
2018-12-18 00:33:25 +01:00
&plaintext_length ),
verify_status );
generate_key tests: exercise the key After generating a key, perform a smoke test: run one operation with it and check that the operation has the expected status.
2018-06-20 18:16:48 +02:00
}
return( 1 );
exit:
return( 0 );
}
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
static int exercise_signature_key( psa_key_handle_t handle,
generate_key tests: exercise the key After generating a key, perform a smoke test: run one operation with it and check that the operation has the expected status.
2018-06-20 18:16:48 +02:00
psa_key_usage_t usage,
psa_algorithm_t alg )
{
Change a generate_key test to exercise with PSS This required tweaking exercise_signature_key to use a payload size for the signature based on the algorithm, since our implementation of PSS requires that the input size matches the hash size. This would also be the case for PKCS#1 v1.5 with a specified hash.
2018-06-30 00:20:25 +02:00
unsigned char payload[PSA_HASH_MAX_SIZE] = {1};
size_t payload_length = 16;
Use PSA_xxx_MAX_SIZE for hash/MAC/signature size in tests In tests that had a hard-coded buffer size, use PSA_MAC_MAX_SIZE or PSA_ASYMMETRIC_SIGNATURE_MAX_SIZE as appropriate. Test that PSA_xxx_MAX_SIZE is larger than the size used in tests that expect a specific output.
2018-06-28 00:07:19 +02:00
unsigned char signature[PSA_ASYMMETRIC_SIGNATURE_MAX_SIZE] = {0};
generate_key tests: exercise the key After generating a key, perform a smoke test: run one operation with it and check that the operation has the expected status.
2018-06-20 18:16:48 +02:00
size_t signature_length = sizeof( signature );
Test psa_copy_key Split the testing into tests that exercise policies in test_suite_psa_crypto and tests that exercise slot content (slot states, key material) in test_suite_psa_crypto_slot_management. Test various cases of source and target policies with and without wildcards. Missing: testing of the policy constraint on psa_copy_key itself. Test several key types (raw data, AES, RSA). Test with the source or target being persistent. Add failure tests (incompatible policies, source slot empty, target slot occupied).
2019-01-28 13:03:09 +01:00
psa_algorithm_t hash_alg = PSA_ALG_SIGN_GET_HASH( alg );
/* If the policy allows signing with any hash, just pick one. */
if( PSA_ALG_IS_HASH_AND_SIGN( alg ) && hash_alg == PSA_ALG_ANY_HASH )
{
Smoke-test operation contexts after setup+abort After a successful setup followed by abort, or after a failed setup from an inactive state, a context must be usable. Test this for hash, MAC and cipher contexts.
2019-02-25 17:42:03 +01:00
#if defined(KNOWN_SUPPORTED_HASH_ALG)
hash_alg = KNOWN_SUPPORTED_HASH_ALG;
alg ^= PSA_ALG_ANY_HASH ^ hash_alg;
Test psa_copy_key Split the testing into tests that exercise policies in test_suite_psa_crypto and tests that exercise slot content (slot states, key material) in test_suite_psa_crypto_slot_management. Test various cases of source and target policies with and without wildcards. Missing: testing of the policy constraint on psa_copy_key itself. Test several key types (raw data, AES, RSA). Test with the source or target being persistent. Add failure tests (incompatible policies, source slot empty, target slot occupied).
2019-01-28 13:03:09 +01:00
#else
test_fail( "No hash algorithm for hash-and-sign testing", __LINE__, __FILE__ );
Smoke-test operation contexts after setup+abort After a successful setup followed by abort, or after a failed setup from an inactive state, a context must be usable. Test this for hash, MAC and cipher contexts.
2019-02-25 17:42:03 +01:00
return( 1 );
Test psa_copy_key Split the testing into tests that exercise policies in test_suite_psa_crypto and tests that exercise slot content (slot states, key material) in test_suite_psa_crypto_slot_management. Test various cases of source and target policies with and without wildcards. Missing: testing of the policy constraint on psa_copy_key itself. Test several key types (raw data, AES, RSA). Test with the source or target being persistent. Add failure tests (incompatible policies, source slot empty, target slot occupied).
2019-01-28 13:03:09 +01:00
#endif
}
generate_key tests: exercise the key After generating a key, perform a smoke test: run one operation with it and check that the operation has the expected status.
2018-06-20 18:16:48 +02:00
if( usage & PSA_KEY_USAGE_SIGN )
{
Change a generate_key test to exercise with PSS This required tweaking exercise_signature_key to use a payload size for the signature based on the algorithm, since our implementation of PSS requires that the input size matches the hash size. This would also be the case for PKCS#1 v1.5 with a specified hash.
2018-06-30 00:20:25 +02:00
/* Some algorithms require the payload to have the size of
* the hash encoded in the algorithm. Use this input size
* even for algorithms that allow other input sizes. */
if( hash_alg != 0 )
payload_length = PSA_HASH_SIZE( hash_alg );
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_asymmetric_sign( handle, alg,
payload, payload_length,
signature, sizeof( signature ),
&signature_length ) );
generate_key tests: exercise the key After generating a key, perform a smoke test: run one operation with it and check that the operation has the expected status.
2018-06-20 18:16:48 +02:00
}
if( usage & PSA_KEY_USAGE_VERIFY )
{
psa_status_t verify_status =
( usage & PSA_KEY_USAGE_SIGN ?
PSA_SUCCESS :
PSA_ERROR_INVALID_SIGNATURE );
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( psa_asymmetric_verify( handle, alg,
payload, payload_length,
signature, signature_length ),
verify_status );
generate_key tests: exercise the key After generating a key, perform a smoke test: run one operation with it and check that the operation has the expected status.
2018-06-20 18:16:48 +02:00
}
return( 1 );
exit:
return( 0 );
}
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
static int exercise_asymmetric_encryption_key( psa_key_handle_t handle,
generate_key tests: exercise the key After generating a key, perform a smoke test: run one operation with it and check that the operation has the expected status.
2018-06-20 18:16:48 +02:00
psa_key_usage_t usage,
psa_algorithm_t alg )
{
unsigned char plaintext[256] = "Hello, world...";
unsigned char ciphertext[256] = "(wabblewebblewibblewobblewubble)";
size_t ciphertext_length = sizeof( ciphertext );
size_t plaintext_length = 16;
if( usage & PSA_KEY_USAGE_ENCRYPT )
{
Rewrap some lines after the macro changes Change the way some lines are wrapped to cut at a more logical place. This commit mainly rewrites multi-line calls to TEST_EQUAL, and also a few calls to PSA_ASSERT.
2018-12-18 00:33:25 +01:00
PSA_ASSERT( psa_asymmetric_encrypt( handle, alg,
plaintext, plaintext_length,
NULL, 0,
ciphertext, sizeof( ciphertext ),
&ciphertext_length ) );
generate_key tests: exercise the key After generating a key, perform a smoke test: run one operation with it and check that the operation has the expected status.
2018-06-20 18:16:48 +02:00
}
if( usage & PSA_KEY_USAGE_DECRYPT )
{
psa_status_t status =
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_asymmetric_decrypt( handle, alg,
generate_key tests: exercise the key After generating a key, perform a smoke test: run one operation with it and check that the operation has the expected status.
2018-06-20 18:16:48 +02:00
ciphertext, ciphertext_length,
NULL, 0,
plaintext, sizeof( plaintext ),
&plaintext_length );
TEST_ASSERT( status == PSA_SUCCESS ||
( ( usage & PSA_KEY_USAGE_ENCRYPT ) == 0 &&
( status == PSA_ERROR_INVALID_ARGUMENT ||
status == PSA_ERROR_INVALID_PADDING ) ) );
}
return( 1 );
exit:
return( 0 );
}
Factor duplicated code into exercise_key Also fail the test if the test code lacks a way to exercise the key.
2018-07-01 22:31:34 +02:00
Refactor key derivation setup in tests
2019-07-03 13:41:36 +02:00
static int setup_key_derivation_wrap( psa_key_derivation_operation_t* operation,
psa_key_handle_t handle,
psa_algorithm_t alg,
unsigned char* input1, size_t input1_length,
unsigned char* input2, size_t input2_length,
size_t capacity )
{
PSA_ASSERT( psa_key_derivation_setup( operation, alg ) );
if( PSA_ALG_IS_HKDF( alg ) )
{
PSA_ASSERT( psa_key_derivation_input_bytes( operation,
PSA_KEY_DERIVATION_INPUT_SALT,
input1, input1_length ) );
PSA_ASSERT( psa_key_derivation_input_key( operation,
PSA_KEY_DERIVATION_INPUT_SECRET,
handle ) );
PSA_ASSERT( psa_key_derivation_input_bytes( operation,
PSA_KEY_DERIVATION_INPUT_INFO,
input2,
input2_length ) );
}
else if( PSA_ALG_IS_TLS12_PRF( alg ) ||
PSA_ALG_IS_TLS12_PSK_TO_MS( alg ) )
{
PSA_ASSERT( psa_key_derivation_input_bytes( operation,
PSA_KEY_DERIVATION_INPUT_SEED,
input1, input1_length ) );
PSA_ASSERT( psa_key_derivation_input_key( operation,
PSA_KEY_DERIVATION_INPUT_SECRET,
handle ) );
PSA_ASSERT( psa_key_derivation_input_bytes( operation,
PSA_KEY_DERIVATION_INPUT_LABEL,
input2, input2_length ) );
}
else
{
TEST_ASSERT( ! "Key derivation algorithm not supported" );
}
Limit keys to 65528 bits 65528 bits is more than any reasonable key until we start supporting post-quantum cryptography. This limit is chosen to allow bit-sizes to be stored in 16 bits, with 65535 left to indicate an invalid value. It's a whole number of bytes, which facilitates some calculations, in particular allowing a key of exactly PSA_CRYPTO_MAX_STORAGE_SIZE to be created but not one bit more. As a resource usage limit, this is arguably too large, but that's out of scope of the current commit. Test that key import, generation and derivation reject overly large sizes.
2019-07-30 17:26:54 +02:00
if( capacity != SIZE_MAX )
PSA_ASSERT( psa_key_derivation_set_capacity( operation, capacity ) );
Refactor key derivation setup in tests
2019-07-03 13:41:36 +02:00
return( 1 );
exit:
return( 0 );
}
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
static int exercise_key_derivation_key( psa_key_handle_t handle,
Add framework for simple key derivation New key type PSA_KEY_TYPE_DERIVE. New usage flag PSA_KEY_USAGE_DERIVE. New function psa_key_derivation. No key derivation algorithm is implemented yet. The code may not compile with -Wunused. Write some unit test code for psa_key_derivation. Most of it cannot be used yet due to the lack of a key derivation algorithm.
2018-07-12 17:17:20 +02:00
psa_key_usage_t usage,
psa_algorithm_t alg )
{
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
psa_key_derivation_operation_t operation = PSA_KEY_DERIVATION_OPERATION_INIT;
Refactor key derivation setup in tests
2019-07-03 13:41:36 +02:00
unsigned char input1[] = "Input 1";
size_t input1_length = sizeof( input1 );
unsigned char input2[] = "Input 2";
size_t input2_length = sizeof( input2 );
Add framework for simple key derivation New key type PSA_KEY_TYPE_DERIVE. New usage flag PSA_KEY_USAGE_DERIVE. New function psa_key_derivation. No key derivation algorithm is implemented yet. The code may not compile with -Wunused. Write some unit test code for psa_key_derivation. Most of it cannot be used yet due to the lack of a key derivation algorithm.
2018-07-12 17:17:20 +02:00
unsigned char output[1];
Refactor key derivation setup in tests
2019-07-03 13:41:36 +02:00
size_t capacity = sizeof( output );
Add framework for simple key derivation New key type PSA_KEY_TYPE_DERIVE. New usage flag PSA_KEY_USAGE_DERIVE. New function psa_key_derivation. No key derivation algorithm is implemented yet. The code may not compile with -Wunused. Write some unit test code for psa_key_derivation. Most of it cannot be used yet due to the lack of a key derivation algorithm.
2018-07-12 17:17:20 +02:00
if( usage & PSA_KEY_USAGE_DERIVE )
{
Refactor key derivation setup in tests
2019-07-03 13:41:36 +02:00
if( !setup_key_derivation_wrap( &operation, handle, alg,
input1, input1_length,
input2, input2_length, capacity ) )
goto exit;
Convert exercise_key_derivation_key to the new KDF API
2019-05-29 17:35:00 +02:00
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
PSA_ASSERT( psa_key_derivation_output_bytes( &operation,
Wrap and reindent some lines After renaming several identifiers, re-wrap and re-indent some lines to make the code prettier.
2019-05-16 17:53:40 +02:00
output,
Refactor key derivation setup in tests
2019-07-03 13:41:36 +02:00
capacity ) );
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
PSA_ASSERT( psa_key_derivation_abort( &operation ) );
Add framework for simple key derivation New key type PSA_KEY_TYPE_DERIVE. New usage flag PSA_KEY_USAGE_DERIVE. New function psa_key_derivation. No key derivation algorithm is implemented yet. The code may not compile with -Wunused. Write some unit test code for psa_key_derivation. Most of it cannot be used yet due to the lack of a key derivation algorithm.
2018-07-12 17:17:20 +02:00
}
return( 1 );
exit:
return( 0 );
}
Factor common code into key_agreement_with_self
2018-11-07 18:45:02 +01:00
/* We need two keys to exercise key agreement. Exercise the
* private key against its own public key. */
Wrap and reindent some lines After renaming several identifiers, re-wrap and re-indent some lines to make the code prettier.
2019-05-16 17:53:40 +02:00
static psa_status_t key_agreement_with_self(
psa_key_derivation_operation_t *operation,
psa_key_handle_t handle )
New API function: psa_key_agreement Set up a generator from a key agreement.
2018-09-18 12:01:02 +02:00
{
Factor common code into key_agreement_with_self
2018-11-07 18:45:02 +01:00
psa_key_type_t private_key_type;
New API function: psa_key_agreement Set up a generator from a key agreement.
2018-09-18 12:01:02 +02:00
psa_key_type_t public_key_type;
size_t key_bits;
uint8_t *public_key = NULL;
size_t public_key_length;
Replace PSA error code definitions with the ones defined in PSA spec
2019-02-14 12:48:10 +01:00
/* Return GENERIC_ERROR if something other than the final call to
Wrap and reindent some lines After renaming several identifiers, re-wrap and re-indent some lines to make the code prettier.
2019-05-16 17:53:40 +02:00
* psa_key_derivation_key_agreement fails. This isn't fully satisfactory,
* but it's good enough: callers will report it as a failed test anyway. */
Replace PSA error code definitions with the ones defined in PSA spec
2019-02-14 12:48:10 +01:00
psa_status_t status = PSA_ERROR_GENERIC_ERROR;
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
Factor common code into key_agreement_with_self
2018-11-07 18:45:02 +01:00
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
PSA_ASSERT( psa_get_key_attributes( handle, &attributes ) );
private_key_type = psa_get_key_type( &attributes );
key_bits = psa_get_key_bits( &attributes );
Rename *KEYPAIR* to *KEY_PAIR* Be consistent with PUBLIC_KEY. perl -i -pe 's/KEYPAIR/KEY_PAIR/g' $(git ls-files)
2019-05-16 19:39:54 +02:00
public_key_type = PSA_KEY_TYPE_PUBLIC_KEY_OF_KEY_PAIR( private_key_type );
Factor common code into key_agreement_with_self
2018-11-07 18:45:02 +01:00
public_key_length = PSA_KEY_EXPORT_MAX_SIZE( public_key_type, key_bits );
ASSERT_ALLOC( public_key, public_key_length );
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_export_public_key( handle,
public_key, public_key_length,
&public_key_length ) );
Factor common code into key_agreement_with_self
2018-11-07 18:45:02 +01:00
Wrap and reindent some lines After renaming several identifiers, re-wrap and re-indent some lines to make the code prettier.
2019-05-16 17:53:40 +02:00
status = psa_key_derivation_key_agreement(
operation, PSA_KEY_DERIVATION_INPUT_SECRET, handle,
public_key, public_key_length );
Factor common code into key_agreement_with_self
2018-11-07 18:45:02 +01:00
exit:
mbedtls_free( public_key );
Call psa_reset_key_attributes after psa_get_key_attributes After calling psa_get_key_attributes(), call psa_reset_key_attributes() if the key may have domain parameters, because that's the way to free the domain parameter substructure in the attribute structure. Keep not calling reset() in some places where the key can only be a symmetric key which doesn't have domain parameters.
2019-04-26 16:03:33 +02:00
psa_reset_key_attributes( &attributes );
Factor common code into key_agreement_with_self
2018-11-07 18:45:02 +01:00
return( status );
}
Add exercise_key for raw key agreement
2019-04-11 21:24:55 +02:00
/* We need two keys to exercise key agreement. Exercise the
* private key against its own public key. */
static psa_status_t raw_key_agreement_with_self( psa_algorithm_t alg,
psa_key_handle_t handle )
{
psa_key_type_t private_key_type;
psa_key_type_t public_key_type;
size_t key_bits;
uint8_t *public_key = NULL;
size_t public_key_length;
uint8_t output[1024];
size_t output_length;
/* Return GENERIC_ERROR if something other than the final call to
Wrap and reindent some lines After renaming several identifiers, re-wrap and re-indent some lines to make the code prettier.
2019-05-16 17:53:40 +02:00
* psa_key_derivation_key_agreement fails. This isn't fully satisfactory,
* but it's good enough: callers will report it as a failed test anyway. */
Add exercise_key for raw key agreement
2019-04-11 21:24:55 +02:00
psa_status_t status = PSA_ERROR_GENERIC_ERROR;
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
Add exercise_key for raw key agreement
2019-04-11 21:24:55 +02:00
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
PSA_ASSERT( psa_get_key_attributes( handle, &attributes ) );
private_key_type = psa_get_key_type( &attributes );
key_bits = psa_get_key_bits( &attributes );
Rename *KEYPAIR* to *KEY_PAIR* Be consistent with PUBLIC_KEY. perl -i -pe 's/KEYPAIR/KEY_PAIR/g' $(git ls-files)
2019-05-16 19:39:54 +02:00
public_key_type = PSA_KEY_TYPE_PUBLIC_KEY_OF_KEY_PAIR( private_key_type );
Add exercise_key for raw key agreement
2019-04-11 21:24:55 +02:00
public_key_length = PSA_KEY_EXPORT_MAX_SIZE( public_key_type, key_bits );
ASSERT_ALLOC( public_key, public_key_length );
PSA_ASSERT( psa_export_public_key( handle,
public_key, public_key_length,
&public_key_length ) );
Shorten the name of psa_key_agreement_raw_shared_secret There is less of a risk of confusion with the KA+KDF function now.
2019-05-16 18:00:41 +02:00
status = psa_raw_key_agreement( alg, handle,
public_key, public_key_length,
output, sizeof( output ), &output_length );
Add exercise_key for raw key agreement
2019-04-11 21:24:55 +02:00
exit:
mbedtls_free( public_key );
Call psa_reset_key_attributes after psa_get_key_attributes After calling psa_get_key_attributes(), call psa_reset_key_attributes() if the key may have domain parameters, because that's the way to free the domain parameter substructure in the attribute structure. Keep not calling reset() in some places where the key can only be a symmetric key which doesn't have domain parameters.
2019-04-26 16:03:33 +02:00
psa_reset_key_attributes( &attributes );
Add exercise_key for raw key agreement
2019-04-11 21:24:55 +02:00
return( status );
}
static int exercise_raw_key_agreement_key( psa_key_handle_t handle,
psa_key_usage_t usage,
psa_algorithm_t alg )
{
int ok = 0;
if( usage & PSA_KEY_USAGE_DERIVE )
{
/* We need two keys to exercise key agreement. Exercise the
* private key against its own public key. */
PSA_ASSERT( raw_key_agreement_with_self( alg, handle ) );
}
ok = 1;
exit:
return( ok );
}
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
static int exercise_key_agreement_key( psa_key_handle_t handle,
Factor common code into key_agreement_with_self
2018-11-07 18:45:02 +01:00
psa_key_usage_t usage,
psa_algorithm_t alg )
{
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
psa_key_derivation_operation_t operation = PSA_KEY_DERIVATION_OPERATION_INIT;
New API function: psa_key_agreement Set up a generator from a key agreement.
2018-09-18 12:01:02 +02:00
unsigned char output[1];
int ok = 0;
if( usage & PSA_KEY_USAGE_DERIVE )
{
/* We need two keys to exercise key agreement. Exercise the
* private key against its own public key. */
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
PSA_ASSERT( psa_key_derivation_setup( &operation, alg ) );
PSA_ASSERT( key_agreement_with_self( &operation, handle ) );
PSA_ASSERT( psa_key_derivation_output_bytes( &operation,
Wrap and reindent some lines After renaming several identifiers, re-wrap and re-indent some lines to make the code prettier.
2019-05-16 17:53:40 +02:00
output,
sizeof( output ) ) );
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
PSA_ASSERT( psa_key_derivation_abort( &operation ) );
New API function: psa_key_agreement Set up a generator from a key agreement.
2018-09-18 12:01:02 +02:00
}
ok = 1;
exit:
return( ok );
}
Allow building with the -Wunused flag Make some functions non-static, to avoid Wunused function warnings. Make a function scoped variable block scoped instead, to avoid Wunused variable warnings in some configurations.
2019-06-27 18:31:33 +02:00
int asn1_skip_integer( unsigned char **p, const unsigned char *end,
size_t min_bits, size_t max_bits,
int must_be_odd )
Improve and augment export sanity checks Implement sanity checks of exported public keys, using ASN.1 parsing. Rewrite the sanity checks of key pairs using ASN.1 parsing, so as to check more things with simpler code.
2018-08-11 01:22:42 +02:00
{
size_t len;
size_t actual_bits;
unsigned char msb;
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( mbedtls_asn1_get_tag( p, end, &len,
Rewrap some lines after the macro changes Change the way some lines are wrapped to cut at a more logical place. This commit mainly rewrites multi-line calls to TEST_EQUAL, and also a few calls to PSA_ASSERT.
2018-12-18 00:33:25 +01:00
MBEDTLS_ASN1_INTEGER ),
0 );
Improve and augment export sanity checks Implement sanity checks of exported public keys, using ASN.1 parsing. Rewrite the sanity checks of key pairs using ASN.1 parsing, so as to check more things with simpler code.
2018-08-11 01:22:42 +02:00
/* Tolerate a slight departure from DER encoding:
* - 0 may be represented by an empty string or a 1-byte string.
* - The sign bit may be used as a value bit. */
if( ( len == 1 && ( *p )[0] == 0 ) ||
( len > 1 && ( *p )[0] == 0 && ( ( *p )[1] & 0x80 ) != 0 ) )
{
++( *p );
--len;
}
if( min_bits == 0 && len == 0 )
return( 1 );
msb = ( *p )[0];
TEST_ASSERT( msb != 0 );
actual_bits = 8 * ( len - 1 );
while( msb != 0 )
{
msb >>= 1;
++actual_bits;
}
TEST_ASSERT( actual_bits >= min_bits );
TEST_ASSERT( actual_bits <= max_bits );
if( must_be_odd )
TEST_ASSERT( ( ( *p )[len-1] & 1 ) != 0 );
*p += len;
return( 1 );
exit:
return( 0 );
}
static int exported_key_sanity_check( psa_key_type_t type, size_t bits,
uint8_t *exported, size_t exported_length )
Move export key sanity check from generate to exercise Move the code to perform sanity checks on the exported key from generate_key to exercise_key. This way the sanity checks can be performed after importing or deriving a key as well. In addition to checking the exported key if its usage allows it, check the exported public key if the key is asymmetric.
2018-08-10 19:07:32 +02:00
{
Improve and augment export sanity checks Implement sanity checks of exported public keys, using ASN.1 parsing. Rewrite the sanity checks of key pairs using ASN.1 parsing, so as to check more things with simpler code.
2018-08-11 01:22:42 +02:00
if( PSA_KEY_TYPE_IS_UNSTRUCTURED( type ) )
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( exported_length, ( bits + 7 ) / 8 );
Improve and augment export sanity checks Implement sanity checks of exported public keys, using ASN.1 parsing. Rewrite the sanity checks of key pairs using ASN.1 parsing, so as to check more things with simpler code.
2018-08-11 01:22:42 +02:00
else
TEST_ASSERT( exported_length <= PSA_KEY_EXPORT_MAX_SIZE( type, bits ) );
Move export key sanity check from generate to exercise Move the code to perform sanity checks on the exported key from generate_key to exercise_key. This way the sanity checks can be performed after importing or deriving a key as well. In addition to checking the exported key if its usage allows it, check the exported public key if the key is asymmetric.
2018-08-10 19:07:32 +02:00
#if defined(MBEDTLS_DES_C)
if( type == PSA_KEY_TYPE_DES )
{
/* Check the parity bits. */
unsigned i;
for( i = 0; i < bits / 8; i++ )
{
unsigned bit_count = 0;
unsigned m;
for( m = 1; m <= 0x100; m <<= 1 )
{
if( exported[i] & m )
++bit_count;
}
TEST_ASSERT( bit_count % 2 != 0 );
}
}
Improve and augment export sanity checks Implement sanity checks of exported public keys, using ASN.1 parsing. Rewrite the sanity checks of key pairs using ASN.1 parsing, so as to check more things with simpler code.
2018-08-11 01:22:42 +02:00
else
Move export key sanity check from generate to exercise Move the code to perform sanity checks on the exported key from generate_key to exercise_key. This way the sanity checks can be performed after importing or deriving a key as well. In addition to checking the exported key if its usage allows it, check the exported public key if the key is asymmetric.
2018-08-10 19:07:32 +02:00
#endif
#if defined(MBEDTLS_RSA_C) && defined(MBEDTLS_PK_PARSE_C)
Rename *KEYPAIR* to *KEY_PAIR* Be consistent with PUBLIC_KEY. perl -i -pe 's/KEYPAIR/KEY_PAIR/g' $(git ls-files)
2019-05-16 19:39:54 +02:00
if( type == PSA_KEY_TYPE_RSA_KEY_PAIR )
Move export key sanity check from generate to exercise Move the code to perform sanity checks on the exported key from generate_key to exercise_key. This way the sanity checks can be performed after importing or deriving a key as well. In addition to checking the exported key if its usage allows it, check the exported public key if the key is asymmetric.
2018-08-10 19:07:32 +02:00
{
Improve and augment export sanity checks Implement sanity checks of exported public keys, using ASN.1 parsing. Rewrite the sanity checks of key pairs using ASN.1 parsing, so as to check more things with simpler code.
2018-08-11 01:22:42 +02:00
uint8_t *p = exported;
uint8_t *end = exported + exported_length;
size_t len;
/* RSAPrivateKey ::= SEQUENCE {
Clarify comment in test In RSAPrivateKey, Version is an INTEGER. The version must be 0.
2018-08-21 16:12:54 +02:00
* version INTEGER, -- must be 0
Improve and augment export sanity checks Implement sanity checks of exported public keys, using ASN.1 parsing. Rewrite the sanity checks of key pairs using ASN.1 parsing, so as to check more things with simpler code.
2018-08-11 01:22:42 +02:00
* modulus INTEGER, -- n
* publicExponent INTEGER, -- e
* privateExponent INTEGER, -- d
* prime1 INTEGER, -- p
* prime2 INTEGER, -- q
* exponent1 INTEGER, -- d mod (p-1)
* exponent2 INTEGER, -- d mod (q-1)
* coefficient INTEGER, -- (inverse of q) mod p
* }
*/
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( mbedtls_asn1_get_tag( &p, end, &len,
MBEDTLS_ASN1_SEQUENCE |
MBEDTLS_ASN1_CONSTRUCTED ), 0 );
TEST_EQUAL( p + len, end );
Improve and augment export sanity checks Implement sanity checks of exported public keys, using ASN.1 parsing. Rewrite the sanity checks of key pairs using ASN.1 parsing, so as to check more things with simpler code.
2018-08-11 01:22:42 +02:00
if( ! asn1_skip_integer( &p, end, 0, 0, 0 ) )
goto exit;
if( ! asn1_skip_integer( &p, end, bits, bits, 1 ) )
goto exit;
if( ! asn1_skip_integer( &p, end, 2, bits, 1 ) )
goto exit;
/* Require d to be at least half the size of n. */
if( ! asn1_skip_integer( &p, end, bits / 2, bits, 1 ) )
goto exit;
/* Require p and q to be at most half the size of n, rounded up. */
if( ! asn1_skip_integer( &p, end, bits / 2, bits / 2 + 1, 1 ) )
goto exit;
if( ! asn1_skip_integer( &p, end, bits / 2, bits / 2 + 1, 1 ) )
goto exit;
if( ! asn1_skip_integer( &p, end, 1, bits / 2 + 1, 0 ) )
goto exit;
if( ! asn1_skip_integer( &p, end, 1, bits / 2 + 1, 0 ) )
goto exit;
if( ! asn1_skip_integer( &p, end, 1, bits / 2 + 1, 0 ) )
goto exit;
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( p, end );
Indent PSA tests according to K&R rules with Mbed TLS tweaks Only whitespace changes in this commit.
2018-12-17 23:35:42 +01:00
}
Improve and augment export sanity checks Implement sanity checks of exported public keys, using ASN.1 parsing. Rewrite the sanity checks of key pairs using ASN.1 parsing, so as to check more things with simpler code.
2018-08-11 01:22:42 +02:00
else
Move export key sanity check from generate to exercise Move the code to perform sanity checks on the exported key from generate_key to exercise_key. This way the sanity checks can be performed after importing or deriving a key as well. In addition to checking the exported key if its usage allows it, check the exported public key if the key is asymmetric.
2018-08-10 19:07:32 +02:00
#endif /* MBEDTLS_RSA_C */
#if defined(MBEDTLS_ECP_C)
Rename *KEYPAIR* to *KEY_PAIR* Be consistent with PUBLIC_KEY. perl -i -pe 's/KEYPAIR/KEY_PAIR/g' $(git ls-files)
2019-05-16 19:39:54 +02:00
if( PSA_KEY_TYPE_IS_ECC_KEY_PAIR( type ) )
Move export key sanity check from generate to exercise Move the code to perform sanity checks on the exported key from generate_key to exercise_key. This way the sanity checks can be performed after importing or deriving a key as well. In addition to checking the exported key if its usage allows it, check the exported public key if the key is asymmetric.
2018-08-10 19:07:32 +02:00
{
Private EC key format: remove ASN.1-based sanity checks In preparation for the import/export format change for private elliptic curve keys from RFC 5915 to the raw secret value, remove ASN.1-based sanity checks. For the raw secret value, most byte strings of the correct length are valid (the details depend on the curve), so as a sanity check, just check the length.
2018-10-29 19:21:41 +01:00
/* Just the secret value */
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( exported_length, PSA_BITS_TO_BYTES( bits ) );
Private EC key format: remove ASN.1-based sanity checks In preparation for the import/export format change for private elliptic curve keys from RFC 5915 to the raw secret value, remove ASN.1-based sanity checks. For the raw secret value, most byte strings of the correct length are valid (the details depend on the curve), so as a sanity check, just check the length.
2018-10-29 19:21:41 +01:00
}
Improve and augment export sanity checks Implement sanity checks of exported public keys, using ASN.1 parsing. Rewrite the sanity checks of key pairs using ASN.1 parsing, so as to check more things with simpler code.
2018-08-11 01:22:42 +02:00
else
Move export key sanity check from generate to exercise Move the code to perform sanity checks on the exported key from generate_key to exercise_key. This way the sanity checks can be performed after importing or deriving a key as well. In addition to checking the exported key if its usage allows it, check the exported public key if the key is asymmetric.
2018-08-10 19:07:32 +02:00
#endif /* MBEDTLS_ECP_C */
Improve and augment export sanity checks Implement sanity checks of exported public keys, using ASN.1 parsing. Rewrite the sanity checks of key pairs using ASN.1 parsing, so as to check more things with simpler code.
2018-08-11 01:22:42 +02:00
if( PSA_KEY_TYPE_IS_PUBLIC_KEY( type ) )
{
uint8_t *p = exported;
uint8_t *end = exported + exported_length;
#if defined(MBEDTLS_RSA_C)
if( type == PSA_KEY_TYPE_RSA_PUBLIC_KEY )
{
Allow building with the -Wunused flag Make some functions non-static, to avoid Wunused function warnings. Make a function scoped variable block scoped instead, to avoid Wunused variable warnings in some configurations.
2019-06-27 18:31:33 +02:00
size_t len;
Improve and augment export sanity checks Implement sanity checks of exported public keys, using ASN.1 parsing. Rewrite the sanity checks of key pairs using ASN.1 parsing, so as to check more things with simpler code.
2018-08-11 01:22:42 +02:00
/* RSAPublicKey ::= SEQUENCE {
* modulus INTEGER, -- n
* publicExponent INTEGER } -- e
*/
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( mbedtls_asn1_get_tag( &p, end, &len,
MBEDTLS_ASN1_SEQUENCE |
Rewrap some lines after the macro changes Change the way some lines are wrapped to cut at a more logical place. This commit mainly rewrites multi-line calls to TEST_EQUAL, and also a few calls to PSA_ASSERT.
2018-12-18 00:33:25 +01:00
MBEDTLS_ASN1_CONSTRUCTED ),
0 );
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( p + len, end );
Improve and augment export sanity checks Implement sanity checks of exported public keys, using ASN.1 parsing. Rewrite the sanity checks of key pairs using ASN.1 parsing, so as to check more things with simpler code.
2018-08-11 01:22:42 +02:00
if( ! asn1_skip_integer( &p, end, bits, bits, 1 ) )
goto exit;
if( ! asn1_skip_integer( &p, end, 2, bits, 1 ) )
goto exit;
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( p, end );
Improve and augment export sanity checks Implement sanity checks of exported public keys, using ASN.1 parsing. Rewrite the sanity checks of key pairs using ASN.1 parsing, so as to check more things with simpler code.
2018-08-11 01:22:42 +02:00
}
else
#endif /* MBEDTLS_RSA_C */
#if defined(MBEDTLS_ECP_C)
if( PSA_KEY_TYPE_IS_ECC_PUBLIC_KEY( type ) )
{
psa: Simplify EC public key format Remove front matter from our EC key format, to make it just the contents of an ECPoint as defined by SEC1 section 2.3.3. As a consequence of the simplification, remove the restriction on not being able to use an ECDH key with ECDSA. There is no longer any OID specified when importing a key, so we can't reject importing of an ECDH key for the purpose of ECDSA based on the OID.
2019-01-10 12:42:27 +01:00
/* The representation of an ECC public key is:
* - The byte 0x04;
* - `x_P` as a `ceiling(m/8)`-byte string, big-endian;
* - `y_P` as a `ceiling(m/8)`-byte string, big-endian;
* - where m is the bit size associated with the curve.
psa: Simplify RSA public key format Remove pkcs-1 and rsaEncryption front matter from RSA public keys. Move code that was shared between RSA and other key types (like EC keys) to be used only with non-RSA keys.
2019-01-10 11:23:21 +01:00
*/
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( p + 1 + 2 * PSA_BITS_TO_BYTES( bits ), end );
TEST_EQUAL( p[0], 4 );
Improve and augment export sanity checks Implement sanity checks of exported public keys, using ASN.1 parsing. Rewrite the sanity checks of key pairs using ASN.1 parsing, so as to check more things with simpler code.
2018-08-11 01:22:42 +02:00
}
else
#endif /* MBEDTLS_ECP_C */
{
psa: test: Fix truncation of message by snprintf We had only allocated 40 bytes for printing into, but we wanted to print 46 bytes. Update the buffer to be 47 bytes, which is large enough to hold what we want to print plus a terminating null byte.
2018-10-26 18:07:22 +02:00
char message[47];
Improve and augment export sanity checks Implement sanity checks of exported public keys, using ASN.1 parsing. Rewrite the sanity checks of key pairs using ASN.1 parsing, so as to check more things with simpler code.
2018-08-11 01:22:42 +02:00
mbedtls_snprintf( message, sizeof( message ),
"No sanity check for public key type=0x%08lx",
(unsigned long) type );
test_fail( message, __LINE__, __FILE__ );
return( 0 );
}
}
else
{
/* No sanity checks for other types */
}
Move export key sanity check from generate to exercise Move the code to perform sanity checks on the exported key from generate_key to exercise_key. This way the sanity checks can be performed after importing or deriving a key as well. In addition to checking the exported key if its usage allows it, check the exported public key if the key is asymmetric.
2018-08-10 19:07:32 +02:00
return( 1 );
Improve and augment export sanity checks Implement sanity checks of exported public keys, using ASN.1 parsing. Rewrite the sanity checks of key pairs using ASN.1 parsing, so as to check more things with simpler code.
2018-08-11 01:22:42 +02:00
exit:
return( 0 );
Move export key sanity check from generate to exercise Move the code to perform sanity checks on the exported key from generate_key to exercise_key. This way the sanity checks can be performed after importing or deriving a key as well. In addition to checking the exported key if its usage allows it, check the exported public key if the key is asymmetric.
2018-08-10 19:07:32 +02:00
}
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
static int exercise_export_key( psa_key_handle_t handle,
Move export key sanity check from generate to exercise Move the code to perform sanity checks on the exported key from generate_key to exercise_key. This way the sanity checks can be performed after importing or deriving a key as well. In addition to checking the exported key if its usage allows it, check the exported public key if the key is asymmetric.
2018-08-10 19:07:32 +02:00
psa_key_usage_t usage )
{
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
Move export key sanity check from generate to exercise Move the code to perform sanity checks on the exported key from generate_key to exercise_key. This way the sanity checks can be performed after importing or deriving a key as well. In addition to checking the exported key if its usage allows it, check the exported public key if the key is asymmetric.
2018-08-10 19:07:32 +02:00
uint8_t *exported = NULL;
size_t exported_size = 0;
size_t exported_length = 0;
int ok = 0;
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
PSA_ASSERT( psa_get_key_attributes( handle, &attributes ) );
exercise_export_key: fix public key case Public keys are always exportable, even if their usage doesn't include the EXPORT flag.
2018-09-13 20:34:11 +02:00
if( ( usage & PSA_KEY_USAGE_EXPORT ) == 0 &&
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
! PSA_KEY_TYPE_IS_PUBLIC_KEY( psa_get_key_type( &attributes ) ) )
Move export key sanity check from generate to exercise Move the code to perform sanity checks on the exported key from generate_key to exercise_key. This way the sanity checks can be performed after importing or deriving a key as well. In addition to checking the exported key if its usage allows it, check the exported public key if the key is asymmetric.
2018-08-10 19:07:32 +02:00
{
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( psa_export_key( handle, NULL, 0, &exported_length ),
PSA_ERROR_NOT_PERMITTED );
Call psa_reset_key_attributes after psa_get_key_attributes After calling psa_get_key_attributes(), call psa_reset_key_attributes() if the key may have domain parameters, because that's the way to free the domain parameter substructure in the attribute structure. Keep not calling reset() in some places where the key can only be a symmetric key which doesn't have domain parameters.
2019-04-26 16:03:33 +02:00
ok = 1;
goto exit;
Move export key sanity check from generate to exercise Move the code to perform sanity checks on the exported key from generate_key to exercise_key. This way the sanity checks can be performed after importing or deriving a key as well. In addition to checking the exported key if its usage allows it, check the exported public key if the key is asymmetric.
2018-08-10 19:07:32 +02:00
}
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
exported_size = PSA_KEY_EXPORT_MAX_SIZE( psa_get_key_type( &attributes ),
psa_get_key_bits( &attributes ) );
Use ASSERT_ALLOC instead of mbedtls_calloc in PSA tests This commit resolves a bug whereby some test cases failed on systems where mbedtls_calloc returns NULL when the size of 0, because the test case asserted `pointer != NULL` regardless of the size.
2018-09-27 13:54:18 +02:00
ASSERT_ALLOC( exported, exported_size );
Move export key sanity check from generate to exercise Move the code to perform sanity checks on the exported key from generate_key to exercise_key. This way the sanity checks can be performed after importing or deriving a key as well. In addition to checking the exported key if its usage allows it, check the exported public key if the key is asymmetric.
2018-08-10 19:07:32 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_export_key( handle,
exported, exported_size,
&exported_length ) );
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
ok = exported_key_sanity_check( psa_get_key_type( &attributes ),
psa_get_key_bits( &attributes ),
exported, exported_length );
Move export key sanity check from generate to exercise Move the code to perform sanity checks on the exported key from generate_key to exercise_key. This way the sanity checks can be performed after importing or deriving a key as well. In addition to checking the exported key if its usage allows it, check the exported public key if the key is asymmetric.
2018-08-10 19:07:32 +02:00
exit:
mbedtls_free( exported );
Call psa_reset_key_attributes after psa_get_key_attributes After calling psa_get_key_attributes(), call psa_reset_key_attributes() if the key may have domain parameters, because that's the way to free the domain parameter substructure in the attribute structure. Keep not calling reset() in some places where the key can only be a symmetric key which doesn't have domain parameters.
2019-04-26 16:03:33 +02:00
psa_reset_key_attributes( &attributes );
Move export key sanity check from generate to exercise Move the code to perform sanity checks on the exported key from generate_key to exercise_key. This way the sanity checks can be performed after importing or deriving a key as well. In addition to checking the exported key if its usage allows it, check the exported public key if the key is asymmetric.
2018-08-10 19:07:32 +02:00
return( ok );
}
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
static int exercise_export_public_key( psa_key_handle_t handle )
Move export key sanity check from generate to exercise Move the code to perform sanity checks on the exported key from generate_key to exercise_key. This way the sanity checks can be performed after importing or deriving a key as well. In addition to checking the exported key if its usage allows it, check the exported public key if the key is asymmetric.
2018-08-10 19:07:32 +02:00
{
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
Move export key sanity check from generate to exercise Move the code to perform sanity checks on the exported key from generate_key to exercise_key. This way the sanity checks can be performed after importing or deriving a key as well. In addition to checking the exported key if its usage allows it, check the exported public key if the key is asymmetric.
2018-08-10 19:07:32 +02:00
psa_key_type_t public_type;
uint8_t *exported = NULL;
size_t exported_size = 0;
size_t exported_length = 0;
int ok = 0;
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
PSA_ASSERT( psa_get_key_attributes( handle, &attributes ) );
if( ! PSA_KEY_TYPE_IS_ASYMMETRIC( psa_get_key_type( &attributes ) ) )
Move export key sanity check from generate to exercise Move the code to perform sanity checks on the exported key from generate_key to exercise_key. This way the sanity checks can be performed after importing or deriving a key as well. In addition to checking the exported key if its usage allows it, check the exported public key if the key is asymmetric.
2018-08-10 19:07:32 +02:00
{
Rewrap some lines after the macro changes Change the way some lines are wrapped to cut at a more logical place. This commit mainly rewrites multi-line calls to TEST_EQUAL, and also a few calls to PSA_ASSERT.
2018-12-18 00:33:25 +01:00
TEST_EQUAL( psa_export_public_key( handle, NULL, 0, &exported_length ),
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
PSA_ERROR_INVALID_ARGUMENT );
Move export key sanity check from generate to exercise Move the code to perform sanity checks on the exported key from generate_key to exercise_key. This way the sanity checks can be performed after importing or deriving a key as well. In addition to checking the exported key if its usage allows it, check the exported public key if the key is asymmetric.
2018-08-10 19:07:32 +02:00
return( 1 );
}
Rename *KEYPAIR* to *KEY_PAIR* Be consistent with PUBLIC_KEY. perl -i -pe 's/KEYPAIR/KEY_PAIR/g' $(git ls-files)
2019-05-16 19:39:54 +02:00
public_type = PSA_KEY_TYPE_PUBLIC_KEY_OF_KEY_PAIR(
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
psa_get_key_type( &attributes ) );
exported_size = PSA_KEY_EXPORT_MAX_SIZE( public_type,
psa_get_key_bits( &attributes ) );
Use ASSERT_ALLOC instead of mbedtls_calloc in PSA tests This commit resolves a bug whereby some test cases failed on systems where mbedtls_calloc returns NULL when the size of 0, because the test case asserted `pointer != NULL` regardless of the size.
2018-09-27 13:54:18 +02:00
ASSERT_ALLOC( exported, exported_size );
Move export key sanity check from generate to exercise Move the code to perform sanity checks on the exported key from generate_key to exercise_key. This way the sanity checks can be performed after importing or deriving a key as well. In addition to checking the exported key if its usage allows it, check the exported public key if the key is asymmetric.
2018-08-10 19:07:32 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_export_public_key( handle,
exported, exported_size,
&exported_length ) );
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
ok = exported_key_sanity_check( public_type,
psa_get_key_bits( &attributes ),
Move export key sanity check from generate to exercise Move the code to perform sanity checks on the exported key from generate_key to exercise_key. This way the sanity checks can be performed after importing or deriving a key as well. In addition to checking the exported key if its usage allows it, check the exported public key if the key is asymmetric.
2018-08-10 19:07:32 +02:00
exported, exported_length );
exit:
mbedtls_free( exported );
Call psa_reset_key_attributes after psa_get_key_attributes After calling psa_get_key_attributes(), call psa_reset_key_attributes() if the key may have domain parameters, because that's the way to free the domain parameter substructure in the attribute structure. Keep not calling reset() in some places where the key can only be a symmetric key which doesn't have domain parameters.
2019-04-26 16:03:33 +02:00
psa_reset_key_attributes( &attributes );
Move export key sanity check from generate to exercise Move the code to perform sanity checks on the exported key from generate_key to exercise_key. This way the sanity checks can be performed after importing or deriving a key as well. In addition to checking the exported key if its usage allows it, check the exported public key if the key is asymmetric.
2018-08-10 19:07:32 +02:00
return( ok );
}
Document exercise_key and fix one incorrect usage In one place, exercise_key was used in a such a way that if the test failed inside exercise_key, the test suite would correctly report the test as failed but would not report the exact location of the failure. Fix this. Add documentation for exercise_key that explains how to use it.
2019-02-05 20:32:06 +01:00
/** Do smoke tests on a key.
*
* Perform one of each operation indicated by \p alg (decrypt/encrypt,
* sign/verify, or derivation) that is permitted according to \p usage.
* \p usage and \p alg should correspond to the expected policy on the
* key.
*
* Export the key if permitted by \p usage, and check that the output
* looks sensible. If \p usage forbids export, check that
* \p psa_export_key correctly rejects the attempt. If the key is
* asymmetric, also check \p psa_export_public_key.
*
* If the key fails the tests, this function calls the test framework's
* `test_fail` function and returns false. Otherwise this function returns
* true. Therefore it should be used as follows:
* ```
* if( ! exercise_key( ... ) ) goto exit;
* ```
*
* \param handle The key to exercise. It should be capable of performing
* \p alg.
* \param usage The usage flags to assume.
* \param alg The algorithm to exercise.
*
* \retval 0 The key failed the smoke tests.
* \retval 1 The key passed the smoke tests.
*/
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
static int exercise_key( psa_key_handle_t handle,
Factor duplicated code into exercise_key Also fail the test if the test code lacks a way to exercise the key.
2018-07-01 22:31:34 +02:00
psa_key_usage_t usage,
psa_algorithm_t alg )
{
int ok;
if( alg == 0 )
ok = 1; /* If no algorihm, do nothing (used for raw data "keys"). */
else if( PSA_ALG_IS_MAC( alg ) )
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
ok = exercise_mac_key( handle, usage, alg );
Factor duplicated code into exercise_key Also fail the test if the test code lacks a way to exercise the key.
2018-07-01 22:31:34 +02:00
else if( PSA_ALG_IS_CIPHER( alg ) )
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
ok = exercise_cipher_key( handle, usage, alg );
Factor duplicated code into exercise_key Also fail the test if the test code lacks a way to exercise the key.
2018-07-01 22:31:34 +02:00
else if( PSA_ALG_IS_AEAD( alg ) )
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
ok = exercise_aead_key( handle, usage, alg );
Factor duplicated code into exercise_key Also fail the test if the test code lacks a way to exercise the key.
2018-07-01 22:31:34 +02:00
else if( PSA_ALG_IS_SIGN( alg ) )
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
ok = exercise_signature_key( handle, usage, alg );
Factor duplicated code into exercise_key Also fail the test if the test code lacks a way to exercise the key.
2018-07-01 22:31:34 +02:00
else if( PSA_ALG_IS_ASYMMETRIC_ENCRYPTION( alg ) )
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
ok = exercise_asymmetric_encryption_key( handle, usage, alg );
Add framework for simple key derivation New key type PSA_KEY_TYPE_DERIVE. New usage flag PSA_KEY_USAGE_DERIVE. New function psa_key_derivation. No key derivation algorithm is implemented yet. The code may not compile with -Wunused. Write some unit test code for psa_key_derivation. Most of it cannot be used yet due to the lack of a key derivation algorithm.
2018-07-12 17:17:20 +02:00
else if( PSA_ALG_IS_KEY_DERIVATION( alg ) )
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
ok = exercise_key_derivation_key( handle, usage, alg );
Add exercise_key for raw key agreement
2019-04-11 21:24:55 +02:00
else if( PSA_ALG_IS_RAW_KEY_AGREEMENT( alg ) )
ok = exercise_raw_key_agreement_key( handle, usage, alg );
New API function: psa_key_agreement Set up a generator from a key agreement.
2018-09-18 12:01:02 +02:00
else if( PSA_ALG_IS_KEY_AGREEMENT( alg ) )
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
ok = exercise_key_agreement_key( handle, usage, alg );
Factor duplicated code into exercise_key Also fail the test if the test code lacks a way to exercise the key.
2018-07-01 22:31:34 +02:00
else
{
char message[40];
mbedtls_snprintf( message, sizeof( message ),
"No code to exercise alg=0x%08lx",
(unsigned long) alg );
test_fail( message, __LINE__, __FILE__ );
ok = 0;
}
Move export key sanity check from generate to exercise Move the code to perform sanity checks on the exported key from generate_key to exercise_key. This way the sanity checks can be performed after importing or deriving a key as well. In addition to checking the exported key if its usage allows it, check the exported public key if the key is asymmetric.
2018-08-10 19:07:32 +02:00
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
ok = ok && exercise_export_key( handle, usage );
ok = ok && exercise_export_public_key( handle );
Move export key sanity check from generate to exercise Move the code to perform sanity checks on the exported key from generate_key to exercise_key. This way the sanity checks can be performed after importing or deriving a key as well. In addition to checking the exported key if its usage allows it, check the exported public key if the key is asymmetric.
2018-08-10 19:07:32 +02:00
Factor duplicated code into exercise_key Also fail the test if the test code lacks a way to exercise the key.
2018-07-01 22:31:34 +02:00
return( ok );
}
Factor usage_to_exercise into its own function
2018-10-25 22:35:43 +02:00
static psa_key_usage_t usage_to_exercise( psa_key_type_t type,
psa_algorithm_t alg )
{
if( PSA_ALG_IS_MAC( alg ) || PSA_ALG_IS_SIGN( alg ) )
{
return( PSA_KEY_TYPE_IS_PUBLIC_KEY( type ) ?
PSA_KEY_USAGE_VERIFY :
PSA_KEY_USAGE_SIGN | PSA_KEY_USAGE_VERIFY );
}
else if( PSA_ALG_IS_CIPHER( alg ) || PSA_ALG_IS_AEAD( alg ) ||
PSA_ALG_IS_ASYMMETRIC_ENCRYPTION( alg ) )
{
return( PSA_KEY_TYPE_IS_PUBLIC_KEY( type ) ?
PSA_KEY_USAGE_ENCRYPT :
PSA_KEY_USAGE_ENCRYPT | PSA_KEY_USAGE_DECRYPT );
}
else if( PSA_ALG_IS_KEY_DERIVATION( alg ) ||
PSA_ALG_IS_KEY_AGREEMENT( alg ) )
{
return( PSA_KEY_USAGE_DERIVE );
}
else
{
return( 0 );
}
}
psa: Extend psa_generate_key to support persistent lifetimes
2018-11-07 17:05:30 +01:00
Simplify and expand invalid-handle tests Simplify invalid-handle tests and make them test more things. Call these tests in several test functions after destroying a key.
2019-04-18 22:28:52 +02:00
static int test_operations_on_invalid_handle( psa_key_handle_t handle )
{
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
uint8_t buffer[1];
size_t length;
int ok = 0;
Replace psa_make_key_persistent by id/lifetime setters in tests Remove all internal uses of psa_make_key_persistent.
2019-05-15 16:12:22 +02:00
psa_set_key_id( &attributes, 0x6964 );
Simplify and expand invalid-handle tests Simplify invalid-handle tests and make them test more things. Call these tests in several test functions after destroying a key.
2019-04-18 22:28:52 +02:00
psa_set_key_usage_flags( &attributes, PSA_KEY_USAGE_ENCRYPT );
psa_set_key_algorithm( &attributes, PSA_ALG_CTR );
psa_set_key_type( &attributes, PSA_KEY_TYPE_AES );
TEST_EQUAL( psa_get_key_attributes( handle, &attributes ),
PSA_ERROR_INVALID_HANDLE );
TEST_EQUAL( psa_get_key_id( &attributes ), 0 );
Update persistent_key_load_key_from_storage to use attributes Update persistent_key_load_key_from_storage to the new attribute-based key creation interface. I tweaked the code a little to make it simpler and more robust without changing the core logic.
2019-04-19 14:06:53 +02:00
TEST_EQUAL( psa_get_key_lifetime( &attributes ), 0 );
Simplify and expand invalid-handle tests Simplify invalid-handle tests and make them test more things. Call these tests in several test functions after destroying a key.
2019-04-18 22:28:52 +02:00
TEST_EQUAL( psa_get_key_usage_flags( &attributes ), 0 );
TEST_EQUAL( psa_get_key_algorithm( &attributes ), 0 );
TEST_EQUAL( psa_get_key_type( &attributes ), 0 );
TEST_EQUAL( psa_get_key_bits( &attributes ), 0 );
TEST_EQUAL( psa_export_key( handle,
buffer, sizeof( buffer ), &length ),
PSA_ERROR_INVALID_HANDLE );
TEST_EQUAL( psa_export_public_key( handle,
buffer, sizeof( buffer ), &length ),
PSA_ERROR_INVALID_HANDLE );
ok = 1;
exit:
psa_reset_key_attributes( &attributes );
return( ok );
}
Test slot_number attribute Test the behavior of the getter/setter functions. Test that psa_get_key_slot_number() reports a slot number for a key in a secure element, and doesn't report a slot number for a key that is not in a secure element. Test that psa_get_key_slot_number() reports the correct slot number for a key in a secure element.
2019-08-02 20:30:01 +02:00
/* Assert that a key isn't reported as having a slot number. */
#if defined(MBEDTLS_PSA_CRYPTO_SE_C)
#define ASSERT_NO_SLOT_NUMBER( attributes ) \
do \
{ \
psa_key_slot_number_t ASSERT_NO_SLOT_NUMBER_slot_number; \
TEST_EQUAL( psa_get_key_slot_number( \
attributes, \
&ASSERT_NO_SLOT_NUMBER_slot_number ), \
PSA_ERROR_INVALID_ARGUMENT ); \
} \
while( 0 )
#else /* MBEDTLS_PSA_CRYPTO_SE_C */
#define ASSERT_NO_SLOT_NUMBER( attributes ) \
( (void) 0 )
#endif /* MBEDTLS_PSA_CRYPTO_SE_C */
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
/* An overapproximation of the amount of storage needed for a key of the
* given type and with the given content. The API doesn't make it easy
* to find a good value for the size. The current implementation doesn't
* care about the value anyway. */
#define KEY_BITS_FROM_DATA( type, data ) \
( data )->len
psa: Extend psa_generate_key to support persistent lifetimes
2018-11-07 17:05:30 +01:00
typedef enum {
IMPORT_KEY = 0,
GENERATE_KEY = 1,
DERIVE_KEY = 2
} generate_method;
Add PSA crypto module New module psa_crypto.c (MBEDTLS_PSA_CRYPTO_C): Platform Security Architecture compatibility layer on top of libmedcrypto. Implement psa_crypto_init function which sets up a RNG. Add a mbedtls_psa_crypto_free function which deinitializes the library. Define a first batch of error codes.
2018-01-27 23:32:46 +01:00
/* END_HEADER */
/* BEGIN_DEPENDENCIES
* depends_on:MBEDTLS_PSA_CRYPTO_C
* END_DEPENDENCIES
*/
Document and check the consistency of truncated MAC encodings Add comments noting that the maximum length of a MAC must fit in PSA_ALG_MAC_TRUNCATION_MASK. Add a unit test that verifies that the maximum MAC size fits.
2018-08-21 14:54:54 +02:00
/* BEGIN_CASE */
void static_checks( )
{
size_t max_truncated_mac_size =
PSA_ALG_MAC_TRUNCATION_MASK >> PSA_MAC_TRUNCATION_OFFSET;
/* Check that the length for a truncated MAC always fits in the algorithm
* encoding. The shifted mask is the maximum truncated value. The
* untruncated algorithm may be one byte larger. */
TEST_ASSERT( PSA_MAC_MAX_SIZE <= 1 + max_truncated_mac_size );
}
/* END_CASE */
Add key management functions Define psa_key_type_t and a first stab at a few values. New functions psa_import_key, psa_export_key, psa_destroy_key, psa_get_key_information. Implement them for raw data and RSA. Under the hood, create an in-memory, fixed-size keystore with room for MBEDTLS_PSA_KEY_SLOT_COUNT - 1 keys.
2018-01-28 13:16:24 +01:00
/* BEGIN_CASE */
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
void attributes_set_get( int id_arg, int lifetime_arg,
int usage_flags_arg, int alg_arg,
Set the key size as an attribute Instead of passing a separate parameter for the key size to psa_generate_key and psa_generator_import_key, set it through the attributes, like the key type and other metadata.
2019-04-26 13:49:28 +02:00
int type_arg, int bits_arg )
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
{
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
psa_key_id_t id = id_arg;
psa_key_lifetime_t lifetime = lifetime_arg;
psa_key_usage_t usage_flags = usage_flags_arg;
psa_algorithm_t alg = alg_arg;
psa_key_type_t type = type_arg;
Set the key size as an attribute Instead of passing a separate parameter for the key size to psa_generate_key and psa_generator_import_key, set it through the attributes, like the key type and other metadata.
2019-04-26 13:49:28 +02:00
size_t bits = bits_arg;
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
TEST_EQUAL( psa_get_key_id( &attributes ), 0 );
TEST_EQUAL( psa_get_key_lifetime( &attributes ), 0 );
TEST_EQUAL( psa_get_key_usage_flags( &attributes ), 0 );
TEST_EQUAL( psa_get_key_algorithm( &attributes ), 0 );
TEST_EQUAL( psa_get_key_type( &attributes ), 0 );
Set the key size as an attribute Instead of passing a separate parameter for the key size to psa_generate_key and psa_generator_import_key, set it through the attributes, like the key type and other metadata.
2019-04-26 13:49:28 +02:00
TEST_EQUAL( psa_get_key_bits( &attributes ), 0 );
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
Replace psa_make_key_persistent by id/lifetime setters in tests Remove all internal uses of psa_make_key_persistent.
2019-05-15 16:12:22 +02:00
psa_set_key_id( &attributes, id );
psa_set_key_lifetime( &attributes, lifetime );
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
psa_set_key_usage_flags( &attributes, usage_flags );
psa_set_key_algorithm( &attributes, alg );
psa_set_key_type( &attributes, type );
Set the key size as an attribute Instead of passing a separate parameter for the key size to psa_generate_key and psa_generator_import_key, set it through the attributes, like the key type and other metadata.
2019-04-26 13:49:28 +02:00
psa_set_key_bits( &attributes, bits );
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
TEST_EQUAL( psa_get_key_id( &attributes ), id );
TEST_EQUAL( psa_get_key_lifetime( &attributes ), lifetime );
TEST_EQUAL( psa_get_key_usage_flags( &attributes ), usage_flags );
TEST_EQUAL( psa_get_key_algorithm( &attributes ), alg );
TEST_EQUAL( psa_get_key_type( &attributes ), type );
Set the key size as an attribute Instead of passing a separate parameter for the key size to psa_generate_key and psa_generator_import_key, set it through the attributes, like the key type and other metadata.
2019-04-26 13:49:28 +02:00
TEST_EQUAL( psa_get_key_bits( &attributes ), bits );
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
psa_reset_key_attributes( &attributes );
TEST_EQUAL( psa_get_key_id( &attributes ), 0 );
TEST_EQUAL( psa_get_key_lifetime( &attributes ), 0 );
TEST_EQUAL( psa_get_key_usage_flags( &attributes ), 0 );
TEST_EQUAL( psa_get_key_algorithm( &attributes ), 0 );
TEST_EQUAL( psa_get_key_type( &attributes ), 0 );
Set the key size as an attribute Instead of passing a separate parameter for the key size to psa_generate_key and psa_generator_import_key, set it through the attributes, like the key type and other metadata.
2019-04-26 13:49:28 +02:00
TEST_EQUAL( psa_get_key_bits( &attributes ), 0 );
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
}
/* END_CASE */
Add a few tests for persistent attributes psa_set_key_lifetime and psa_set_key_id aren't pure setters: they also set the other attribute in some conditions. Add dedicated tests for this behavior.
2019-05-15 16:14:57 +02:00
/* BEGIN_CASE */
void persistence_attributes( int id1_arg, int lifetime_arg, int id2_arg,
int expected_id_arg, int expected_lifetime_arg )
{
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
psa_key_id_t id1 = id1_arg;
psa_key_lifetime_t lifetime = lifetime_arg;
psa_key_id_t id2 = id2_arg;
psa_key_id_t expected_id = expected_id_arg;
psa_key_lifetime_t expected_lifetime = expected_lifetime_arg;
if( id1_arg != -1 )
psa_set_key_id( &attributes, id1 );
if( lifetime_arg != -1 )
psa_set_key_lifetime( &attributes, lifetime );
if( id2_arg != -1 )
psa_set_key_id( &attributes, id2 );
TEST_EQUAL( psa_get_key_id( &attributes ), expected_id );
TEST_EQUAL( psa_get_key_lifetime( &attributes ), expected_lifetime );
}
/* END_CASE */
Test slot_number attribute Test the behavior of the getter/setter functions. Test that psa_get_key_slot_number() reports a slot number for a key in a secure element, and doesn't report a slot number for a key that is not in a secure element. Test that psa_get_key_slot_number() reports the correct slot number for a key in a secure element.
2019-08-02 20:30:01 +02:00
/* BEGIN_CASE depends_on:MBEDTLS_PSA_CRYPTO_SE_C */
void slot_number_attribute( )
{
psa_key_slot_number_t slot_number = 0xdeadbeef;
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
/* Initially, there is no slot number. */
TEST_EQUAL( psa_get_key_slot_number( &attributes, &slot_number ),
PSA_ERROR_INVALID_ARGUMENT );
/* Test setting a slot number. */
psa_set_key_slot_number( &attributes, 0 );
PSA_ASSERT( psa_get_key_slot_number( &attributes, &slot_number ) );
TEST_EQUAL( slot_number, 0 );
/* Test changing the slot number. */
psa_set_key_slot_number( &attributes, 42 );
PSA_ASSERT( psa_get_key_slot_number( &attributes, &slot_number ) );
TEST_EQUAL( slot_number, 42 );
/* Test clearing the slot number. */
psa_clear_key_slot_number( &attributes );
TEST_EQUAL( psa_get_key_slot_number( &attributes, &slot_number ),
PSA_ERROR_INVALID_ARGUMENT );
/* Clearing again should have no effect. */
psa_clear_key_slot_number( &attributes );
TEST_EQUAL( psa_get_key_slot_number( &attributes, &slot_number ),
PSA_ERROR_INVALID_ARGUMENT );
/* Test that reset clears the slot number. */
psa_set_key_slot_number( &attributes, 42 );
PSA_ASSERT( psa_get_key_slot_number( &attributes, &slot_number ) );
TEST_EQUAL( slot_number, 42 );
psa_reset_key_attributes( &attributes );
TEST_EQUAL( psa_get_key_slot_number( &attributes, &slot_number ),
PSA_ERROR_INVALID_ARGUMENT );
}
/* END_CASE */
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
/* BEGIN_CASE */
Add test function for import with a bad policy
2019-07-31 15:53:45 +02:00
void import_with_policy( int type_arg,
int usage_arg, int alg_arg,
int expected_status_arg )
{
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
psa_key_attributes_t got_attributes = PSA_KEY_ATTRIBUTES_INIT;
psa_key_handle_t handle = 0;
psa_key_type_t type = type_arg;
psa_key_usage_t usage = usage_arg;
psa_algorithm_t alg = alg_arg;
psa_status_t expected_status = expected_status_arg;
const uint8_t key_material[16] = {0};
psa_status_t status;
PSA_ASSERT( psa_crypto_init( ) );
psa_set_key_type( &attributes, type );
psa_set_key_usage_flags( &attributes, usage );
psa_set_key_algorithm( &attributes, alg );
status = psa_import_key( &attributes,
key_material, sizeof( key_material ),
&handle );
TEST_EQUAL( status, expected_status );
if( status != PSA_SUCCESS )
goto exit;
PSA_ASSERT( psa_get_key_attributes( handle, &got_attributes ) );
TEST_EQUAL( psa_get_key_type( &got_attributes ), type );
TEST_EQUAL( psa_get_key_usage_flags( &got_attributes ), usage );
TEST_EQUAL( psa_get_key_algorithm( &got_attributes ), alg );
Test slot_number attribute Test the behavior of the getter/setter functions. Test that psa_get_key_slot_number() reports a slot number for a key in a secure element, and doesn't report a slot number for a key that is not in a secure element. Test that psa_get_key_slot_number() reports the correct slot number for a key in a secure element.
2019-08-02 20:30:01 +02:00
ASSERT_NO_SLOT_NUMBER( &got_attributes );
Add test function for import with a bad policy
2019-07-31 15:53:45 +02:00
PSA_ASSERT( psa_destroy_key( handle ) );
test_operations_on_invalid_handle( handle );
exit:
psa_destroy_key( handle );
psa_reset_key_attributes( &got_attributes );
PSA_DONE( );
}
/* END_CASE */
/* BEGIN_CASE */
void import_with_data( data_t *data, int type_arg,
int attr_bits_arg,
int expected_status_arg )
Add key management functions Define psa_key_type_t and a first stab at a few values. New functions psa_import_key, psa_export_key, psa_destroy_key, psa_get_key_information. Implement them for raw data and RSA. Under the hood, create an in-memory, fixed-size keystore with room for MBEDTLS_PSA_KEY_SLOT_COUNT - 1 keys.
2018-01-28 13:16:24 +01:00
{
Implement atomic-creation psa_import_key Implement the new, attribute-based psa_import_key and some basic functions to access psa_key_attributes_t. Replace psa_import_key_to_handle by psa_import_key in a few test functions. This commit does not handle persistence attributes yet.
2019-04-17 15:05:45 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
psa_key_attributes_t got_attributes = PSA_KEY_ATTRIBUTES_INIT;
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t handle = 0;
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
psa_key_type_t type = type_arg;
Test psa_import_key: test for size in attributes Add tests where psa_import_key is called with attributes specifying an incorrect size.
2019-05-03 16:59:21 +02:00
size_t attr_bits = attr_bits_arg;
Get rid of some casts in test_suite_psa_crypto Use more auxiliary variables to unmarshall int values.
2018-06-21 09:25:10 +02:00
psa_status_t expected_status = expected_status_arg;
Add key management functions Define psa_key_type_t and a first stab at a few values. New functions psa_import_key, psa_export_key, psa_destroy_key, psa_get_key_information. Implement them for raw data and RSA. Under the hood, create an in-memory, fixed-size keystore with room for MBEDTLS_PSA_KEY_SLOT_COUNT - 1 keys.
2018-01-28 13:16:24 +01:00
psa_status_t status;
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
Add key management functions Define psa_key_type_t and a first stab at a few values. New functions psa_import_key, psa_export_key, psa_destroy_key, psa_get_key_information. Implement them for raw data and RSA. Under the hood, create an in-memory, fixed-size keystore with room for MBEDTLS_PSA_KEY_SLOT_COUNT - 1 keys.
2018-01-28 13:16:24 +01:00
Implement atomic-creation psa_import_key Implement the new, attribute-based psa_import_key and some basic functions to access psa_key_attributes_t. Replace psa_import_key_to_handle by psa_import_key in a few test functions. This commit does not handle persistence attributes yet.
2019-04-17 15:05:45 +02:00
psa_set_key_type( &attributes, type );
Test psa_import_key: test for size in attributes Add tests where psa_import_key is called with attributes specifying an incorrect size.
2019-05-03 16:59:21 +02:00
psa_set_key_bits( &attributes, attr_bits );
Add test function for import with a bad policy
2019-07-31 15:53:45 +02:00
Put handle parameter last: psa_import_key In psa_import_key, change the order of parameters to pass the pointer where the newly created handle will be stored last. This is consistent with most other library functions that put inputs before outputs.
2019-05-15 20:15:10 +02:00
status = psa_import_key( &attributes, data->x, data->len, &handle );
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( status, expected_status );
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
if( status != PSA_SUCCESS )
goto exit;
PSA_ASSERT( psa_get_key_attributes( handle, &got_attributes ) );
TEST_EQUAL( psa_get_key_type( &got_attributes ), type );
Test psa_import_key: test for size in attributes Add tests where psa_import_key is called with attributes specifying an incorrect size.
2019-05-03 16:59:21 +02:00
if( attr_bits != 0 )
Move attribute fields to a substructure Move the "core attributes" to a substructure of psa_key_attribute_t. The motivation is to be able to use the new structure psa_core_key_attributes_t internally.
2019-07-30 13:48:52 +02:00
TEST_EQUAL( attr_bits, psa_get_key_bits( &got_attributes ) );
Test slot_number attribute Test the behavior of the getter/setter functions. Test that psa_get_key_slot_number() reports a slot number for a key in a secure element, and doesn't report a slot number for a key that is not in a secure element. Test that psa_get_key_slot_number() reports the correct slot number for a key in a secure element.
2019-08-02 20:30:01 +02:00
ASSERT_NO_SLOT_NUMBER( &got_attributes );
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
PSA_ASSERT( psa_destroy_key( handle ) );
Simplify and expand invalid-handle tests Simplify invalid-handle tests and make them test more things. Call these tests in several test functions after destroying a key.
2019-04-18 22:28:52 +02:00
test_operations_on_invalid_handle( handle );
Add key management functions Define psa_key_type_t and a first stab at a few values. New functions psa_import_key, psa_export_key, psa_destroy_key, psa_get_key_information. Implement them for raw data and RSA. Under the hood, create an in-memory, fixed-size keystore with room for MBEDTLS_PSA_KEY_SLOT_COUNT - 1 keys.
2018-01-28 13:16:24 +01:00
exit:
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
psa_destroy_key( handle );
Call psa_reset_key_attributes after psa_get_key_attributes After calling psa_get_key_attributes(), call psa_reset_key_attributes() if the key may have domain parameters, because that's the way to free the domain parameter substructure in the attribute structure. Keep not calling reset() in some places where the key can only be a symmetric key which doesn't have domain parameters.
2019-04-26 16:03:33 +02:00
psa_reset_key_attributes( &got_attributes );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
Add key management functions Define psa_key_type_t and a first stab at a few values. New functions psa_import_key, psa_export_key, psa_destroy_key, psa_get_key_information. Implement them for raw data and RSA. Under the hood, create an in-memory, fixed-size keystore with room for MBEDTLS_PSA_KEY_SLOT_COUNT - 1 keys.
2018-01-28 13:16:24 +01:00
}
/* END_CASE */
Limit keys to 65528 bits 65528 bits is more than any reasonable key until we start supporting post-quantum cryptography. This limit is chosen to allow bit-sizes to be stored in 16 bits, with 65535 left to indicate an invalid value. It's a whole number of bytes, which facilitates some calculations, in particular allowing a key of exactly PSA_CRYPTO_MAX_STORAGE_SIZE to be created but not one bit more. As a resource usage limit, this is arguably too large, but that's out of scope of the current commit. Test that key import, generation and derivation reject overly large sizes.
2019-07-30 17:26:54 +02:00
/* BEGIN_CASE */
void import_large_key( int type_arg, int byte_size_arg,
int expected_status_arg )
{
psa_key_type_t type = type_arg;
size_t byte_size = byte_size_arg;
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
psa_status_t expected_status = expected_status_arg;
psa_key_handle_t handle = 0;
psa_status_t status;
uint8_t *buffer = NULL;
size_t buffer_size = byte_size + 1;
size_t n;
/* It would be better to skip the test than fail it if the allocation
* fails, but the test framework doesn't support this yet. */
ASSERT_ALLOC( buffer, buffer_size );
memset( buffer, 'K', byte_size );
PSA_ASSERT( psa_crypto_init( ) );
/* Try importing the key */
psa_set_key_usage_flags( &attributes, PSA_KEY_USAGE_EXPORT );
psa_set_key_type( &attributes, type );
status = psa_import_key( &attributes, buffer, byte_size, &handle );
TEST_EQUAL( status, expected_status );
if( status == PSA_SUCCESS )
{
PSA_ASSERT( psa_get_key_attributes( handle, &attributes ) );
TEST_EQUAL( psa_get_key_type( &attributes ), type );
TEST_EQUAL( psa_get_key_bits( &attributes ),
PSA_BYTES_TO_BITS( byte_size ) );
Test slot_number attribute Test the behavior of the getter/setter functions. Test that psa_get_key_slot_number() reports a slot number for a key in a secure element, and doesn't report a slot number for a key that is not in a secure element. Test that psa_get_key_slot_number() reports the correct slot number for a key in a secure element.
2019-08-02 20:30:01 +02:00
ASSERT_NO_SLOT_NUMBER( &attributes );
Limit keys to 65528 bits 65528 bits is more than any reasonable key until we start supporting post-quantum cryptography. This limit is chosen to allow bit-sizes to be stored in 16 bits, with 65535 left to indicate an invalid value. It's a whole number of bytes, which facilitates some calculations, in particular allowing a key of exactly PSA_CRYPTO_MAX_STORAGE_SIZE to be created but not one bit more. As a resource usage limit, this is arguably too large, but that's out of scope of the current commit. Test that key import, generation and derivation reject overly large sizes.
2019-07-30 17:26:54 +02:00
memset( buffer, 0, byte_size + 1 );
PSA_ASSERT( psa_export_key( handle, buffer, byte_size, &n ) );
for( n = 0; n < byte_size; n++ )
TEST_EQUAL( buffer[n], 'K' );
for( n = byte_size; n < buffer_size; n++ )
TEST_EQUAL( buffer[n], 0 );
}
exit:
psa_destroy_key( handle );
PSA_DONE( );
mbedtls_free( buffer );
}
/* END_CASE */
Test that creating RSA keys larger than the maximum fails Test keypair import, public key import and key generation.
2018-06-28 00:16:11 +02:00
/* BEGIN_CASE */
void import_rsa_made_up( int bits_arg, int keypair, int expected_status_arg )
{
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t handle = 0;
Test that creating RSA keys larger than the maximum fails Test keypair import, public key import and key generation.
2018-06-28 00:16:11 +02:00
size_t bits = bits_arg;
psa_status_t expected_status = expected_status_arg;
psa_status_t status;
psa_key_type_t type =
Rename *KEYPAIR* to *KEY_PAIR* Be consistent with PUBLIC_KEY. perl -i -pe 's/KEYPAIR/KEY_PAIR/g' $(git ls-files)
2019-05-16 19:39:54 +02:00
keypair ? PSA_KEY_TYPE_RSA_KEY_PAIR : PSA_KEY_TYPE_RSA_PUBLIC_KEY;
Test that creating RSA keys larger than the maximum fails Test keypair import, public key import and key generation.
2018-06-28 00:16:11 +02:00
size_t buffer_size = /* Slight overapproximations */
keypair ? bits * 9 / 16 + 80 : bits / 8 + 20;
Use ASSERT_ALLOC instead of mbedtls_calloc in PSA tests This commit resolves a bug whereby some test cases failed on systems where mbedtls_calloc returns NULL when the size of 0, because the test case asserted `pointer != NULL` regardless of the size.
2018-09-27 13:54:18 +02:00
unsigned char *buffer = NULL;
Test that creating RSA keys larger than the maximum fails Test keypair import, public key import and key generation.
2018-06-28 00:16:11 +02:00
unsigned char *p;
int ret;
size_t length;
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
Test that creating RSA keys larger than the maximum fails Test keypair import, public key import and key generation.
2018-06-28 00:16:11 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
Use ASSERT_ALLOC instead of mbedtls_calloc in PSA tests This commit resolves a bug whereby some test cases failed on systems where mbedtls_calloc returns NULL when the size of 0, because the test case asserted `pointer != NULL` regardless of the size.
2018-09-27 13:54:18 +02:00
ASSERT_ALLOC( buffer, buffer_size );
Test that creating RSA keys larger than the maximum fails Test keypair import, public key import and key generation.
2018-06-28 00:16:11 +02:00
TEST_ASSERT( ( ret = construct_fake_rsa_key( buffer, buffer_size, &p,
bits, keypair ) ) >= 0 );
length = ret;
/* Try importing the key */
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_set_key_type( &attributes, type );
Put handle parameter last: psa_import_key In psa_import_key, change the order of parameters to pass the pointer where the newly created handle will be stored last. This is consistent with most other library functions that put inputs before outputs.
2019-05-15 20:15:10 +02:00
status = psa_import_key( &attributes, p, length, &handle );
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( status, expected_status );
Close or destroy keys explicitly in tests
2019-05-28 14:08:50 +02:00
Test that creating RSA keys larger than the maximum fails Test keypair import, public key import and key generation.
2018-06-28 00:16:11 +02:00
if( status == PSA_SUCCESS )
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_destroy_key( handle ) );
Test that creating RSA keys larger than the maximum fails Test keypair import, public key import and key generation.
2018-06-28 00:16:11 +02:00
exit:
mbedtls_free( buffer );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
Test that creating RSA keys larger than the maximum fails Test keypair import, public key import and key generation.
2018-06-28 00:16:11 +02:00
}
/* END_CASE */
Add key management functions Define psa_key_type_t and a first stab at a few values. New functions psa_import_key, psa_export_key, psa_destroy_key, psa_get_key_information. Implement them for raw data and RSA. Under the hood, create an in-memory, fixed-size keystore with room for MBEDTLS_PSA_KEY_SLOT_COUNT - 1 keys.
2018-01-28 13:16:24 +01:00
/* BEGIN_CASE */
On target testing tests adaptation Updated all psa crypto tests to use the new test format
2018-06-12 16:06:52 +02:00
void import_export( data_t *data,
add non-regression tests for export public/non public key
2018-06-04 17:42:36 +02:00
int type_arg,
Align test functions to usage/alg parameter order Manually cherry-picked from ca5bed742fd1a886284baca48b2b7574d875630c by taking that patch, replacing KEYPAIR by KEY_PAIR throughout (renaming applied in this branch), and discarding parts about import_twice in test_suite_psa_crypto (this test function was removed from this branch).
2019-05-24 15:00:06 +02:00
int usage_arg, int alg_arg,
Add key management functions Define psa_key_type_t and a first stab at a few values. New functions psa_import_key, psa_export_key, psa_destroy_key, psa_get_key_information. Implement them for raw data and RSA. Under the hood, create an in-memory, fixed-size keystore with room for MBEDTLS_PSA_KEY_SLOT_COUNT - 1 keys.
2018-01-28 13:16:24 +01:00
int expected_bits,
int export_size_delta,
Get rid of some casts in test_suite_psa_crypto Use more auxiliary variables to unmarshall int values.
2018-06-21 09:25:10 +02:00
int expected_export_status_arg,
Add key management functions Define psa_key_type_t and a first stab at a few values. New functions psa_import_key, psa_export_key, psa_destroy_key, psa_get_key_information. Implement them for raw data and RSA. Under the hood, create an in-memory, fixed-size keystore with room for MBEDTLS_PSA_KEY_SLOT_COUNT - 1 keys.
2018-01-28 13:16:24 +01:00
int canonical_input )
{
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t handle = 0;
Add key management functions Define psa_key_type_t and a first stab at a few values. New functions psa_import_key, psa_export_key, psa_destroy_key, psa_get_key_information. Implement them for raw data and RSA. Under the hood, create an in-memory, fixed-size keystore with room for MBEDTLS_PSA_KEY_SLOT_COUNT - 1 keys.
2018-01-28 13:16:24 +01:00
psa_key_type_t type = type_arg;
Hygiene improvements in PSA crypto test code Get rid of many redundant casts. In particular, it is not useful to cast uint32_t values to size_t before performing arithmetic or comparisons on them. Rewrap a number of function calls, many of which now have narrower arguments thanks to the removed casts. When a function call doesn't fit on a single line, avoid grouping unrelated parameters together, but do try to group a buffer pointer and the associated size. Define more auxiliary variables xxx of a particular integer type (psa_algorithm_t, psa_key_usage_t, etc.) corresponding to a test function xxx_arg which has the type int. This avoids the need to cast xxx_arg to an unsigned type sometimes in the code.
2018-06-18 16:35:34 +02:00
psa_algorithm_t alg = alg_arg;
Get rid of some casts in test_suite_psa_crypto Use more auxiliary variables to unmarshall int values.
2018-06-21 09:25:10 +02:00
psa_status_t expected_export_status = expected_export_status_arg;
Add key management functions Define psa_key_type_t and a first stab at a few values. New functions psa_import_key, psa_export_key, psa_destroy_key, psa_get_key_information. Implement them for raw data and RSA. Under the hood, create an in-memory, fixed-size keystore with room for MBEDTLS_PSA_KEY_SLOT_COUNT - 1 keys.
2018-01-28 13:16:24 +01:00
psa_status_t status;
unsigned char *exported = NULL;
unsigned char *reexported = NULL;
size_t export_size;
psa_export_key: Always set a valid data length Make psa_export_key() always set a valid data_length when exporting, even when there are errors. This makes the API easier to use for buggy programs (like our test code). Our test code previously used exported_length uninitialized when checking to see that the buffer returned was all zero in import_export() in the case where an error was returned from psa_export_key(). Initialize exported_length to an invalid length, and check that it gets set properly by psa_export_key(), to avoid this using export_length uninitialized. Note that the mem_is_zero() check is still valid when psa_export_key() returns an error, e.g. where exported_length is 0, as we want to check that nothing was written to the buffer on error. Out test code also previous passed NULL for the data_length parameter of psa_export_key() when it expected a failure (in key_policy_fail()). However, data_length is not allowed to be NULL, especially now that we write to data_length from psa_export_key() even when there are errors. Update the test code to not pass in a NULL data_length.
2018-06-27 18:20:43 +02:00
size_t exported_length = INVALID_EXPORT_LENGTH;
Add key management functions Define psa_key_type_t and a first stab at a few values. New functions psa_import_key, psa_export_key, psa_destroy_key, psa_get_key_information. Implement them for raw data and RSA. Under the hood, create an in-memory, fixed-size keystore with room for MBEDTLS_PSA_KEY_SLOT_COUNT - 1 keys.
2018-01-28 13:16:24 +01:00
size_t reexported_length;
Implement atomic-creation psa_import_key Implement the new, attribute-based psa_import_key and some basic functions to access psa_key_attributes_t. Replace psa_import_key_to_handle by psa_import_key in a few test functions. This commit does not handle persistence attributes yet.
2019-04-17 15:05:45 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
psa_key_attributes_t got_attributes = PSA_KEY_ATTRIBUTES_INIT;
Add key management functions Define psa_key_type_t and a first stab at a few values. New functions psa_import_key, psa_export_key, psa_destroy_key, psa_get_key_information. Implement them for raw data and RSA. Under the hood, create an in-memory, fixed-size keystore with room for MBEDTLS_PSA_KEY_SLOT_COUNT - 1 keys.
2018-01-28 13:16:24 +01:00
Replace ssize_t by ptrdiff_t ssize_t is a POSIX thing, not standard C
2018-07-17 16:36:59 +02:00
export_size = (ptrdiff_t) data->len + export_size_delta;
Use ASSERT_ALLOC instead of mbedtls_calloc in PSA tests This commit resolves a bug whereby some test cases failed on systems where mbedtls_calloc returns NULL when the size of 0, because the test case asserted `pointer != NULL` regardless of the size.
2018-09-27 13:54:18 +02:00
ASSERT_ALLOC( exported, export_size );
Add key management functions Define psa_key_type_t and a first stab at a few values. New functions psa_import_key, psa_export_key, psa_destroy_key, psa_get_key_information. Implement them for raw data and RSA. Under the hood, create an in-memory, fixed-size keystore with room for MBEDTLS_PSA_KEY_SLOT_COUNT - 1 keys.
2018-01-28 13:16:24 +01:00
if( ! canonical_input )
Use ASSERT_ALLOC instead of mbedtls_calloc in PSA tests This commit resolves a bug whereby some test cases failed on systems where mbedtls_calloc returns NULL when the size of 0, because the test case asserted `pointer != NULL` regardless of the size.
2018-09-27 13:54:18 +02:00
ASSERT_ALLOC( reexported, export_size );
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
Add key management functions Define psa_key_type_t and a first stab at a few values. New functions psa_import_key, psa_export_key, psa_destroy_key, psa_get_key_information. Implement them for raw data and RSA. Under the hood, create an in-memory, fixed-size keystore with room for MBEDTLS_PSA_KEY_SLOT_COUNT - 1 keys.
2018-01-28 13:16:24 +01:00
Implement atomic-creation psa_import_key Implement the new, attribute-based psa_import_key and some basic functions to access psa_key_attributes_t. Replace psa_import_key_to_handle by psa_import_key in a few test functions. This commit does not handle persistence attributes yet.
2019-04-17 15:05:45 +02:00
psa_set_key_usage_flags( &attributes, usage_arg );
psa_set_key_algorithm( &attributes, alg );
psa_set_key_type( &attributes, type );
Add calls for set policy in export/sign tests Add calls for set policy in export/sign tests
2018-03-28 12:46:26 +02:00
Add key management functions Define psa_key_type_t and a first stab at a few values. New functions psa_import_key, psa_export_key, psa_destroy_key, psa_get_key_information. Implement them for raw data and RSA. Under the hood, create an in-memory, fixed-size keystore with room for MBEDTLS_PSA_KEY_SLOT_COUNT - 1 keys.
2018-01-28 13:16:24 +01:00
/* Import the key */
Put handle parameter last: psa_import_key In psa_import_key, change the order of parameters to pass the pointer where the newly created handle will be stored last. This is consistent with most other library functions that put inputs before outputs.
2019-05-15 20:15:10 +02:00
PSA_ASSERT( psa_import_key( &attributes, data->x, data->len, &handle ) );
Add key management functions Define psa_key_type_t and a first stab at a few values. New functions psa_import_key, psa_export_key, psa_destroy_key, psa_get_key_information. Implement them for raw data and RSA. Under the hood, create an in-memory, fixed-size keystore with room for MBEDTLS_PSA_KEY_SLOT_COUNT - 1 keys.
2018-01-28 13:16:24 +01:00
/* Test the key information */
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
PSA_ASSERT( psa_get_key_attributes( handle, &got_attributes ) );
TEST_EQUAL( psa_get_key_type( &got_attributes ), type );
TEST_EQUAL( psa_get_key_bits( &got_attributes ), (size_t) expected_bits );
Test slot_number attribute Test the behavior of the getter/setter functions. Test that psa_get_key_slot_number() reports a slot number for a key in a secure element, and doesn't report a slot number for a key that is not in a secure element. Test that psa_get_key_slot_number() reports the correct slot number for a key in a secure element.
2019-08-02 20:30:01 +02:00
ASSERT_NO_SLOT_NUMBER( &got_attributes );
Add key management functions Define psa_key_type_t and a first stab at a few values. New functions psa_import_key, psa_export_key, psa_destroy_key, psa_get_key_information. Implement them for raw data and RSA. Under the hood, create an in-memory, fixed-size keystore with room for MBEDTLS_PSA_KEY_SLOT_COUNT - 1 keys.
2018-01-28 13:16:24 +01:00
/* Export the key */
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
status = psa_export_key( handle,
Add key management functions Define psa_key_type_t and a first stab at a few values. New functions psa_import_key, psa_export_key, psa_destroy_key, psa_get_key_information. Implement them for raw data and RSA. Under the hood, create an in-memory, fixed-size keystore with room for MBEDTLS_PSA_KEY_SLOT_COUNT - 1 keys.
2018-01-28 13:16:24 +01:00
exported, export_size,
&exported_length );
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( status, expected_export_status );
psa_export_key: Always set a valid data length Make psa_export_key() always set a valid data_length when exporting, even when there are errors. This makes the API easier to use for buggy programs (like our test code). Our test code previously used exported_length uninitialized when checking to see that the buffer returned was all zero in import_export() in the case where an error was returned from psa_export_key(). Initialize exported_length to an invalid length, and check that it gets set properly by psa_export_key(), to avoid this using export_length uninitialized. Note that the mem_is_zero() check is still valid when psa_export_key() returns an error, e.g. where exported_length is 0, as we want to check that nothing was written to the buffer on error. Out test code also previous passed NULL for the data_length parameter of psa_export_key() when it expected a failure (in key_policy_fail()). However, data_length is not allowed to be NULL, especially now that we write to data_length from psa_export_key() even when there are errors. Update the test code to not pass in a NULL data_length.
2018-06-27 18:20:43 +02:00
/* The exported length must be set by psa_export_key() to a value between 0
* and export_size. On errors, the exported length must be 0. */
TEST_ASSERT( exported_length != INVALID_EXPORT_LENGTH );
TEST_ASSERT( status == PSA_SUCCESS || exported_length == 0 );
TEST_ASSERT( exported_length <= export_size );
Add tests for mac_sign
2018-08-14 15:17:54 +02:00
TEST_ASSERT( mem_is_char( exported + exported_length, 0,
Simplify mem_is_nonzero to mem_is_zero This also fixes a bug that the value that mem_is_nonzero tried to return could overflow int.
2018-06-21 09:21:51 +02:00
export_size - exported_length ) );
Add key management functions Define psa_key_type_t and a first stab at a few values. New functions psa_import_key, psa_export_key, psa_destroy_key, psa_get_key_information. Implement them for raw data and RSA. Under the hood, create an in-memory, fixed-size keystore with room for MBEDTLS_PSA_KEY_SLOT_COUNT - 1 keys.
2018-01-28 13:16:24 +01:00
if( status != PSA_SUCCESS )
psa_export_key: zero out potential garbage in the output buffer In psa_export_key, ensure that each byte of the output buffer either contains its original value, is zero, or is part of the actual output. Specifically, don't risk having partial output on error, and don't leave extra data at the end of the buffer when exporting an asymmetric key. Test that exporting to a previously zeroed buffer leaves the buffer zeroed outside the actual output if any.
2018-06-20 00:11:45 +02:00
{
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( exported_length, 0 );
Add key management functions Define psa_key_type_t and a first stab at a few values. New functions psa_import_key, psa_export_key, psa_destroy_key, psa_get_key_information. Implement them for raw data and RSA. Under the hood, create an in-memory, fixed-size keystore with room for MBEDTLS_PSA_KEY_SLOT_COUNT - 1 keys.
2018-01-28 13:16:24 +01:00
goto destroy;
psa_export_key: zero out potential garbage in the output buffer In psa_export_key, ensure that each byte of the output buffer either contains its original value, is zero, or is part of the actual output. Specifically, don't risk having partial output on error, and don't leave extra data at the end of the buffer when exporting an asymmetric key. Test that exporting to a previously zeroed buffer leaves the buffer zeroed outside the actual output if any.
2018-06-20 00:11:45 +02:00
}
Add key management functions Define psa_key_type_t and a first stab at a few values. New functions psa_import_key, psa_export_key, psa_destroy_key, psa_get_key_information. Implement them for raw data and RSA. Under the hood, create an in-memory, fixed-size keystore with room for MBEDTLS_PSA_KEY_SLOT_COUNT - 1 keys.
2018-01-28 13:16:24 +01:00
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
if( ! exercise_export_key( handle, usage_arg ) )
Do export sanity checks in import_export as well This is not useful to validate the implementation when importing canonical input, which is the case for most import/export test cases, but it helps validate the sanity checks themselves.
2018-08-11 01:24:55 +02:00
goto exit;
Add key management functions Define psa_key_type_t and a first stab at a few values. New functions psa_import_key, psa_export_key, psa_destroy_key, psa_get_key_information. Implement them for raw data and RSA. Under the hood, create an in-memory, fixed-size keystore with room for MBEDTLS_PSA_KEY_SLOT_COUNT - 1 keys.
2018-01-28 13:16:24 +01:00
if( canonical_input )
Use ASSERT_COMPARE instead of memcmp in PSA tests This commit fixes some missing size comparison. In aead_encrypt_decrypt, aead_encrypt and aead_decrypt, the test code would not have noticed if the library function had reported an output length that was not the expected length.
2018-09-27 13:57:19 +02:00
ASSERT_COMPARE( data->x, data->len, exported, exported_length );
Add key management functions Define psa_key_type_t and a first stab at a few values. New functions psa_import_key, psa_export_key, psa_destroy_key, psa_get_key_information. Implement them for raw data and RSA. Under the hood, create an in-memory, fixed-size keystore with room for MBEDTLS_PSA_KEY_SLOT_COUNT - 1 keys.
2018-01-28 13:16:24 +01:00
else
{
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t handle2;
Split long lines after psa_import_key refactoring
2019-05-15 20:22:09 +02:00
PSA_ASSERT( psa_import_key( &attributes, exported, exported_length,
&handle2 ) );
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_export_key( handle2,
reexported,
export_size,
&reexported_length ) );
Use ASSERT_COMPARE instead of memcmp in PSA tests This commit fixes some missing size comparison. In aead_encrypt_decrypt, aead_encrypt and aead_decrypt, the test code would not have noticed if the library function had reported an output length that was not the expected length.
2018-09-27 13:57:19 +02:00
ASSERT_COMPARE( exported, exported_length,
reexported, reexported_length );
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_close_key( handle2 ) );
Add key management functions Define psa_key_type_t and a first stab at a few values. New functions psa_import_key, psa_export_key, psa_destroy_key, psa_get_key_information. Implement them for raw data and RSA. Under the hood, create an in-memory, fixed-size keystore with room for MBEDTLS_PSA_KEY_SLOT_COUNT - 1 keys.
2018-01-28 13:16:24 +01:00
}
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
TEST_ASSERT( exported_length <= PSA_KEY_EXPORT_MAX_SIZE( type, psa_get_key_bits( &got_attributes ) ) );
Add key management functions Define psa_key_type_t and a first stab at a few values. New functions psa_import_key, psa_export_key, psa_destroy_key, psa_get_key_information. Implement them for raw data and RSA. Under the hood, create an in-memory, fixed-size keystore with room for MBEDTLS_PSA_KEY_SLOT_COUNT - 1 keys.
2018-01-28 13:16:24 +01:00
destroy:
/* Destroy the key */
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_destroy_key( handle ) );
Simplify and expand invalid-handle tests Simplify invalid-handle tests and make them test more things. Call these tests in several test functions after destroying a key.
2019-04-18 22:28:52 +02:00
test_operations_on_invalid_handle( handle );
Add key management functions Define psa_key_type_t and a first stab at a few values. New functions psa_import_key, psa_export_key, psa_destroy_key, psa_get_key_information. Implement them for raw data and RSA. Under the hood, create an in-memory, fixed-size keystore with room for MBEDTLS_PSA_KEY_SLOT_COUNT - 1 keys.
2018-01-28 13:16:24 +01:00
exit:
On target testing tests adaptation Updated all psa crypto tests to use the new test format
2018-06-12 16:06:52 +02:00
mbedtls_free( exported );
mbedtls_free( reexported );
Call psa_reset_key_attributes after psa_get_key_attributes After calling psa_get_key_attributes(), call psa_reset_key_attributes() if the key may have domain parameters, because that's the way to free the domain parameter substructure in the attribute structure. Keep not calling reset() in some places where the key can only be a symmetric key which doesn't have domain parameters.
2019-04-26 16:03:33 +02:00
psa_reset_key_attributes( &got_attributes );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
Add key management functions Define psa_key_type_t and a first stab at a few values. New functions psa_import_key, psa_export_key, psa_destroy_key, psa_get_key_information. Implement them for raw data and RSA. Under the hood, create an in-memory, fixed-size keystore with room for MBEDTLS_PSA_KEY_SLOT_COUNT - 1 keys.
2018-01-28 13:16:24 +01:00
}
/* END_CASE */
PSA crypto: asymmetric signature (RSA PKCS#1v1.5 only) Define hash algorithms and RSA signature algorithms. New function psa_asymmetric_sign. Implement psa_asymmetric_sign for RSA PKCS#1 v1.5.
2018-02-03 22:44:14 +01:00
move import_export_public_key func place
2018-06-06 16:26:04 +02:00
/* BEGIN_CASE */
On target testing tests adaptation Updated all psa crypto tests to use the new test format
2018-06-12 16:06:52 +02:00
void import_export_public_key( data_t *data,
Normalize whitespace Normalize whitespace to Mbed TLS standards. There are only whitespace changes in this commit.
2018-06-18 15:41:12 +02:00
int type_arg,
int alg_arg,
Improve export_public_key test function In the test function for export_public_key, don't just check the length of the result. Compare the actual result to the expected result. Take an extra argument that allows using an export buffer that's larger or smaller than needed. Zero is the size given by PSA_KEY_EXPORT_MAX_SIZE. Don't check the output of psa_get_key_information. That's useful in import_export because it tests both import and export, but not in import_export_public_key whose goal is only to test public key export. This commit adjusts the existing test data but does not add new test cases.
2018-10-29 15:15:31 +01:00
int export_size_delta,
int expected_export_status_arg,
data_t *expected_public_key )
move import_export_public_key func place
2018-06-06 16:26:04 +02:00
{
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t handle = 0;
move import_export_public_key func place
2018-06-06 16:26:04 +02:00
psa_key_type_t type = type_arg;
Hygiene improvements in PSA crypto test code Get rid of many redundant casts. In particular, it is not useful to cast uint32_t values to size_t before performing arithmetic or comparisons on them. Rewrap a number of function calls, many of which now have narrower arguments thanks to the removed casts. When a function call doesn't fit on a single line, avoid grouping unrelated parameters together, but do try to group a buffer pointer and the associated size. Define more auxiliary variables xxx of a particular integer type (psa_algorithm_t, psa_key_usage_t, etc.) corresponding to a test function xxx_arg which has the type int. This avoids the need to cast xxx_arg to an unsigned type sometimes in the code.
2018-06-18 16:35:34 +02:00
psa_algorithm_t alg = alg_arg;
Get rid of some casts in test_suite_psa_crypto Use more auxiliary variables to unmarshall int values.
2018-06-21 09:25:10 +02:00
psa_status_t expected_export_status = expected_export_status_arg;
move import_export_public_key func place
2018-06-06 16:26:04 +02:00
psa_status_t status;
unsigned char *exported = NULL;
Improve export_public_key test function In the test function for export_public_key, don't just check the length of the result. Compare the actual result to the expected result. Take an extra argument that allows using an export buffer that's larger or smaller than needed. Zero is the size given by PSA_KEY_EXPORT_MAX_SIZE. Don't check the output of psa_get_key_information. That's useful in import_export because it tests both import and export, but not in import_export_public_key whose goal is only to test public key export. This commit adjusts the existing test data but does not add new test cases.
2018-10-29 15:15:31 +01:00
size_t export_size = expected_public_key->len + export_size_delta;
psa: export_public_key: Check for all zero on error
2018-06-27 18:47:40 +02:00
size_t exported_length = INVALID_EXPORT_LENGTH;
Implement atomic-creation psa_import_key Implement the new, attribute-based psa_import_key and some basic functions to access psa_key_attributes_t. Replace psa_import_key_to_handle by psa_import_key in a few test functions. This commit does not handle persistence attributes yet.
2019-04-17 15:05:45 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
move import_export_public_key func place
2018-06-06 16:26:04 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
move import_export_public_key func place
2018-06-06 16:26:04 +02:00
Implement atomic-creation psa_import_key Implement the new, attribute-based psa_import_key and some basic functions to access psa_key_attributes_t. Replace psa_import_key_to_handle by psa_import_key in a few test functions. This commit does not handle persistence attributes yet.
2019-04-17 15:05:45 +02:00
psa_set_key_usage_flags( &attributes, PSA_KEY_USAGE_EXPORT );
psa_set_key_algorithm( &attributes, alg );
psa_set_key_type( &attributes, type );
move import_export_public_key func place
2018-06-06 16:26:04 +02:00
/* Import the key */
Put handle parameter last: psa_import_key In psa_import_key, change the order of parameters to pass the pointer where the newly created handle will be stored last. This is consistent with most other library functions that put inputs before outputs.
2019-05-15 20:15:10 +02:00
PSA_ASSERT( psa_import_key( &attributes, data->x, data->len, &handle ) );
move import_export_public_key func place
2018-06-06 16:26:04 +02:00
Improve export_public_key test function In the test function for export_public_key, don't just check the length of the result. Compare the actual result to the expected result. Take an extra argument that allows using an export buffer that's larger or smaller than needed. Zero is the size given by PSA_KEY_EXPORT_MAX_SIZE. Don't check the output of psa_get_key_information. That's useful in import_export because it tests both import and export, but not in import_export_public_key whose goal is only to test public key export. This commit adjusts the existing test data but does not add new test cases.
2018-10-29 15:15:31 +01:00
/* Export the public key */
ASSERT_ALLOC( exported, export_size );
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
status = psa_export_public_key( handle,
Normalize whitespace Normalize whitespace to Mbed TLS standards. There are only whitespace changes in this commit.
2018-06-18 15:41:12 +02:00
exported, export_size,
&exported_length );
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( status, expected_export_status );
Improve export_public_key test function In the test function for export_public_key, don't just check the length of the result. Compare the actual result to the expected result. Take an extra argument that allows using an export buffer that's larger or smaller than needed. Zero is the size given by PSA_KEY_EXPORT_MAX_SIZE. Don't check the output of psa_get_key_information. That's useful in import_export because it tests both import and export, but not in import_export_public_key whose goal is only to test public key export. This commit adjusts the existing test data but does not add new test cases.
2018-10-29 15:15:31 +01:00
if( status == PSA_SUCCESS )
In export tests, also test PSA_KEY_EXPORT_MAX_SIZE When testing psa_export_key or psa_export_public_key, test that the expected result fits in the size given by PSA_KEY_EXPORT_MAX_SIZE.
2018-10-29 15:18:41 +01:00
{
Rename *KEYPAIR* to *KEY_PAIR* Be consistent with PUBLIC_KEY. perl -i -pe 's/KEYPAIR/KEY_PAIR/g' $(git ls-files)
2019-05-16 19:39:54 +02:00
psa_key_type_t public_type = PSA_KEY_TYPE_PUBLIC_KEY_OF_KEY_PAIR( type );
In export tests, also test PSA_KEY_EXPORT_MAX_SIZE When testing psa_export_key or psa_export_public_key, test that the expected result fits in the size given by PSA_KEY_EXPORT_MAX_SIZE.
2018-10-29 15:18:41 +01:00
size_t bits;
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
PSA_ASSERT( psa_get_key_attributes( handle, &attributes ) );
bits = psa_get_key_bits( &attributes );
In export tests, also test PSA_KEY_EXPORT_MAX_SIZE When testing psa_export_key or psa_export_public_key, test that the expected result fits in the size given by PSA_KEY_EXPORT_MAX_SIZE.
2018-10-29 15:18:41 +01:00
TEST_ASSERT( expected_public_key->len <=
PSA_KEY_EXPORT_MAX_SIZE( public_type, bits ) );
Improve export_public_key test function In the test function for export_public_key, don't just check the length of the result. Compare the actual result to the expected result. Take an extra argument that allows using an export buffer that's larger or smaller than needed. Zero is the size given by PSA_KEY_EXPORT_MAX_SIZE. Don't check the output of psa_get_key_information. That's useful in import_export because it tests both import and export, but not in import_export_public_key whose goal is only to test public key export. This commit adjusts the existing test data but does not add new test cases.
2018-10-29 15:15:31 +01:00
ASSERT_COMPARE( expected_public_key->x, expected_public_key->len,
exported, exported_length );
In export tests, also test PSA_KEY_EXPORT_MAX_SIZE When testing psa_export_key or psa_export_public_key, test that the expected result fits in the size given by PSA_KEY_EXPORT_MAX_SIZE.
2018-10-29 15:18:41 +01:00
}
move import_export_public_key func place
2018-06-06 16:26:04 +02:00
exit:
On target testing tests adaptation Updated all psa crypto tests to use the new test format
2018-06-12 16:06:52 +02:00
mbedtls_free( exported );
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_destroy_key( handle );
Call psa_reset_key_attributes after psa_get_key_attributes After calling psa_get_key_attributes(), call psa_reset_key_attributes() if the key may have domain parameters, because that's the way to free the domain parameter substructure in the attribute structure. Keep not calling reset() in some places where the key can only be a symmetric key which doesn't have domain parameters.
2019-04-26 16:03:33 +02:00
psa_reset_key_attributes( &attributes );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
move import_export_public_key func place
2018-06-06 16:26:04 +02:00
}
/* END_CASE */
Add import-and-exercise tests for some signature algorithms
2018-06-26 16:12:43 +02:00
/* BEGIN_CASE */
void import_and_exercise_key( data_t *data,
int type_arg,
int bits_arg,
int alg_arg )
{
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t handle = 0;
Add import-and-exercise tests for some signature algorithms
2018-06-26 16:12:43 +02:00
psa_key_type_t type = type_arg;
size_t bits = bits_arg;
psa_algorithm_t alg = alg_arg;
Factor usage_to_exercise into its own function
2018-10-25 22:35:43 +02:00
psa_key_usage_t usage = usage_to_exercise( type, alg );
Implement atomic-creation psa_import_key Implement the new, attribute-based psa_import_key and some basic functions to access psa_key_attributes_t. Replace psa_import_key_to_handle by psa_import_key in a few test functions. This commit does not handle persistence attributes yet.
2019-04-17 15:05:45 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
psa_key_attributes_t got_attributes = PSA_KEY_ATTRIBUTES_INIT;
Add import-and-exercise tests for some signature algorithms
2018-06-26 16:12:43 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
Add import-and-exercise tests for some signature algorithms
2018-06-26 16:12:43 +02:00
Implement atomic-creation psa_import_key Implement the new, attribute-based psa_import_key and some basic functions to access psa_key_attributes_t. Replace psa_import_key_to_handle by psa_import_key in a few test functions. This commit does not handle persistence attributes yet.
2019-04-17 15:05:45 +02:00
psa_set_key_usage_flags( &attributes, usage );
psa_set_key_algorithm( &attributes, alg );
psa_set_key_type( &attributes, type );
Add import-and-exercise tests for some signature algorithms
2018-06-26 16:12:43 +02:00
/* Import the key */
Put handle parameter last: psa_import_key In psa_import_key, change the order of parameters to pass the pointer where the newly created handle will be stored last. This is consistent with most other library functions that put inputs before outputs.
2019-05-15 20:15:10 +02:00
PSA_ASSERT( psa_import_key( &attributes, data->x, data->len, &handle ) );
Add import-and-exercise tests for some signature algorithms
2018-06-26 16:12:43 +02:00
/* Test the key information */
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
PSA_ASSERT( psa_get_key_attributes( handle, &got_attributes ) );
TEST_EQUAL( psa_get_key_type( &got_attributes ), type );
TEST_EQUAL( psa_get_key_bits( &got_attributes ), bits );
Add import-and-exercise tests for some signature algorithms
2018-06-26 16:12:43 +02:00
/* Do something with the key according to its type and permitted usage. */
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
if( ! exercise_key( handle, usage, alg ) )
Factor duplicated code into exercise_key Also fail the test if the test code lacks a way to exercise the key.
2018-07-01 22:31:34 +02:00
goto exit;
Add import-and-exercise tests for some signature algorithms
2018-06-26 16:12:43 +02:00
Simplify and expand invalid-handle tests Simplify invalid-handle tests and make them test more things. Call these tests in several test functions after destroying a key.
2019-04-18 22:28:52 +02:00
PSA_ASSERT( psa_destroy_key( handle ) );
test_operations_on_invalid_handle( handle );
Add import-and-exercise tests for some signature algorithms
2018-06-26 16:12:43 +02:00
exit:
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_destroy_key( handle );
Call psa_reset_key_attributes after psa_get_key_attributes After calling psa_get_key_attributes(), call psa_reset_key_attributes() if the key may have domain parameters, because that's the way to free the domain parameter substructure in the attribute structure. Keep not calling reset() in some places where the key can only be a symmetric key which doesn't have domain parameters.
2019-04-26 16:03:33 +02:00
psa_reset_key_attributes( &got_attributes );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
Add import-and-exercise tests for some signature algorithms
2018-06-26 16:12:43 +02:00
}
/* END_CASE */
Reorder PSA test cases to group them by topic * init-deinit * import-export * policies * lifetime * hash * MAC * cipher * AEAD * asymmetric sign * asymmetric verify * asymmetric encrypt-decrypt This commit only moves test functions and test cases around. It does not modify, add or remove tests.
2018-06-18 22:20:03 +02:00
/* BEGIN_CASE */
void key_policy( int usage_arg, int alg_arg )
{
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t handle = 0;
Reorder PSA test cases to group them by topic * init-deinit * import-export * policies * lifetime * hash * MAC * cipher * AEAD * asymmetric sign * asymmetric verify * asymmetric encrypt-decrypt This commit only moves test functions and test cases around. It does not modify, add or remove tests.
2018-06-18 22:20:03 +02:00
psa_algorithm_t alg = alg_arg;
psa_key_usage_t usage = usage_arg;
psa_key_type_t key_type = PSA_KEY_TYPE_AES;
unsigned char key[32] = {0};
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
Reorder PSA test cases to group them by topic * init-deinit * import-export * policies * lifetime * hash * MAC * cipher * AEAD * asymmetric sign * asymmetric verify * asymmetric encrypt-decrypt This commit only moves test functions and test cases around. It does not modify, add or remove tests.
2018-06-18 22:20:03 +02:00
memset( key, 0x2a, sizeof( key ) );
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
Reorder PSA test cases to group them by topic * init-deinit * import-export * policies * lifetime * hash * MAC * cipher * AEAD * asymmetric sign * asymmetric verify * asymmetric encrypt-decrypt This commit only moves test functions and test cases around. It does not modify, add or remove tests.
2018-06-18 22:20:03 +02:00
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
psa_set_key_usage_flags( &attributes, usage );
psa_set_key_algorithm( &attributes, alg );
psa_set_key_type( &attributes, key_type );
Reorder PSA test cases to group them by topic * init-deinit * import-export * policies * lifetime * hash * MAC * cipher * AEAD * asymmetric sign * asymmetric verify * asymmetric encrypt-decrypt This commit only moves test functions and test cases around. It does not modify, add or remove tests.
2018-06-18 22:20:03 +02:00
Put handle parameter last: psa_import_key In psa_import_key, change the order of parameters to pass the pointer where the newly created handle will be stored last. This is consistent with most other library functions that put inputs before outputs.
2019-05-15 20:15:10 +02:00
PSA_ASSERT( psa_import_key( &attributes, key, sizeof( key ), &handle ) );
Reorder PSA test cases to group them by topic * init-deinit * import-export * policies * lifetime * hash * MAC * cipher * AEAD * asymmetric sign * asymmetric verify * asymmetric encrypt-decrypt This commit only moves test functions and test cases around. It does not modify, add or remove tests.
2018-06-18 22:20:03 +02:00
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
PSA_ASSERT( psa_get_key_attributes( handle, &attributes ) );
TEST_EQUAL( psa_get_key_type( &attributes ), key_type );
TEST_EQUAL( psa_get_key_usage_flags( &attributes ), usage );
TEST_EQUAL( psa_get_key_algorithm( &attributes ), alg );
Reorder PSA test cases to group them by topic * init-deinit * import-export * policies * lifetime * hash * MAC * cipher * AEAD * asymmetric sign * asymmetric verify * asymmetric encrypt-decrypt This commit only moves test functions and test cases around. It does not modify, add or remove tests.
2018-06-18 22:20:03 +02:00
exit:
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_destroy_key( handle );
Call psa_reset_key_attributes after psa_get_key_attributes After calling psa_get_key_attributes(), call psa_reset_key_attributes() if the key may have domain parameters, because that's the way to free the domain parameter substructure in the attribute structure. Keep not calling reset() in some places where the key can only be a symmetric key which doesn't have domain parameters.
2019-04-26 16:03:33 +02:00
psa_reset_key_attributes( &attributes );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
Reorder PSA test cases to group them by topic * init-deinit * import-export * policies * lifetime * hash * MAC * cipher * AEAD * asymmetric sign * asymmetric verify * asymmetric encrypt-decrypt This commit only moves test functions and test cases around. It does not modify, add or remove tests.
2018-06-18 22:20:03 +02:00
}
/* END_CASE */
psa: Add initializers for key policies Add new initializers for key policies and use them in our docs, example programs, tests, and library code. Prefer using the macro initializers due to their straightforwardness.
2019-01-04 12:47:20 +01:00
/* BEGIN_CASE */
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
void key_attributes_init( )
psa: Add initializers for key policies Add new initializers for key policies and use them in our docs, example programs, tests, and library code. Prefer using the macro initializers due to their straightforwardness.
2019-01-04 12:47:20 +01:00
{
/* Test each valid way of initializing the object, except for `= {0}`, as
* Clang 5 complains when `-Wmissing-field-initializers` is used, even
* though it's OK by the C standard. We could test for this, but we'd need
* to supress the Clang warning for the test. */
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_key_attributes_t func = psa_key_attributes_init( );
psa_key_attributes_t init = PSA_KEY_ATTRIBUTES_INIT;
psa_key_attributes_t zero;
psa: Add initializers for key policies Add new initializers for key policies and use them in our docs, example programs, tests, and library code. Prefer using the macro initializers due to their straightforwardness.
2019-01-04 12:47:20 +01:00
memset( &zero, 0, sizeof( zero ) );
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
TEST_EQUAL( psa_get_key_lifetime( &func ), PSA_KEY_LIFETIME_VOLATILE );
TEST_EQUAL( psa_get_key_lifetime( &init ), PSA_KEY_LIFETIME_VOLATILE );
TEST_EQUAL( psa_get_key_lifetime( &zero ), PSA_KEY_LIFETIME_VOLATILE );
TEST_EQUAL( psa_get_key_type( &func ), 0 );
TEST_EQUAL( psa_get_key_type( &init ), 0 );
TEST_EQUAL( psa_get_key_type( &zero ), 0 );
psa: Test fresh contexts have default behavior Test that freshly-initialized contexts exhibit default behavior through the API. Do this without depending on the internal representation of the contexts. This provides better portability of our tests on compilers like MSVC.
2019-02-07 17:33:37 +01:00
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
TEST_EQUAL( psa_get_key_bits( &func ), 0 );
TEST_EQUAL( psa_get_key_bits( &init ), 0 );
TEST_EQUAL( psa_get_key_bits( &zero ), 0 );
TEST_EQUAL( psa_get_key_usage_flags( &func ), 0 );
TEST_EQUAL( psa_get_key_usage_flags( &init ), 0 );
TEST_EQUAL( psa_get_key_usage_flags( &zero ), 0 );
TEST_EQUAL( psa_get_key_algorithm( &func ), 0 );
TEST_EQUAL( psa_get_key_algorithm( &init ), 0 );
TEST_EQUAL( psa_get_key_algorithm( &zero ), 0 );
psa: Add initializers for key policies Add new initializers for key policies and use them in our docs, example programs, tests, and library code. Prefer using the macro initializers due to their straightforwardness.
2019-01-04 12:47:20 +01:00
}
/* END_CASE */
Reorder PSA test cases to group them by topic * init-deinit * import-export * policies * lifetime * hash * MAC * cipher * AEAD * asymmetric sign * asymmetric verify * asymmetric encrypt-decrypt This commit only moves test functions and test cases around. It does not modify, add or remove tests.
2018-06-18 22:20:03 +02:00
/* BEGIN_CASE */
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
void mac_key_policy( int policy_usage,
int policy_alg,
int key_type,
data_t *key_data,
int exercise_alg )
Reorder PSA test cases to group them by topic * init-deinit * import-export * policies * lifetime * hash * MAC * cipher * AEAD * asymmetric sign * asymmetric verify * asymmetric encrypt-decrypt This commit only moves test functions and test cases around. It does not modify, add or remove tests.
2018-06-18 22:20:03 +02:00
{
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t handle = 0;
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
psa: Add initializers for MAC operation objects Add new initializers for MAC operation objects and use them in our tests and library code. Prefer using the macro initializers due to their straightforwardness.
2019-01-04 12:48:03 +01:00
psa_mac_operation_t operation = PSA_MAC_OPERATION_INIT;
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
psa_status_t status;
unsigned char mac[PSA_MAC_MAX_SIZE];
Reorder PSA test cases to group them by topic * init-deinit * import-export * policies * lifetime * hash * MAC * cipher * AEAD * asymmetric sign * asymmetric verify * asymmetric encrypt-decrypt This commit only moves test functions and test cases around. It does not modify, add or remove tests.
2018-06-18 22:20:03 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
Reorder PSA test cases to group them by topic * init-deinit * import-export * policies * lifetime * hash * MAC * cipher * AEAD * asymmetric sign * asymmetric verify * asymmetric encrypt-decrypt This commit only moves test functions and test cases around. It does not modify, add or remove tests.
2018-06-18 22:20:03 +02:00
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_set_key_usage_flags( &attributes, policy_usage );
psa_set_key_algorithm( &attributes, policy_alg );
psa_set_key_type( &attributes, key_type );
Reorder PSA test cases to group them by topic * init-deinit * import-export * policies * lifetime * hash * MAC * cipher * AEAD * asymmetric sign * asymmetric verify * asymmetric encrypt-decrypt This commit only moves test functions and test cases around. It does not modify, add or remove tests.
2018-06-18 22:20:03 +02:00
Split long lines after psa_import_key refactoring
2019-05-15 20:22:09 +02:00
PSA_ASSERT( psa_import_key( &attributes, key_data->x, key_data->len,
&handle ) );
Reorder PSA test cases to group them by topic * init-deinit * import-export * policies * lifetime * hash * MAC * cipher * AEAD * asymmetric sign * asymmetric verify * asymmetric encrypt-decrypt This commit only moves test functions and test cases around. It does not modify, add or remove tests.
2018-06-18 22:20:03 +02:00
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
status = psa_mac_sign_setup( &operation, handle, exercise_alg );
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
if( policy_alg == exercise_alg &&
( policy_usage & PSA_KEY_USAGE_SIGN ) != 0 )
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( status );
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
else
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( status, PSA_ERROR_NOT_PERMITTED );
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
psa_mac_abort( &operation );
memset( mac, 0, sizeof( mac ) );
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
status = psa_mac_verify_setup( &operation, handle, exercise_alg );
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
if( policy_alg == exercise_alg &&
( policy_usage & PSA_KEY_USAGE_VERIFY ) != 0 )
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( status );
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
else
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( status, PSA_ERROR_NOT_PERMITTED );
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
exit:
psa_mac_abort( &operation );
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_destroy_key( handle );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
}
/* END_CASE */
/* BEGIN_CASE */
void cipher_key_policy( int policy_usage,
int policy_alg,
int key_type,
data_t *key_data,
int exercise_alg )
{
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t handle = 0;
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
psa: Add initializers for cipher operation objects Add new initializers for cipher operation objects and use them in our tests and library code. Prefer using the macro initializers due to their straightforwardness.
2019-01-04 12:48:27 +01:00
psa_cipher_operation_t operation = PSA_CIPHER_OPERATION_INIT;
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
psa_status_t status;
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_set_key_usage_flags( &attributes, policy_usage );
psa_set_key_algorithm( &attributes, policy_alg );
psa_set_key_type( &attributes, key_type );
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
Split long lines after psa_import_key refactoring
2019-05-15 20:22:09 +02:00
PSA_ASSERT( psa_import_key( &attributes, key_data->x, key_data->len,
&handle ) );
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
status = psa_cipher_encrypt_setup( &operation, handle, exercise_alg );
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
if( policy_alg == exercise_alg &&
( policy_usage & PSA_KEY_USAGE_ENCRYPT ) != 0 )
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( status );
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
else
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( status, PSA_ERROR_NOT_PERMITTED );
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
psa_cipher_abort( &operation );
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
status = psa_cipher_decrypt_setup( &operation, handle, exercise_alg );
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
if( policy_alg == exercise_alg &&
( policy_usage & PSA_KEY_USAGE_DECRYPT ) != 0 )
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( status );
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
else
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( status, PSA_ERROR_NOT_PERMITTED );
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
exit:
psa_cipher_abort( &operation );
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_destroy_key( handle );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
}
/* END_CASE */
/* BEGIN_CASE */
void aead_key_policy( int policy_usage,
int policy_alg,
int key_type,
data_t *key_data,
int nonce_length_arg,
int tag_length_arg,
int exercise_alg )
{
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t handle = 0;
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
psa_status_t status;
unsigned char nonce[16] = {0};
size_t nonce_length = nonce_length_arg;
unsigned char tag[16];
size_t tag_length = tag_length_arg;
size_t output_length;
TEST_ASSERT( nonce_length <= sizeof( nonce ) );
TEST_ASSERT( tag_length <= sizeof( tag ) );
Reorder PSA test cases to group them by topic * init-deinit * import-export * policies * lifetime * hash * MAC * cipher * AEAD * asymmetric sign * asymmetric verify * asymmetric encrypt-decrypt This commit only moves test functions and test cases around. It does not modify, add or remove tests.
2018-06-18 22:20:03 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_set_key_usage_flags( &attributes, policy_usage );
psa_set_key_algorithm( &attributes, policy_alg );
psa_set_key_type( &attributes, key_type );
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
Split long lines after psa_import_key refactoring
2019-05-15 20:22:09 +02:00
PSA_ASSERT( psa_import_key( &attributes, key_data->x, key_data->len,
&handle ) );
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
status = psa_aead_encrypt( handle, exercise_alg,
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
nonce, nonce_length,
NULL, 0,
NULL, 0,
tag, tag_length,
&output_length );
if( policy_alg == exercise_alg &&
( policy_usage & PSA_KEY_USAGE_ENCRYPT ) != 0 )
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( status );
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
else
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( status, PSA_ERROR_NOT_PERMITTED );
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
memset( tag, 0, sizeof( tag ) );
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
status = psa_aead_decrypt( handle, exercise_alg,
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
nonce, nonce_length,
NULL, 0,
tag, tag_length,
NULL, 0,
&output_length );
if( policy_alg == exercise_alg &&
( policy_usage & PSA_KEY_USAGE_DECRYPT ) != 0 )
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( status, PSA_ERROR_INVALID_SIGNATURE );
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
else
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( status, PSA_ERROR_NOT_PERMITTED );
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
exit:
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_destroy_key( handle );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
}
/* END_CASE */
/* BEGIN_CASE */
void asymmetric_encryption_key_policy( int policy_usage,
int policy_alg,
int key_type,
data_t *key_data,
int exercise_alg )
{
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t handle = 0;
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
psa_status_t status;
size_t key_bits;
size_t buffer_length;
unsigned char *buffer = NULL;
size_t output_length;
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_set_key_usage_flags( &attributes, policy_usage );
psa_set_key_algorithm( &attributes, policy_alg );
psa_set_key_type( &attributes, key_type );
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
Split long lines after psa_import_key refactoring
2019-05-15 20:22:09 +02:00
PSA_ASSERT( psa_import_key( &attributes, key_data->x, key_data->len,
&handle ) );
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
PSA_ASSERT( psa_get_key_attributes( handle, &attributes ) );
key_bits = psa_get_key_bits( &attributes );
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
buffer_length = PSA_ASYMMETRIC_ENCRYPT_OUTPUT_SIZE( key_type, key_bits,
exercise_alg );
Use ASSERT_ALLOC instead of mbedtls_calloc in PSA tests This commit resolves a bug whereby some test cases failed on systems where mbedtls_calloc returns NULL when the size of 0, because the test case asserted `pointer != NULL` regardless of the size.
2018-09-27 13:54:18 +02:00
ASSERT_ALLOC( buffer, buffer_length );
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
status = psa_asymmetric_encrypt( handle, exercise_alg,
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
NULL, 0,
NULL, 0,
buffer, buffer_length,
&output_length );
if( policy_alg == exercise_alg &&
( policy_usage & PSA_KEY_USAGE_ENCRYPT ) != 0 )
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( status );
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
else
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( status, PSA_ERROR_NOT_PERMITTED );
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
Skip calling memset when the size is 0 memset(NULL, c, 0) has undefined behavior, so don't do it. clang-asan complains.
2018-09-26 18:19:24 +02:00
if( buffer_length != 0 )
memset( buffer, 0, buffer_length );
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
status = psa_asymmetric_decrypt( handle, exercise_alg,
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
buffer, buffer_length,
NULL, 0,
buffer, buffer_length,
&output_length );
if( policy_alg == exercise_alg &&
( policy_usage & PSA_KEY_USAGE_DECRYPT ) != 0 )
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( status, PSA_ERROR_INVALID_PADDING );
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
else
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( status, PSA_ERROR_NOT_PERMITTED );
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
exit:
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_destroy_key( handle );
Call psa_reset_key_attributes after psa_get_key_attributes After calling psa_get_key_attributes(), call psa_reset_key_attributes() if the key may have domain parameters, because that's the way to free the domain parameter substructure in the attribute structure. Keep not calling reset() in some places where the key can only be a symmetric key which doesn't have domain parameters.
2019-04-26 16:03:33 +02:00
psa_reset_key_attributes( &attributes );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
mbedtls_free( buffer );
}
/* END_CASE */
/* BEGIN_CASE */
void asymmetric_signature_key_policy( int policy_usage,
int policy_alg,
int key_type,
data_t *key_data,
Add a hash wildcard value for hash-and-sign algorithm You can use PSA_ALG_ANY_HASH to build the algorithm value for a hash-and-sign algorithm in a policy. Then the policy allows usage with this hash-and-sign family with any hash. Test that PSA_ALG_ANY_HASH-based policies allow a specific hash, but not a different hash-and-sign family. Test that PSA_ALG_ANY_HASH is not valid for operations, only in policies.
2019-01-14 16:06:39 +01:00
int exercise_alg,
int payload_length_arg )
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
{
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t handle = 0;
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
psa_status_t status;
Add a hash wildcard value for hash-and-sign algorithm You can use PSA_ALG_ANY_HASH to build the algorithm value for a hash-and-sign algorithm in a policy. Then the policy allows usage with this hash-and-sign family with any hash. Test that PSA_ALG_ANY_HASH-based policies allow a specific hash, but not a different hash-and-sign family. Test that PSA_ALG_ANY_HASH is not valid for operations, only in policies.
2019-01-14 16:06:39 +01:00
unsigned char payload[PSA_HASH_MAX_SIZE] = {1};
/* If `payload_length_arg > 0`, `exercise_alg` is supposed to be
* compatible with the policy and `payload_length_arg` is supposed to be
* a valid input length to sign. If `payload_length_arg <= 0`,
* `exercise_alg` is supposed to be forbidden by the policy. */
int compatible_alg = payload_length_arg > 0;
size_t payload_length = compatible_alg ? payload_length_arg : 0;
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
unsigned char signature[PSA_ASYMMETRIC_SIGNATURE_MAX_SIZE] = {0};
size_t signature_length;
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_set_key_usage_flags( &attributes, policy_usage );
psa_set_key_algorithm( &attributes, policy_alg );
psa_set_key_type( &attributes, key_type );
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
Split long lines after psa_import_key refactoring
2019-05-15 20:22:09 +02:00
PSA_ASSERT( psa_import_key( &attributes, key_data->x, key_data->len,
&handle ) );
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
status = psa_asymmetric_sign( handle, exercise_alg,
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
payload, payload_length,
signature, sizeof( signature ),
&signature_length );
Add a hash wildcard value for hash-and-sign algorithm You can use PSA_ALG_ANY_HASH to build the algorithm value for a hash-and-sign algorithm in a policy. Then the policy allows usage with this hash-and-sign family with any hash. Test that PSA_ALG_ANY_HASH-based policies allow a specific hash, but not a different hash-and-sign family. Test that PSA_ALG_ANY_HASH is not valid for operations, only in policies.
2019-01-14 16:06:39 +01:00
if( compatible_alg && ( policy_usage & PSA_KEY_USAGE_SIGN ) != 0 )
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( status );
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
else
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( status, PSA_ERROR_NOT_PERMITTED );
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
memset( signature, 0, sizeof( signature ) );
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
status = psa_asymmetric_verify( handle, exercise_alg,
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
payload, payload_length,
signature, sizeof( signature ) );
Add a hash wildcard value for hash-and-sign algorithm You can use PSA_ALG_ANY_HASH to build the algorithm value for a hash-and-sign algorithm in a policy. Then the policy allows usage with this hash-and-sign family with any hash. Test that PSA_ALG_ANY_HASH-based policies allow a specific hash, but not a different hash-and-sign family. Test that PSA_ALG_ANY_HASH is not valid for operations, only in policies.
2019-01-14 16:06:39 +01:00
if( compatible_alg && ( policy_usage & PSA_KEY_USAGE_VERIFY ) != 0 )
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( status, PSA_ERROR_INVALID_SIGNATURE );
Tests: cover policy checks for all operations Add tests of key policy checks for MAC, cipher, AEAD, asymmetric encryption and asymmetric signature. For each category, test with/without the requisite usage flag in each direction, and test algorithm mismatch.
2018-07-06 16:53:09 +02:00
else
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( status, PSA_ERROR_NOT_PERMITTED );
Reorder PSA test cases to group them by topic * init-deinit * import-export * policies * lifetime * hash * MAC * cipher * AEAD * asymmetric sign * asymmetric verify * asymmetric encrypt-decrypt This commit only moves test functions and test cases around. It does not modify, add or remove tests.
2018-06-18 22:20:03 +02:00
exit:
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_destroy_key( handle );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
Reorder PSA test cases to group them by topic * init-deinit * import-export * policies * lifetime * hash * MAC * cipher * AEAD * asymmetric sign * asymmetric verify * asymmetric encrypt-decrypt This commit only moves test functions and test cases around. It does not modify, add or remove tests.
2018-06-18 22:20:03 +02:00
}
/* END_CASE */
Adapt derive_key_policy test to the new API
2019-06-11 15:50:16 +02:00
/* BEGIN_CASE */
Add framework for simple key derivation New key type PSA_KEY_TYPE_DERIVE. New usage flag PSA_KEY_USAGE_DERIVE. New function psa_key_derivation. No key derivation algorithm is implemented yet. The code may not compile with -Wunused. Write some unit test code for psa_key_derivation. Most of it cannot be used yet due to the lack of a key derivation algorithm.
2018-07-12 17:17:20 +02:00
void derive_key_policy( int policy_usage,
int policy_alg,
int key_type,
data_t *key_data,
int exercise_alg )
{
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t handle = 0;
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
psa_key_derivation_operation_t operation = PSA_KEY_DERIVATION_OPERATION_INIT;
Add framework for simple key derivation New key type PSA_KEY_TYPE_DERIVE. New usage flag PSA_KEY_USAGE_DERIVE. New function psa_key_derivation. No key derivation algorithm is implemented yet. The code may not compile with -Wunused. Write some unit test code for psa_key_derivation. Most of it cannot be used yet due to the lack of a key derivation algorithm.
2018-07-12 17:17:20 +02:00
psa_status_t status;
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
Add framework for simple key derivation New key type PSA_KEY_TYPE_DERIVE. New usage flag PSA_KEY_USAGE_DERIVE. New function psa_key_derivation. No key derivation algorithm is implemented yet. The code may not compile with -Wunused. Write some unit test code for psa_key_derivation. Most of it cannot be used yet due to the lack of a key derivation algorithm.
2018-07-12 17:17:20 +02:00
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_set_key_usage_flags( &attributes, policy_usage );
psa_set_key_algorithm( &attributes, policy_alg );
psa_set_key_type( &attributes, key_type );
Add framework for simple key derivation New key type PSA_KEY_TYPE_DERIVE. New usage flag PSA_KEY_USAGE_DERIVE. New function psa_key_derivation. No key derivation algorithm is implemented yet. The code may not compile with -Wunused. Write some unit test code for psa_key_derivation. Most of it cannot be used yet due to the lack of a key derivation algorithm.
2018-07-12 17:17:20 +02:00
Split long lines after psa_import_key refactoring
2019-05-15 20:22:09 +02:00
PSA_ASSERT( psa_import_key( &attributes, key_data->x, key_data->len,
&handle ) );
Add framework for simple key derivation New key type PSA_KEY_TYPE_DERIVE. New usage flag PSA_KEY_USAGE_DERIVE. New function psa_key_derivation. No key derivation algorithm is implemented yet. The code may not compile with -Wunused. Write some unit test code for psa_key_derivation. Most of it cannot be used yet due to the lack of a key derivation algorithm.
2018-07-12 17:17:20 +02:00
Adapt derive_key_policy test to the new API
2019-06-11 15:50:16 +02:00
PSA_ASSERT( psa_key_derivation_setup( &operation, exercise_alg ) );
if( PSA_ALG_IS_TLS12_PRF( exercise_alg ) ||
PSA_ALG_IS_TLS12_PSK_TO_MS( exercise_alg ) )
Improve style
2019-06-28 14:35:36 +02:00
{
Adapt derive_key_policy test to the new API
2019-06-11 15:50:16 +02:00
PSA_ASSERT( psa_key_derivation_input_bytes(
&operation,
PSA_KEY_DERIVATION_INPUT_SEED,
(const uint8_t*) "", 0) );
Improve style
2019-06-28 14:35:36 +02:00
}
Adapt derive_key_policy test to the new API
2019-06-11 15:50:16 +02:00
status = psa_key_derivation_input_key( &operation,
PSA_KEY_DERIVATION_INPUT_SECRET,
handle );
Add framework for simple key derivation New key type PSA_KEY_TYPE_DERIVE. New usage flag PSA_KEY_USAGE_DERIVE. New function psa_key_derivation. No key derivation algorithm is implemented yet. The code may not compile with -Wunused. Write some unit test code for psa_key_derivation. Most of it cannot be used yet due to the lack of a key derivation algorithm.
2018-07-12 17:17:20 +02:00
if( policy_alg == exercise_alg &&
( policy_usage & PSA_KEY_USAGE_DERIVE ) != 0 )
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( status );
Add framework for simple key derivation New key type PSA_KEY_TYPE_DERIVE. New usage flag PSA_KEY_USAGE_DERIVE. New function psa_key_derivation. No key derivation algorithm is implemented yet. The code may not compile with -Wunused. Write some unit test code for psa_key_derivation. Most of it cannot be used yet due to the lack of a key derivation algorithm.
2018-07-12 17:17:20 +02:00
else
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( status, PSA_ERROR_NOT_PERMITTED );
Add framework for simple key derivation New key type PSA_KEY_TYPE_DERIVE. New usage flag PSA_KEY_USAGE_DERIVE. New function psa_key_derivation. No key derivation algorithm is implemented yet. The code may not compile with -Wunused. Write some unit test code for psa_key_derivation. Most of it cannot be used yet due to the lack of a key derivation algorithm.
2018-07-12 17:17:20 +02:00
exit:
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
psa_key_derivation_abort( &operation );
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_destroy_key( handle );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
Add framework for simple key derivation New key type PSA_KEY_TYPE_DERIVE. New usage flag PSA_KEY_USAGE_DERIVE. New function psa_key_derivation. No key derivation algorithm is implemented yet. The code may not compile with -Wunused. Write some unit test code for psa_key_derivation. Most of it cannot be used yet due to the lack of a key derivation algorithm.
2018-07-12 17:17:20 +02:00
}
/* END_CASE */
New API function: psa_key_agreement Set up a generator from a key agreement.
2018-09-18 12:01:02 +02:00
/* BEGIN_CASE */
void agreement_key_policy( int policy_usage,
int policy_alg,
int key_type_arg,
data_t *key_data,
int exercise_alg )
{
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t handle = 0;
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
New API function: psa_key_agreement Set up a generator from a key agreement.
2018-09-18 12:01:02 +02:00
psa_key_type_t key_type = key_type_arg;
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
psa_key_derivation_operation_t operation = PSA_KEY_DERIVATION_OPERATION_INIT;
New API function: psa_key_agreement Set up a generator from a key agreement.
2018-09-18 12:01:02 +02:00
psa_status_t status;
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
New API function: psa_key_agreement Set up a generator from a key agreement.
2018-09-18 12:01:02 +02:00
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_set_key_usage_flags( &attributes, policy_usage );
psa_set_key_algorithm( &attributes, policy_alg );
psa_set_key_type( &attributes, key_type );
New API function: psa_key_agreement Set up a generator from a key agreement.
2018-09-18 12:01:02 +02:00
Split long lines after psa_import_key refactoring
2019-05-15 20:22:09 +02:00
PSA_ASSERT( psa_import_key( &attributes, key_data->x, key_data->len,
&handle ) );
New API function: psa_key_agreement Set up a generator from a key agreement.
2018-09-18 12:01:02 +02:00
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
PSA_ASSERT( psa_key_derivation_setup( &operation, exercise_alg ) );
status = key_agreement_with_self( &operation, handle );
New API function: psa_key_agreement Set up a generator from a key agreement.
2018-09-18 12:01:02 +02:00
if( policy_alg == exercise_alg &&
( policy_usage & PSA_KEY_USAGE_DERIVE ) != 0 )
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( status );
New API function: psa_key_agreement Set up a generator from a key agreement.
2018-09-18 12:01:02 +02:00
else
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( status, PSA_ERROR_NOT_PERMITTED );
New API function: psa_key_agreement Set up a generator from a key agreement.
2018-09-18 12:01:02 +02:00
exit:
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
psa_key_derivation_abort( &operation );
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_destroy_key( handle );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
New API function: psa_key_agreement Set up a generator from a key agreement.
2018-09-18 12:01:02 +02:00
}
/* END_CASE */
Keys may allow a second algorithm Add a second permitted algorithm to key policies. This commit includes smoke tests that do not cover psa_copy_key.
2019-05-10 19:33:38 +02:00
/* BEGIN_CASE */
void key_policy_alg2( int key_type_arg, data_t *key_data,
int usage_arg, int alg_arg, int alg2_arg )
{
psa_key_handle_t handle = 0;
psa_key_type_t key_type = key_type_arg;
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
psa_key_attributes_t got_attributes = PSA_KEY_ATTRIBUTES_INIT;
psa_key_usage_t usage = usage_arg;
psa_algorithm_t alg = alg_arg;
psa_algorithm_t alg2 = alg2_arg;
PSA_ASSERT( psa_crypto_init( ) );
psa_set_key_usage_flags( &attributes, usage );
psa_set_key_algorithm( &attributes, alg );
psa_set_key_enrollment_algorithm( &attributes, alg2 );
psa_set_key_type( &attributes, key_type );
PSA_ASSERT( psa_import_key( &attributes, key_data->x, key_data->len,
&handle ) );
PSA_ASSERT( psa_get_key_attributes( handle, &got_attributes ) );
TEST_EQUAL( psa_get_key_usage_flags( &got_attributes ), usage );
TEST_EQUAL( psa_get_key_algorithm( &got_attributes ), alg );
TEST_EQUAL( psa_get_key_enrollment_algorithm( &got_attributes ), alg2 );
if( ! exercise_key( handle, usage, alg ) )
goto exit;
if( ! exercise_key( handle, usage, alg2 ) )
goto exit;
exit:
psa_destroy_key( handle );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
Keys may allow a second algorithm Add a second permitted algorithm to key policies. This commit includes smoke tests that do not cover psa_copy_key.
2019-05-10 19:33:38 +02:00
}
/* END_CASE */
Update key agreement policy tests for the new derivation API Separate test functions for raw key agreement and key agreement with KDF.
2019-04-11 21:25:46 +02:00
/* BEGIN_CASE */
void raw_agreement_key_policy( int policy_usage,
int policy_alg,
int key_type_arg,
data_t *key_data,
int exercise_alg )
{
psa_key_handle_t handle = 0;
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
Update key agreement policy tests for the new derivation API Separate test functions for raw key agreement and key agreement with KDF.
2019-04-11 21:25:46 +02:00
psa_key_type_t key_type = key_type_arg;
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
psa_key_derivation_operation_t operation = PSA_KEY_DERIVATION_OPERATION_INIT;
Update key agreement policy tests for the new derivation API Separate test functions for raw key agreement and key agreement with KDF.
2019-04-11 21:25:46 +02:00
psa_status_t status;
PSA_ASSERT( psa_crypto_init( ) );
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_set_key_usage_flags( &attributes, policy_usage );
psa_set_key_algorithm( &attributes, policy_alg );
psa_set_key_type( &attributes, key_type );
Update key agreement policy tests for the new derivation API Separate test functions for raw key agreement and key agreement with KDF.
2019-04-11 21:25:46 +02:00
Split long lines after psa_import_key refactoring
2019-05-15 20:22:09 +02:00
PSA_ASSERT( psa_import_key( &attributes, key_data->x, key_data->len,
&handle ) );
Update key agreement policy tests for the new derivation API Separate test functions for raw key agreement and key agreement with KDF.
2019-04-11 21:25:46 +02:00
status = raw_key_agreement_with_self( exercise_alg, handle );
if( policy_alg == exercise_alg &&
( policy_usage & PSA_KEY_USAGE_DERIVE ) != 0 )
PSA_ASSERT( status );
else
TEST_EQUAL( status, PSA_ERROR_NOT_PERMITTED );
exit:
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
psa_key_derivation_abort( &operation );
Update key agreement policy tests for the new derivation API Separate test functions for raw key agreement and key agreement with KDF.
2019-04-11 21:25:46 +02:00
psa_destroy_key( handle );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
Update key agreement policy tests for the new derivation API Separate test functions for raw key agreement and key agreement with KDF.
2019-04-11 21:25:46 +02:00
}
/* END_CASE */
Test psa_copy_key Split the testing into tests that exercise policies in test_suite_psa_crypto and tests that exercise slot content (slot states, key material) in test_suite_psa_crypto_slot_management. Test various cases of source and target policies with and without wildcards. Missing: testing of the policy constraint on psa_copy_key itself. Test several key types (raw data, AES, RSA). Test with the source or target being persistent. Add failure tests (incompatible policies, source slot empty, target slot occupied).
2019-01-28 13:03:09 +01:00
/* BEGIN_CASE */
Enrollment algorithm in policy: add support in psa_copy_key tests Add parameters to psa_copy_key tests for the enrollment algorithm (alg2). This commit only tests with alg2=0, which is equivalent to not setting an enrollment algorithm.
2019-05-20 17:28:11 +02:00
void copy_success( int source_usage_arg,
int source_alg_arg, int source_alg2_arg,
Test psa_copy_key with wrong type or size in attributes Split the test function copy_key into two: one for success and one for failure. Add failure tests where the attributes specify an incorrect type or size.
2019-05-03 17:14:08 +02:00
int type_arg, data_t *material,
int copy_attributes,
Enrollment algorithm in policy: add support in psa_copy_key tests Add parameters to psa_copy_key tests for the enrollment algorithm (alg2). This commit only tests with alg2=0, which is equivalent to not setting an enrollment algorithm.
2019-05-20 17:28:11 +02:00
int target_usage_arg,
int target_alg_arg, int target_alg2_arg,
int expected_usage_arg,
int expected_alg_arg, int expected_alg2_arg )
Test psa_copy_key Split the testing into tests that exercise policies in test_suite_psa_crypto and tests that exercise slot content (slot states, key material) in test_suite_psa_crypto_slot_management. Test various cases of source and target policies with and without wildcards. Missing: testing of the policy constraint on psa_copy_key itself. Test several key types (raw data, AES, RSA). Test with the source or target being persistent. Add failure tests (incompatible policies, source slot empty, target slot occupied).
2019-01-28 13:03:09 +01:00
{
Update copy_key tests to the new attribute-based interface
2019-04-19 11:43:08 +02:00
psa_key_attributes_t source_attributes = PSA_KEY_ATTRIBUTES_INIT;
psa_key_attributes_t target_attributes = PSA_KEY_ATTRIBUTES_INIT;
Test psa_copy_key Split the testing into tests that exercise policies in test_suite_psa_crypto and tests that exercise slot content (slot states, key material) in test_suite_psa_crypto_slot_management. Test various cases of source and target policies with and without wildcards. Missing: testing of the policy constraint on psa_copy_key itself. Test several key types (raw data, AES, RSA). Test with the source or target being persistent. Add failure tests (incompatible policies, source slot empty, target slot occupied).
2019-01-28 13:03:09 +01:00
psa_key_usage_t expected_usage = expected_usage_arg;
psa_algorithm_t expected_alg = expected_alg_arg;
Enrollment algorithm in policy: add support in psa_copy_key tests Add parameters to psa_copy_key tests for the enrollment algorithm (alg2). This commit only tests with alg2=0, which is equivalent to not setting an enrollment algorithm.
2019-05-20 17:28:11 +02:00
psa_algorithm_t expected_alg2 = expected_alg2_arg;
Update copy_key tests to the new attribute-based interface
2019-04-19 11:43:08 +02:00
psa_key_handle_t source_handle = 0;
psa_key_handle_t target_handle = 0;
Test psa_copy_key Split the testing into tests that exercise policies in test_suite_psa_crypto and tests that exercise slot content (slot states, key material) in test_suite_psa_crypto_slot_management. Test various cases of source and target policies with and without wildcards. Missing: testing of the policy constraint on psa_copy_key itself. Test several key types (raw data, AES, RSA). Test with the source or target being persistent. Add failure tests (incompatible policies, source slot empty, target slot occupied).
2019-01-28 13:03:09 +01:00
uint8_t *export_buffer = NULL;
PSA_ASSERT( psa_crypto_init( ) );
Update copy_key tests to the new attribute-based interface
2019-04-19 11:43:08 +02:00
/* Prepare the source key. */
psa_set_key_usage_flags( &source_attributes, source_usage_arg );
psa_set_key_algorithm( &source_attributes, source_alg_arg );
Enrollment algorithm in policy: add support in psa_copy_key tests Add parameters to psa_copy_key tests for the enrollment algorithm (alg2). This commit only tests with alg2=0, which is equivalent to not setting an enrollment algorithm.
2019-05-20 17:28:11 +02:00
psa_set_key_enrollment_algorithm( &source_attributes, source_alg2_arg );
Update copy_key tests to the new attribute-based interface
2019-04-19 11:43:08 +02:00
psa_set_key_type( &source_attributes, type_arg );
Split long lines after psa_import_key refactoring
2019-05-15 20:22:09 +02:00
PSA_ASSERT( psa_import_key( &source_attributes,
material->x, material->len,
&source_handle ) );
Update copy_key tests to the new attribute-based interface
2019-04-19 11:43:08 +02:00
PSA_ASSERT( psa_get_key_attributes( source_handle, &source_attributes ) );
Test psa_copy_key Split the testing into tests that exercise policies in test_suite_psa_crypto and tests that exercise slot content (slot states, key material) in test_suite_psa_crypto_slot_management. Test various cases of source and target policies with and without wildcards. Missing: testing of the policy constraint on psa_copy_key itself. Test several key types (raw data, AES, RSA). Test with the source or target being persistent. Add failure tests (incompatible policies, source slot empty, target slot occupied).
2019-01-28 13:03:09 +01:00
Update copy_key tests to the new attribute-based interface
2019-04-19 11:43:08 +02:00
/* Prepare the target attributes. */
if( copy_attributes )
target_attributes = source_attributes;
if( target_usage_arg != -1 )
psa_set_key_usage_flags( &target_attributes, target_usage_arg );
if( target_alg_arg != -1 )
psa_set_key_algorithm( &target_attributes, target_alg_arg );
Enrollment algorithm in policy: add support in psa_copy_key tests Add parameters to psa_copy_key tests for the enrollment algorithm (alg2). This commit only tests with alg2=0, which is equivalent to not setting an enrollment algorithm.
2019-05-20 17:28:11 +02:00
if( target_alg2_arg != -1 )
psa_set_key_enrollment_algorithm( &target_attributes, target_alg2_arg );
Test psa_copy_key Split the testing into tests that exercise policies in test_suite_psa_crypto and tests that exercise slot content (slot states, key material) in test_suite_psa_crypto_slot_management. Test various cases of source and target policies with and without wildcards. Missing: testing of the policy constraint on psa_copy_key itself. Test several key types (raw data, AES, RSA). Test with the source or target being persistent. Add failure tests (incompatible policies, source slot empty, target slot occupied).
2019-01-28 13:03:09 +01:00
/* Copy the key. */
Test psa_copy_key with wrong type or size in attributes Split the test function copy_key into two: one for success and one for failure. Add failure tests where the attributes specify an incorrect type or size.
2019-05-03 17:14:08 +02:00
PSA_ASSERT( psa_copy_key( source_handle,
&target_attributes, &target_handle ) );
Test psa_copy_key Split the testing into tests that exercise policies in test_suite_psa_crypto and tests that exercise slot content (slot states, key material) in test_suite_psa_crypto_slot_management. Test various cases of source and target policies with and without wildcards. Missing: testing of the policy constraint on psa_copy_key itself. Test several key types (raw data, AES, RSA). Test with the source or target being persistent. Add failure tests (incompatible policies, source slot empty, target slot occupied).
2019-01-28 13:03:09 +01:00
/* Destroy the source to ensure that this doesn't affect the target. */
PSA_ASSERT( psa_destroy_key( source_handle ) );
/* Test that the target slot has the expected content and policy. */
Update copy_key tests to the new attribute-based interface
2019-04-19 11:43:08 +02:00
PSA_ASSERT( psa_get_key_attributes( target_handle, &target_attributes ) );
TEST_EQUAL( psa_get_key_type( &source_attributes ),
psa_get_key_type( &target_attributes ) );
TEST_EQUAL( psa_get_key_bits( &source_attributes ),
psa_get_key_bits( &target_attributes ) );
TEST_EQUAL( expected_usage, psa_get_key_usage_flags( &target_attributes ) );
TEST_EQUAL( expected_alg, psa_get_key_algorithm( &target_attributes ) );
Enrollment algorithm in policy: add support in psa_copy_key tests Add parameters to psa_copy_key tests for the enrollment algorithm (alg2). This commit only tests with alg2=0, which is equivalent to not setting an enrollment algorithm.
2019-05-20 17:28:11 +02:00
TEST_EQUAL( expected_alg2,
psa_get_key_enrollment_algorithm( &target_attributes ) );
Test psa_copy_key Split the testing into tests that exercise policies in test_suite_psa_crypto and tests that exercise slot content (slot states, key material) in test_suite_psa_crypto_slot_management. Test various cases of source and target policies with and without wildcards. Missing: testing of the policy constraint on psa_copy_key itself. Test several key types (raw data, AES, RSA). Test with the source or target being persistent. Add failure tests (incompatible policies, source slot empty, target slot occupied).
2019-01-28 13:03:09 +01:00
if( expected_usage & PSA_KEY_USAGE_EXPORT )
{
size_t length;
ASSERT_ALLOC( export_buffer, material->len );
PSA_ASSERT( psa_export_key( target_handle, export_buffer,
material->len, &length ) );
ASSERT_COMPARE( material->x, material->len,
export_buffer, length );
}
if( ! exercise_key( target_handle, expected_usage, expected_alg ) )
goto exit;
Enrollment algorithm in policy: add support in psa_copy_key tests Add parameters to psa_copy_key tests for the enrollment algorithm (alg2). This commit only tests with alg2=0, which is equivalent to not setting an enrollment algorithm.
2019-05-20 17:28:11 +02:00
if( ! exercise_key( target_handle, expected_usage, expected_alg2 ) )
goto exit;
Test psa_copy_key Split the testing into tests that exercise policies in test_suite_psa_crypto and tests that exercise slot content (slot states, key material) in test_suite_psa_crypto_slot_management. Test various cases of source and target policies with and without wildcards. Missing: testing of the policy constraint on psa_copy_key itself. Test several key types (raw data, AES, RSA). Test with the source or target being persistent. Add failure tests (incompatible policies, source slot empty, target slot occupied).
2019-01-28 13:03:09 +01:00
PSA_ASSERT( psa_close_key( target_handle ) );
exit:
Call psa_reset_key_attributes after psa_get_key_attributes After calling psa_get_key_attributes(), call psa_reset_key_attributes() if the key may have domain parameters, because that's the way to free the domain parameter substructure in the attribute structure. Keep not calling reset() in some places where the key can only be a symmetric key which doesn't have domain parameters.
2019-04-26 16:03:33 +02:00
psa_reset_key_attributes( &source_attributes );
psa_reset_key_attributes( &target_attributes );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
Test psa_copy_key Split the testing into tests that exercise policies in test_suite_psa_crypto and tests that exercise slot content (slot states, key material) in test_suite_psa_crypto_slot_management. Test various cases of source and target policies with and without wildcards. Missing: testing of the policy constraint on psa_copy_key itself. Test several key types (raw data, AES, RSA). Test with the source or target being persistent. Add failure tests (incompatible policies, source slot empty, target slot occupied).
2019-01-28 13:03:09 +01:00
mbedtls_free( export_buffer );
}
/* END_CASE */
Test psa_copy_key with wrong type or size in attributes Split the test function copy_key into two: one for success and one for failure. Add failure tests where the attributes specify an incorrect type or size.
2019-05-03 17:14:08 +02:00
/* BEGIN_CASE */
Enrollment algorithm in policy: add support in psa_copy_key tests Add parameters to psa_copy_key tests for the enrollment algorithm (alg2). This commit only tests with alg2=0, which is equivalent to not setting an enrollment algorithm.
2019-05-20 17:28:11 +02:00
void copy_fail( int source_usage_arg,
int source_alg_arg, int source_alg2_arg,
Test psa_copy_key with wrong type or size in attributes Split the test function copy_key into two: one for success and one for failure. Add failure tests where the attributes specify an incorrect type or size.
2019-05-03 17:14:08 +02:00
int type_arg, data_t *material,
int target_type_arg, int target_bits_arg,
Enrollment algorithm in policy: add support in psa_copy_key tests Add parameters to psa_copy_key tests for the enrollment algorithm (alg2). This commit only tests with alg2=0, which is equivalent to not setting an enrollment algorithm.
2019-05-20 17:28:11 +02:00
int target_usage_arg,
int target_alg_arg, int target_alg2_arg,
Test psa_copy_key with wrong type or size in attributes Split the test function copy_key into two: one for success and one for failure. Add failure tests where the attributes specify an incorrect type or size.
2019-05-03 17:14:08 +02:00
int expected_status_arg )
{
psa_key_attributes_t source_attributes = PSA_KEY_ATTRIBUTES_INIT;
psa_key_attributes_t target_attributes = PSA_KEY_ATTRIBUTES_INIT;
psa_key_handle_t source_handle = 0;
psa_key_handle_t target_handle = 0;
PSA_ASSERT( psa_crypto_init( ) );
/* Prepare the source key. */
psa_set_key_usage_flags( &source_attributes, source_usage_arg );
psa_set_key_algorithm( &source_attributes, source_alg_arg );
Enrollment algorithm in policy: add support in psa_copy_key tests Add parameters to psa_copy_key tests for the enrollment algorithm (alg2). This commit only tests with alg2=0, which is equivalent to not setting an enrollment algorithm.
2019-05-20 17:28:11 +02:00
psa_set_key_enrollment_algorithm( &source_attributes, source_alg2_arg );
Test psa_copy_key with wrong type or size in attributes Split the test function copy_key into two: one for success and one for failure. Add failure tests where the attributes specify an incorrect type or size.
2019-05-03 17:14:08 +02:00
psa_set_key_type( &source_attributes, type_arg );
Split long lines after psa_import_key refactoring
2019-05-15 20:22:09 +02:00
PSA_ASSERT( psa_import_key( &source_attributes,
material->x, material->len,
&source_handle ) );
Test psa_copy_key with wrong type or size in attributes Split the test function copy_key into two: one for success and one for failure. Add failure tests where the attributes specify an incorrect type or size.
2019-05-03 17:14:08 +02:00
/* Prepare the target attributes. */
psa_set_key_type( &target_attributes, target_type_arg );
psa_set_key_bits( &target_attributes, target_bits_arg );
psa_set_key_usage_flags( &target_attributes, target_usage_arg );
psa_set_key_algorithm( &target_attributes, target_alg_arg );
Enrollment algorithm in policy: add support in psa_copy_key tests Add parameters to psa_copy_key tests for the enrollment algorithm (alg2). This commit only tests with alg2=0, which is equivalent to not setting an enrollment algorithm.
2019-05-20 17:28:11 +02:00
psa_set_key_enrollment_algorithm( &target_attributes, target_alg2_arg );
Test psa_copy_key with wrong type or size in attributes Split the test function copy_key into two: one for success and one for failure. Add failure tests where the attributes specify an incorrect type or size.
2019-05-03 17:14:08 +02:00
/* Try to copy the key. */
TEST_EQUAL( psa_copy_key( source_handle,
&target_attributes, &target_handle ),
expected_status_arg );
Close or destroy keys explicitly in tests
2019-05-28 14:08:50 +02:00
PSA_ASSERT( psa_destroy_key( source_handle ) );
Test psa_copy_key with wrong type or size in attributes Split the test function copy_key into two: one for success and one for failure. Add failure tests where the attributes specify an incorrect type or size.
2019-05-03 17:14:08 +02:00
exit:
psa_reset_key_attributes( &source_attributes );
psa_reset_key_attributes( &target_attributes );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
Test psa_copy_key with wrong type or size in attributes Split the test function copy_key into two: one for success and one for failure. Add failure tests where the attributes specify an incorrect type or size.
2019-05-03 17:14:08 +02:00
}
/* END_CASE */
psa: Add initializers for hash operation objects Add new initializers for hash operation objects and use them in our tests and library code. Prefer using the macro initializers due to their straightforwardness.
2019-01-04 12:47:44 +01:00
/* BEGIN_CASE */
void hash_operation_init( )
{
psa: Disallow use of invalid hash contexts If a hash context has not been set up, fail with PSA_ERROR_BAD_STATE as documented in crypto.h and the PSA Crypto specification.
2019-02-15 14:52:25 +01:00
const uint8_t input[1] = { 0 };
psa: Add initializers for hash operation objects Add new initializers for hash operation objects and use them in our tests and library code. Prefer using the macro initializers due to their straightforwardness.
2019-01-04 12:47:44 +01:00
/* Test each valid way of initializing the object, except for `= {0}`, as
* Clang 5 complains when `-Wmissing-field-initializers` is used, even
* though it's OK by the C standard. We could test for this, but we'd need
* to supress the Clang warning for the test. */
psa_hash_operation_t func = psa_hash_operation_init( );
psa_hash_operation_t init = PSA_HASH_OPERATION_INIT;
psa_hash_operation_t zero;
memset( &zero, 0, sizeof( zero ) );
psa: Extend hash bad order test Extend hash bad order test in line with the new bad order tests for MAC and cipher, covering more cases and making comments and test layout consistent. Ensure that when doing hash operations out of order, PSA_ERROR_BAD_STATE is returned as documented in crypto.h and the PSA Crypto specification.
2019-02-19 12:44:55 +01:00
/* A freshly-initialized hash operation should not be usable. */
psa: Disallow use of invalid hash contexts If a hash context has not been set up, fail with PSA_ERROR_BAD_STATE as documented in crypto.h and the PSA Crypto specification.
2019-02-15 14:52:25 +01:00
TEST_EQUAL( psa_hash_update( &func, input, sizeof( input ) ),
PSA_ERROR_BAD_STATE );
TEST_EQUAL( psa_hash_update( &init, input, sizeof( input ) ),
PSA_ERROR_BAD_STATE );
TEST_EQUAL( psa_hash_update( &zero, input, sizeof( input ) ),
PSA_ERROR_BAD_STATE );
psa: Test fresh contexts have default behavior Test that freshly-initialized contexts exhibit default behavior through the API. Do this without depending on the internal representation of the contexts. This provides better portability of our tests on compilers like MSVC.
2019-02-07 17:33:37 +01:00
/* A default hash operation should be abortable without error. */
PSA_ASSERT( psa_hash_abort( &func ) );
PSA_ASSERT( psa_hash_abort( &init ) );
PSA_ASSERT( psa_hash_abort( &zero ) );
psa: Add initializers for hash operation objects Add new initializers for hash operation objects and use them in our tests and library code. Prefer using the macro initializers due to their straightforwardness.
2019-01-04 12:47:44 +01:00
}
/* END_CASE */
Fix potential memory corruption on MAC/cipher setup failure When psa_mac_start(), psa_encrypt_setup() or psa_cipher_setup() failed, depending on when the failure happened, it was possible that psa_mac_abort() or psa_cipher_abort() would crash because it would try to call a free() function uninitialized data in the operation structure. Refactor the functions so that they initialize the operation structure before doing anything else. Add non-regression tests and a few more positive and negative unit tests for psa_mac_start() and psa_cipher_setup() (the latter via psa_encrypt_setip()).
2018-06-20 16:05:20 +02:00
/* BEGIN_CASE */
void hash_setup( int alg_arg,
int expected_status_arg )
{
psa_algorithm_t alg = alg_arg;
Get rid of some casts in test_suite_psa_crypto Use more auxiliary variables to unmarshall int values.
2018-06-21 09:25:10 +02:00
psa_status_t expected_status = expected_status_arg;
psa: Add initializers for hash operation objects Add new initializers for hash operation objects and use them in our tests and library code. Prefer using the macro initializers due to their straightforwardness.
2019-01-04 12:47:44 +01:00
psa_hash_operation_t operation = PSA_HASH_OPERATION_INIT;
Fix potential memory corruption on MAC/cipher setup failure When psa_mac_start(), psa_encrypt_setup() or psa_cipher_setup() failed, depending on when the failure happened, it was possible that psa_mac_abort() or psa_cipher_abort() would crash because it would try to call a free() function uninitialized data in the operation structure. Refactor the functions so that they initialize the operation structure before doing anything else. Add non-regression tests and a few more positive and negative unit tests for psa_mac_start() and psa_cipher_setup() (the latter via psa_encrypt_setip()).
2018-06-20 16:05:20 +02:00
psa_status_t status;
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
Fix potential memory corruption on MAC/cipher setup failure When psa_mac_start(), psa_encrypt_setup() or psa_cipher_setup() failed, depending on when the failure happened, it was possible that psa_mac_abort() or psa_cipher_abort() would crash because it would try to call a free() function uninitialized data in the operation structure. Refactor the functions so that they initialize the operation structure before doing anything else. Add non-regression tests and a few more positive and negative unit tests for psa_mac_start() and psa_cipher_setup() (the latter via psa_encrypt_setip()).
2018-06-20 16:05:20 +02:00
Rename psa_hash_start -> psa_hash_setup Make function names for multipart operations more consistent (hash edition).
2018-07-08 19:46:38 +02:00
status = psa_hash_setup( &operation, alg );
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( status, expected_status );
Fix potential memory corruption on MAC/cipher setup failure When psa_mac_start(), psa_encrypt_setup() or psa_cipher_setup() failed, depending on when the failure happened, it was possible that psa_mac_abort() or psa_cipher_abort() would crash because it would try to call a free() function uninitialized data in the operation structure. Refactor the functions so that they initialize the operation structure before doing anything else. Add non-regression tests and a few more positive and negative unit tests for psa_mac_start() and psa_cipher_setup() (the latter via psa_encrypt_setip()).
2018-06-20 16:05:20 +02:00
Test abort after failed setup Commit "Smoke-test operation contexts after setup+abort" replaced {failed-setup; abort} sequences by {failed-setup; successful-setup}. We want to test that, but we also want to test {failed-setup; abort}. So test {failed-setup; abort; failed-setup; successful-setup}.
2019-02-25 22:11:18 +01:00
/* Whether setup succeeded or failed, abort must succeed. */
PSA_ASSERT( psa_hash_abort( &operation ) );
/* If setup failed, reproduce the failure, so as to
* test the resulting state of the operation object. */
if( status != PSA_SUCCESS )
TEST_EQUAL( psa_hash_setup( &operation, alg ), status );
Smoke-test operation contexts after setup+abort After a successful setup followed by abort, or after a failed setup from an inactive state, a context must be usable. Test this for hash, MAC and cipher contexts.
2019-02-25 17:42:03 +01:00
/* Now the operation object should be reusable. */
#if defined(KNOWN_SUPPORTED_HASH_ALG)
PSA_ASSERT( psa_hash_setup( &operation, KNOWN_SUPPORTED_HASH_ALG ) );
PSA_ASSERT( psa_hash_abort( &operation ) );
#endif
Fix potential memory corruption on MAC/cipher setup failure When psa_mac_start(), psa_encrypt_setup() or psa_cipher_setup() failed, depending on when the failure happened, it was possible that psa_mac_abort() or psa_cipher_abort() would crash because it would try to call a free() function uninitialized data in the operation structure. Refactor the functions so that they initialize the operation structure before doing anything else. Add non-regression tests and a few more positive and negative unit tests for psa_mac_start() and psa_cipher_setup() (the latter via psa_encrypt_setip()).
2018-06-20 16:05:20 +02:00
exit:
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
Fix potential memory corruption on MAC/cipher setup failure When psa_mac_start(), psa_encrypt_setup() or psa_cipher_setup() failed, depending on when the failure happened, it was possible that psa_mac_abort() or psa_cipher_abort() would crash because it would try to call a free() function uninitialized data in the operation structure. Refactor the functions so that they initialize the operation structure before doing anything else. Add non-regression tests and a few more positive and negative unit tests for psa_mac_start() and psa_cipher_setup() (the latter via psa_encrypt_setip()).
2018-06-20 16:05:20 +02:00
}
/* END_CASE */
Add test hash_bad_order 1. New test for testing bad order of hash function calls. 2. Removed test hash_update_bad_paths since it's test scenario was moved to the new test. 3. Moved some scenarios from test hash_verify_bad_paths to the new test.
2018-11-01 09:44:32 +01:00
/* BEGIN_CASE */
void hash_bad_order( )
{
psa: Extend hash bad order test Extend hash bad order test in line with the new bad order tests for MAC and cipher, covering more cases and making comments and test layout consistent. Ensure that when doing hash operations out of order, PSA_ERROR_BAD_STATE is returned as documented in crypto.h and the PSA Crypto specification.
2019-02-19 12:44:55 +01:00
psa_algorithm_t alg = PSA_ALG_SHA_256;
Add test hash_bad_order 1. New test for testing bad order of hash function calls. 2. Removed test hash_update_bad_paths since it's test scenario was moved to the new test. 3. Moved some scenarios from test hash_verify_bad_paths to the new test.
2018-11-01 09:44:32 +01:00
unsigned char input[] = "";
/* SHA-256 hash of an empty string */
psa: Extend hash bad order test Extend hash bad order test in line with the new bad order tests for MAC and cipher, covering more cases and making comments and test layout consistent. Ensure that when doing hash operations out of order, PSA_ERROR_BAD_STATE is returned as documented in crypto.h and the PSA Crypto specification.
2019-02-19 12:44:55 +01:00
const unsigned char valid_hash[] = {
Add test hash_bad_order 1. New test for testing bad order of hash function calls. 2. Removed test hash_update_bad_paths since it's test scenario was moved to the new test. 3. Moved some scenarios from test hash_verify_bad_paths to the new test.
2018-11-01 09:44:32 +01:00
0xe3, 0xb0, 0xc4, 0x42, 0x98, 0xfc, 0x1c, 0x14, 0x9a, 0xfb, 0xf4, 0xc8,
0x99, 0x6f, 0xb9, 0x24, 0x27, 0xae, 0x41, 0xe4, 0x64, 0x9b, 0x93, 0x4c,
0xa4, 0x95, 0x99, 0x1b, 0x78, 0x52, 0xb8, 0x55 };
psa: Extend hash bad order test Extend hash bad order test in line with the new bad order tests for MAC and cipher, covering more cases and making comments and test layout consistent. Ensure that when doing hash operations out of order, PSA_ERROR_BAD_STATE is returned as documented in crypto.h and the PSA Crypto specification.
2019-02-19 12:44:55 +01:00
unsigned char hash[sizeof(valid_hash)] = { 0 };
Add test hash_bad_order 1. New test for testing bad order of hash function calls. 2. Removed test hash_update_bad_paths since it's test scenario was moved to the new test. 3. Moved some scenarios from test hash_verify_bad_paths to the new test.
2018-11-01 09:44:32 +01:00
size_t hash_len;
psa: Add initializers for hash operation objects Add new initializers for hash operation objects and use them in our tests and library code. Prefer using the macro initializers due to their straightforwardness.
2019-01-04 12:47:44 +01:00
psa_hash_operation_t operation = PSA_HASH_OPERATION_INIT;
Add test hash_bad_order 1. New test for testing bad order of hash function calls. 2. Removed test hash_update_bad_paths since it's test scenario was moved to the new test. 3. Moved some scenarios from test hash_verify_bad_paths to the new test.
2018-11-01 09:44:32 +01:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
Add test hash_bad_order 1. New test for testing bad order of hash function calls. 2. Removed test hash_update_bad_paths since it's test scenario was moved to the new test. 3. Moved some scenarios from test hash_verify_bad_paths to the new test.
2018-11-01 09:44:32 +01:00
psa: Disallow repeated setup Calling psa_*_setup() twice on a MAC, cipher, or hash context should result in a PSA_ERROR_BAD_STATE error because the operation has already been set up. Fixes #10
2019-02-19 10:25:10 +01:00
/* Call setup twice in a row. */
PSA_ASSERT( psa_hash_setup( &operation, alg ) );
TEST_EQUAL( psa_hash_setup( &operation, alg ),
PSA_ERROR_BAD_STATE );
PSA_ASSERT( psa_hash_abort( &operation ) );
psa: Extend hash bad order test Extend hash bad order test in line with the new bad order tests for MAC and cipher, covering more cases and making comments and test layout consistent. Ensure that when doing hash operations out of order, PSA_ERROR_BAD_STATE is returned as documented in crypto.h and the PSA Crypto specification.
2019-02-19 12:44:55 +01:00
/* Call update without calling setup beforehand. */
Rewrap some lines after the macro changes Change the way some lines are wrapped to cut at a more logical place. This commit mainly rewrites multi-line calls to TEST_EQUAL, and also a few calls to PSA_ASSERT.
2018-12-18 00:33:25 +01:00
TEST_EQUAL( psa_hash_update( &operation, input, sizeof( input ) ),
psa: Extend hash bad order test Extend hash bad order test in line with the new bad order tests for MAC and cipher, covering more cases and making comments and test layout consistent. Ensure that when doing hash operations out of order, PSA_ERROR_BAD_STATE is returned as documented in crypto.h and the PSA Crypto specification.
2019-02-19 12:44:55 +01:00
PSA_ERROR_BAD_STATE );
PSA_ASSERT( psa_hash_abort( &operation ) );
Add test hash_bad_order 1. New test for testing bad order of hash function calls. 2. Removed test hash_update_bad_paths since it's test scenario was moved to the new test. 3. Moved some scenarios from test hash_verify_bad_paths to the new test.
2018-11-01 09:44:32 +01:00
psa: Extend hash bad order test Extend hash bad order test in line with the new bad order tests for MAC and cipher, covering more cases and making comments and test layout consistent. Ensure that when doing hash operations out of order, PSA_ERROR_BAD_STATE is returned as documented in crypto.h and the PSA Crypto specification.
2019-02-19 12:44:55 +01:00
/* Call update after finish. */
PSA_ASSERT( psa_hash_setup( &operation, alg ) );
PSA_ASSERT( psa_hash_finish( &operation,
hash, sizeof( hash ), &hash_len ) );
Rewrap some lines after the macro changes Change the way some lines are wrapped to cut at a more logical place. This commit mainly rewrites multi-line calls to TEST_EQUAL, and also a few calls to PSA_ASSERT.
2018-12-18 00:33:25 +01:00
TEST_EQUAL( psa_hash_update( &operation, input, sizeof( input ) ),
psa: Disallow use of invalid hash contexts If a hash context has not been set up, fail with PSA_ERROR_BAD_STATE as documented in crypto.h and the PSA Crypto specification.
2019-02-15 14:52:25 +01:00
PSA_ERROR_BAD_STATE );
psa: Extend hash bad order test Extend hash bad order test in line with the new bad order tests for MAC and cipher, covering more cases and making comments and test layout consistent. Ensure that when doing hash operations out of order, PSA_ERROR_BAD_STATE is returned as documented in crypto.h and the PSA Crypto specification.
2019-02-19 12:44:55 +01:00
PSA_ASSERT( psa_hash_abort( &operation ) );
Add test hash_bad_order 1. New test for testing bad order of hash function calls. 2. Removed test hash_update_bad_paths since it's test scenario was moved to the new test. 3. Moved some scenarios from test hash_verify_bad_paths to the new test.
2018-11-01 09:44:32 +01:00
psa: Extend hash bad order test Extend hash bad order test in line with the new bad order tests for MAC and cipher, covering more cases and making comments and test layout consistent. Ensure that when doing hash operations out of order, PSA_ERROR_BAD_STATE is returned as documented in crypto.h and the PSA Crypto specification.
2019-02-19 12:44:55 +01:00
/* Call verify without calling setup beforehand. */
TEST_EQUAL( psa_hash_verify( &operation,
valid_hash, sizeof( valid_hash ) ),
PSA_ERROR_BAD_STATE );
PSA_ASSERT( psa_hash_abort( &operation ) );
/* Call verify after finish. */
PSA_ASSERT( psa_hash_setup( &operation, alg ) );
PSA_ASSERT( psa_hash_finish( &operation,
hash, sizeof( hash ), &hash_len ) );
TEST_EQUAL( psa_hash_verify( &operation,
valid_hash, sizeof( valid_hash ) ),
PSA_ERROR_BAD_STATE );
PSA_ASSERT( psa_hash_abort( &operation ) );
/* Call verify twice in a row. */
PSA_ASSERT( psa_hash_setup( &operation, alg ) );
PSA_ASSERT( psa_hash_verify( &operation,
valid_hash, sizeof( valid_hash ) ) );
TEST_EQUAL( psa_hash_verify( &operation,
valid_hash, sizeof( valid_hash ) ),
PSA_ERROR_BAD_STATE );
PSA_ASSERT( psa_hash_abort( &operation ) );
/* Call finish without calling setup beforehand. */
TEST_EQUAL( psa_hash_finish( &operation,
hash, sizeof( hash ), &hash_len ),
PSA_ERROR_BAD_STATE );
PSA_ASSERT( psa_hash_abort( &operation ) );
/* Call finish twice in a row. */
PSA_ASSERT( psa_hash_setup( &operation, alg ) );
PSA_ASSERT( psa_hash_finish( &operation,
hash, sizeof( hash ), &hash_len ) );
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( psa_hash_finish( &operation,
hash, sizeof( hash ), &hash_len ),
psa: Disallow use of invalid hash contexts If a hash context has not been set up, fail with PSA_ERROR_BAD_STATE as documented in crypto.h and the PSA Crypto specification.
2019-02-15 14:52:25 +01:00
PSA_ERROR_BAD_STATE );
psa: Extend hash bad order test Extend hash bad order test in line with the new bad order tests for MAC and cipher, covering more cases and making comments and test layout consistent. Ensure that when doing hash operations out of order, PSA_ERROR_BAD_STATE is returned as documented in crypto.h and the PSA Crypto specification.
2019-02-19 12:44:55 +01:00
PSA_ASSERT( psa_hash_abort( &operation ) );
Add test hash_bad_order 1. New test for testing bad order of hash function calls. 2. Removed test hash_update_bad_paths since it's test scenario was moved to the new test. 3. Moved some scenarios from test hash_verify_bad_paths to the new test.
2018-11-01 09:44:32 +01:00
psa: Extend hash bad order test Extend hash bad order test in line with the new bad order tests for MAC and cipher, covering more cases and making comments and test layout consistent. Ensure that when doing hash operations out of order, PSA_ERROR_BAD_STATE is returned as documented in crypto.h and the PSA Crypto specification.
2019-02-19 12:44:55 +01:00
/* Call finish after calling verify. */
PSA_ASSERT( psa_hash_setup( &operation, alg ) );
PSA_ASSERT( psa_hash_verify( &operation,
valid_hash, sizeof( valid_hash ) ) );
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( psa_hash_finish( &operation,
hash, sizeof( hash ), &hash_len ),
psa: Disallow use of invalid hash contexts If a hash context has not been set up, fail with PSA_ERROR_BAD_STATE as documented in crypto.h and the PSA Crypto specification.
2019-02-15 14:52:25 +01:00
PSA_ERROR_BAD_STATE );
psa: Extend hash bad order test Extend hash bad order test in line with the new bad order tests for MAC and cipher, covering more cases and making comments and test layout consistent. Ensure that when doing hash operations out of order, PSA_ERROR_BAD_STATE is returned as documented in crypto.h and the PSA Crypto specification.
2019-02-19 12:44:55 +01:00
PSA_ASSERT( psa_hash_abort( &operation ) );
Add test hash_bad_order 1. New test for testing bad order of hash function calls. 2. Removed test hash_update_bad_paths since it's test scenario was moved to the new test. 3. Moved some scenarios from test hash_verify_bad_paths to the new test.
2018-11-01 09:44:32 +01:00
exit:
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
Add test hash_bad_order 1. New test for testing bad order of hash function calls. 2. Removed test hash_update_bad_paths since it's test scenario was moved to the new test. 3. Moved some scenarios from test hash_verify_bad_paths to the new test.
2018-11-01 09:44:32 +01:00
}
/* END_CASE */
Refactor and rename test hash_verify_bad_paths to hash_verify_bad_args 1. Updated test scenarios. 2. Renamed test and updated test description. 3. Some documentation updates. 4. Test dependency moved to .function file.
2018-11-01 13:26:34 +01:00
/* BEGIN_CASE depends_on:MBEDTLS_SHA256_C */
void hash_verify_bad_args( )
Add hash bad paths test Increase code coverage
2018-10-18 17:01:10 +02:00
{
psa_algorithm_t alg = PSA_ALG_SHA_256;
Refactor and rename test hash_verify_bad_paths to hash_verify_bad_args 1. Updated test scenarios. 2. Renamed test and updated test description. 3. Some documentation updates. 4. Test dependency moved to .function file.
2018-11-01 13:26:34 +01:00
/* SHA-256 hash of an empty string with 2 extra bytes (0xaa and 0xbb)
* appended to it */
unsigned char hash[] = {
0xe3, 0xb0, 0xc4, 0x42, 0x98, 0xfc, 0x1c, 0x14, 0x9a, 0xfb, 0xf4, 0xc8,
0x99, 0x6f, 0xb9, 0x24, 0x27, 0xae, 0x41, 0xe4, 0x64, 0x9b, 0x93, 0x4c,
0xa4, 0x95, 0x99, 0x1b, 0x78, 0x52, 0xb8, 0x55, 0xaa, 0xbb };
Add hash bad paths test Increase code coverage
2018-10-18 17:01:10 +02:00
size_t expected_size = PSA_HASH_SIZE( alg );
psa: Add initializers for hash operation objects Add new initializers for hash operation objects and use them in our tests and library code. Prefer using the macro initializers due to their straightforwardness.
2019-01-04 12:47:44 +01:00
psa_hash_operation_t operation = PSA_HASH_OPERATION_INIT;
Add hash bad paths test Increase code coverage
2018-10-18 17:01:10 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
Add hash bad paths test Increase code coverage
2018-10-18 17:01:10 +02:00
Refactor and rename test hash_verify_bad_paths to hash_verify_bad_args 1. Updated test scenarios. 2. Renamed test and updated test description. 3. Some documentation updates. 4. Test dependency moved to .function file.
2018-11-01 13:26:34 +01:00
/* psa_hash_verify with a smaller hash than expected */
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_hash_setup( &operation, alg ) );
Rewrap some lines after the macro changes Change the way some lines are wrapped to cut at a more logical place. This commit mainly rewrites multi-line calls to TEST_EQUAL, and also a few calls to PSA_ASSERT.
2018-12-18 00:33:25 +01:00
TEST_EQUAL( psa_hash_verify( &operation, hash, expected_size - 1 ),
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
PSA_ERROR_INVALID_SIGNATURE );
Add hash bad paths test Increase code coverage
2018-10-18 17:01:10 +02:00
Refactor and rename test hash_verify_bad_paths to hash_verify_bad_args 1. Updated test scenarios. 2. Renamed test and updated test description. 3. Some documentation updates. 4. Test dependency moved to .function file.
2018-11-01 13:26:34 +01:00
/* psa_hash_verify with a non-matching hash */
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_hash_setup( &operation, alg ) );
Rewrap some lines after the macro changes Change the way some lines are wrapped to cut at a more logical place. This commit mainly rewrites multi-line calls to TEST_EQUAL, and also a few calls to PSA_ASSERT.
2018-12-18 00:33:25 +01:00
TEST_EQUAL( psa_hash_verify( &operation, hash + 1, expected_size ),
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
PSA_ERROR_INVALID_SIGNATURE );
Add hash bad paths test Increase code coverage
2018-10-18 17:01:10 +02:00
Refactor and rename test hash_verify_bad_paths to hash_verify_bad_args 1. Updated test scenarios. 2. Renamed test and updated test description. 3. Some documentation updates. 4. Test dependency moved to .function file.
2018-11-01 13:26:34 +01:00
/* psa_hash_verify with a hash longer than expected */
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_hash_setup( &operation, alg ) );
Rewrap some lines after the macro changes Change the way some lines are wrapped to cut at a more logical place. This commit mainly rewrites multi-line calls to TEST_EQUAL, and also a few calls to PSA_ASSERT.
2018-12-18 00:33:25 +01:00
TEST_EQUAL( psa_hash_verify( &operation, hash, sizeof( hash ) ),
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
PSA_ERROR_INVALID_SIGNATURE );
Add scenario to test case hash_bad_paths Tests where the verification hash digest is prefixed with the expected digest but also has extra bytes appended at the end of it.
2018-10-24 17:16:19 +02:00
Add hash bad paths test Increase code coverage
2018-10-18 17:01:10 +02:00
exit:
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
Add hash bad paths test Increase code coverage
2018-10-18 17:01:10 +02:00
}
/* END_CASE */
Rename test hash_finish_bad_paths to hash_finish_bad_args Test dependency moved to .function file.
2018-11-01 10:58:59 +01:00
/* BEGIN_CASE depends_on:MBEDTLS_SHA256_C */
void hash_finish_bad_args( )
Split test hash_bad_paths into 3 different tests 1. Rename hash_bad_paths to hash_verify_bad_paths 2. Add test hash_update_bad_paths 3. Add test hash_finish_bad_paths The different scenarios tested as part of hash_bad_paths are moved to the relevant test.
2018-10-25 09:22:01 +02:00
{
psa_algorithm_t alg = PSA_ALG_SHA_256;
Rename test hash_finish_bad_paths to hash_finish_bad_args Test dependency moved to .function file.
2018-11-01 10:58:59 +01:00
unsigned char hash[PSA_HASH_MAX_SIZE];
Split test hash_bad_paths into 3 different tests 1. Rename hash_bad_paths to hash_verify_bad_paths 2. Add test hash_update_bad_paths 3. Add test hash_finish_bad_paths The different scenarios tested as part of hash_bad_paths are moved to the relevant test.
2018-10-25 09:22:01 +02:00
size_t expected_size = PSA_HASH_SIZE( alg );
psa: Add initializers for hash operation objects Add new initializers for hash operation objects and use them in our tests and library code. Prefer using the macro initializers due to their straightforwardness.
2019-01-04 12:47:44 +01:00
psa_hash_operation_t operation = PSA_HASH_OPERATION_INIT;
Split test hash_bad_paths into 3 different tests 1. Rename hash_bad_paths to hash_verify_bad_paths 2. Add test hash_update_bad_paths 3. Add test hash_finish_bad_paths The different scenarios tested as part of hash_bad_paths are moved to the relevant test.
2018-10-25 09:22:01 +02:00
size_t hash_len;
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
Split test hash_bad_paths into 3 different tests 1. Rename hash_bad_paths to hash_verify_bad_paths 2. Add test hash_update_bad_paths 3. Add test hash_finish_bad_paths The different scenarios tested as part of hash_bad_paths are moved to the relevant test.
2018-10-25 09:22:01 +02:00
/* psa_hash_finish with a smaller hash buffer than expected */
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_hash_setup( &operation, alg ) );
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( psa_hash_finish( &operation,
Rewrap some lines after the macro changes Change the way some lines are wrapped to cut at a more logical place. This commit mainly rewrites multi-line calls to TEST_EQUAL, and also a few calls to PSA_ASSERT.
2018-12-18 00:33:25 +01:00
hash, expected_size - 1, &hash_len ),
PSA_ERROR_BUFFER_TOO_SMALL );
Split test hash_bad_paths into 3 different tests 1. Rename hash_bad_paths to hash_verify_bad_paths 2. Add test hash_update_bad_paths 3. Add test hash_finish_bad_paths The different scenarios tested as part of hash_bad_paths are moved to the relevant test.
2018-10-25 09:22:01 +02:00
exit:
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
Split test hash_bad_paths into 3 different tests 1. Rename hash_bad_paths to hash_verify_bad_paths 2. Add test hash_update_bad_paths 3. Add test hash_finish_bad_paths The different scenarios tested as part of hash_bad_paths are moved to the relevant test.
2018-10-25 09:22:01 +02:00
}
/* END_CASE */
New function psa_hash_clone Clone a hash operation. Test good cases as part as multipart tests. Add new test functions for the state machine.
2019-01-19 12:03:41 +01:00
/* BEGIN_CASE depends_on:MBEDTLS_SHA256_C */
void hash_clone_source_state( )
{
psa_algorithm_t alg = PSA_ALG_SHA_256;
unsigned char hash[PSA_HASH_MAX_SIZE];
psa_hash_operation_t op_source = PSA_HASH_OPERATION_INIT;
psa_hash_operation_t op_init = PSA_HASH_OPERATION_INIT;
psa_hash_operation_t op_setup = PSA_HASH_OPERATION_INIT;
psa_hash_operation_t op_finished = PSA_HASH_OPERATION_INIT;
psa_hash_operation_t op_aborted = PSA_HASH_OPERATION_INIT;
size_t hash_len;
PSA_ASSERT( psa_crypto_init( ) );
PSA_ASSERT( psa_hash_setup( &op_source, alg ) );
PSA_ASSERT( psa_hash_setup( &op_setup, alg ) );
PSA_ASSERT( psa_hash_setup( &op_finished, alg ) );
PSA_ASSERT( psa_hash_finish( &op_finished,
hash, sizeof( hash ), &hash_len ) );
PSA_ASSERT( psa_hash_setup( &op_aborted, alg ) );
PSA_ASSERT( psa_hash_abort( &op_aborted ) );
TEST_EQUAL( psa_hash_clone( &op_source, &op_setup ),
PSA_ERROR_BAD_STATE );
PSA_ASSERT( psa_hash_clone( &op_source, &op_init ) );
PSA_ASSERT( psa_hash_finish( &op_init,
hash, sizeof( hash ), &hash_len ) );
PSA_ASSERT( psa_hash_clone( &op_source, &op_finished ) );
PSA_ASSERT( psa_hash_finish( &op_finished,
hash, sizeof( hash ), &hash_len ) );
PSA_ASSERT( psa_hash_clone( &op_source, &op_aborted ) );
PSA_ASSERT( psa_hash_finish( &op_aborted,
hash, sizeof( hash ), &hash_len ) );
exit:
psa_hash_abort( &op_source );
psa_hash_abort( &op_init );
psa_hash_abort( &op_setup );
psa_hash_abort( &op_finished );
psa_hash_abort( &op_aborted );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
New function psa_hash_clone Clone a hash operation. Test good cases as part as multipart tests. Add new test functions for the state machine.
2019-01-19 12:03:41 +01:00
}
/* END_CASE */
/* BEGIN_CASE depends_on:MBEDTLS_SHA256_C */
void hash_clone_target_state( )
{
psa_algorithm_t alg = PSA_ALG_SHA_256;
unsigned char hash[PSA_HASH_MAX_SIZE];
psa_hash_operation_t op_init = PSA_HASH_OPERATION_INIT;
psa_hash_operation_t op_setup = PSA_HASH_OPERATION_INIT;
psa_hash_operation_t op_finished = PSA_HASH_OPERATION_INIT;
psa_hash_operation_t op_aborted = PSA_HASH_OPERATION_INIT;
psa_hash_operation_t op_target = PSA_HASH_OPERATION_INIT;
size_t hash_len;
PSA_ASSERT( psa_crypto_init( ) );
PSA_ASSERT( psa_hash_setup( &op_setup, alg ) );
PSA_ASSERT( psa_hash_setup( &op_finished, alg ) );
PSA_ASSERT( psa_hash_finish( &op_finished,
hash, sizeof( hash ), &hash_len ) );
PSA_ASSERT( psa_hash_setup( &op_aborted, alg ) );
PSA_ASSERT( psa_hash_abort( &op_aborted ) );
PSA_ASSERT( psa_hash_clone( &op_setup, &op_target ) );
PSA_ASSERT( psa_hash_finish( &op_target,
hash, sizeof( hash ), &hash_len ) );
TEST_EQUAL( psa_hash_clone( &op_init, &op_target ), PSA_ERROR_BAD_STATE );
TEST_EQUAL( psa_hash_clone( &op_finished, &op_target ),
PSA_ERROR_BAD_STATE );
TEST_EQUAL( psa_hash_clone( &op_aborted, &op_target ),
PSA_ERROR_BAD_STATE );
exit:
psa_hash_abort( &op_target );
psa_hash_abort( &op_init );
psa_hash_abort( &op_setup );
psa_hash_abort( &op_finished );
psa_hash_abort( &op_aborted );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
New function psa_hash_clone Clone a hash operation. Test good cases as part as multipart tests. Add new test functions for the state machine.
2019-01-19 12:03:41 +01:00
}
/* END_CASE */
psa: Add initializers for MAC operation objects Add new initializers for MAC operation objects and use them in our tests and library code. Prefer using the macro initializers due to their straightforwardness.
2019-01-04 12:48:03 +01:00
/* BEGIN_CASE */
void mac_operation_init( )
{
psa: Disallow use of invalid MAC contexts Ensure that when doing MAC operations out of order, PSA_ERROR_BAD_STATE is returned as documented in crypto.h and the PSA Crypto specification.
2019-02-15 15:05:35 +01:00
const uint8_t input[1] = { 0 };
psa: Add initializers for MAC operation objects Add new initializers for MAC operation objects and use them in our tests and library code. Prefer using the macro initializers due to their straightforwardness.
2019-01-04 12:48:03 +01:00
/* Test each valid way of initializing the object, except for `= {0}`, as
* Clang 5 complains when `-Wmissing-field-initializers` is used, even
* though it's OK by the C standard. We could test for this, but we'd need
* to supress the Clang warning for the test. */
psa_mac_operation_t func = psa_mac_operation_init( );
psa_mac_operation_t init = PSA_MAC_OPERATION_INIT;
psa_mac_operation_t zero;
memset( &zero, 0, sizeof( zero ) );
psa: Disallow use of invalid MAC contexts Ensure that when doing MAC operations out of order, PSA_ERROR_BAD_STATE is returned as documented in crypto.h and the PSA Crypto specification.
2019-02-15 15:05:35 +01:00
/* A freshly-initialized MAC operation should not be usable. */
TEST_EQUAL( psa_mac_update( &func,
input, sizeof( input ) ),
PSA_ERROR_BAD_STATE );
TEST_EQUAL( psa_mac_update( &init,
input, sizeof( input ) ),
PSA_ERROR_BAD_STATE );
TEST_EQUAL( psa_mac_update( &zero,
input, sizeof( input ) ),
PSA_ERROR_BAD_STATE );
psa: Test fresh contexts have default behavior Test that freshly-initialized contexts exhibit default behavior through the API. Do this without depending on the internal representation of the contexts. This provides better portability of our tests on compilers like MSVC.
2019-02-07 17:33:37 +01:00
/* A default MAC operation should be abortable without error. */
PSA_ASSERT( psa_mac_abort( &func ) );
PSA_ASSERT( psa_mac_abort( &init ) );
PSA_ASSERT( psa_mac_abort( &zero ) );
psa: Add initializers for MAC operation objects Add new initializers for MAC operation objects and use them in our tests and library code. Prefer using the macro initializers due to their straightforwardness.
2019-01-04 12:48:03 +01:00
}
/* END_CASE */
Fix potential memory corruption on MAC/cipher setup failure When psa_mac_start(), psa_encrypt_setup() or psa_cipher_setup() failed, depending on when the failure happened, it was possible that psa_mac_abort() or psa_cipher_abort() would crash because it would try to call a free() function uninitialized data in the operation structure. Refactor the functions so that they initialize the operation structure before doing anything else. Add non-regression tests and a few more positive and negative unit tests for psa_mac_start() and psa_cipher_setup() (the latter via psa_encrypt_setip()).
2018-06-20 16:05:20 +02:00
/* BEGIN_CASE */
void mac_setup( int key_type_arg,
data_t *key,
int alg_arg,
int expected_status_arg )
{
psa_key_type_t key_type = key_type_arg;
psa_algorithm_t alg = alg_arg;
Get rid of some casts in test_suite_psa_crypto Use more auxiliary variables to unmarshall int values.
2018-06-21 09:25:10 +02:00
psa_status_t expected_status = expected_status_arg;
psa: Add initializers for MAC operation objects Add new initializers for MAC operation objects and use them in our tests and library code. Prefer using the macro initializers due to their straightforwardness.
2019-01-04 12:48:03 +01:00
psa_mac_operation_t operation = PSA_MAC_OPERATION_INIT;
Smoke-test operation contexts after setup+abort After a successful setup followed by abort, or after a failed setup from an inactive state, a context must be usable. Test this for hash, MAC and cipher contexts.
2019-02-25 17:42:03 +01:00
psa_status_t status = PSA_ERROR_GENERIC_ERROR;
#if defined(KNOWN_SUPPORTED_MAC_ALG)
const uint8_t smoke_test_key_data[16] = "kkkkkkkkkkkkkkkk";
#endif
Fix potential memory corruption on MAC/cipher setup failure When psa_mac_start(), psa_encrypt_setup() or psa_cipher_setup() failed, depending on when the failure happened, it was possible that psa_mac_abort() or psa_cipher_abort() would crash because it would try to call a free() function uninitialized data in the operation structure. Refactor the functions so that they initialize the operation structure before doing anything else. Add non-regression tests and a few more positive and negative unit tests for psa_mac_start() and psa_cipher_setup() (the latter via psa_encrypt_setip()).
2018-06-20 16:05:20 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
Fix potential memory corruption on MAC/cipher setup failure When psa_mac_start(), psa_encrypt_setup() or psa_cipher_setup() failed, depending on when the failure happened, it was possible that psa_mac_abort() or psa_cipher_abort() would crash because it would try to call a free() function uninitialized data in the operation structure. Refactor the functions so that they initialize the operation structure before doing anything else. Add non-regression tests and a few more positive and negative unit tests for psa_mac_start() and psa_cipher_setup() (the latter via psa_encrypt_setip()).
2018-06-20 16:05:20 +02:00
Smoke-test operation contexts after setup+abort After a successful setup followed by abort, or after a failed setup from an inactive state, a context must be usable. Test this for hash, MAC and cipher contexts.
2019-02-25 17:42:03 +01:00
if( ! exercise_mac_setup( key_type, key->x, key->len, alg,
&operation, &status ) )
goto exit;
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( status, expected_status );
Fix potential memory corruption on MAC/cipher setup failure When psa_mac_start(), psa_encrypt_setup() or psa_cipher_setup() failed, depending on when the failure happened, it was possible that psa_mac_abort() or psa_cipher_abort() would crash because it would try to call a free() function uninitialized data in the operation structure. Refactor the functions so that they initialize the operation structure before doing anything else. Add non-regression tests and a few more positive and negative unit tests for psa_mac_start() and psa_cipher_setup() (the latter via psa_encrypt_setip()).
2018-06-20 16:05:20 +02:00
Smoke-test operation contexts after setup+abort After a successful setup followed by abort, or after a failed setup from an inactive state, a context must be usable. Test this for hash, MAC and cipher contexts.
2019-02-25 17:42:03 +01:00
/* The operation object should be reusable. */
#if defined(KNOWN_SUPPORTED_MAC_ALG)
if( ! exercise_mac_setup( KNOWN_SUPPORTED_MAC_KEY_TYPE,
smoke_test_key_data,
sizeof( smoke_test_key_data ),
KNOWN_SUPPORTED_MAC_ALG,
&operation, &status ) )
goto exit;
TEST_EQUAL( status, PSA_SUCCESS );
#endif
Add tests for mac_sign
2018-08-14 15:17:54 +02:00
exit:
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
Add tests for mac_sign
2018-08-14 15:17:54 +02:00
}
/* END_CASE */
psa: Disallow use of invalid MAC contexts Ensure that when doing MAC operations out of order, PSA_ERROR_BAD_STATE is returned as documented in crypto.h and the PSA Crypto specification.
2019-02-15 15:05:35 +01:00
/* BEGIN_CASE */
void mac_bad_order( )
{
psa_key_handle_t handle = 0;
psa_key_type_t key_type = PSA_KEY_TYPE_HMAC;
psa_algorithm_t alg = PSA_ALG_HMAC(PSA_ALG_SHA_256);
const uint8_t key[] = {
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, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa };
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
psa: Disallow use of invalid MAC contexts Ensure that when doing MAC operations out of order, PSA_ERROR_BAD_STATE is returned as documented in crypto.h and the PSA Crypto specification.
2019-02-15 15:05:35 +01:00
psa_mac_operation_t operation = PSA_MAC_OPERATION_INIT;
uint8_t sign_mac[PSA_MAC_MAX_SIZE + 10] = { 0 };
size_t sign_mac_length = 0;
const uint8_t input[] = { 0xbb, 0xbb, 0xbb, 0xbb };
const uint8_t verify_mac[] = {
0x74, 0x65, 0x93, 0x8c, 0xeb, 0x1d, 0xb3, 0x76, 0x5a, 0x38, 0xe7, 0xdd,
0x85, 0xc5, 0xad, 0x4f, 0x07, 0xe7, 0xd5, 0xb2, 0x64, 0xf0, 0x1a, 0x1a,
0x2c, 0xf9, 0x18, 0xca, 0x59, 0x7e, 0x5d, 0xf6 };
PSA_ASSERT( psa_crypto_init( ) );
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_set_key_usage_flags( &attributes, PSA_KEY_USAGE_SIGN | PSA_KEY_USAGE_VERIFY );
psa_set_key_algorithm( &attributes, alg );
psa_set_key_type( &attributes, key_type );
Fix potential memory corruption on MAC/cipher setup failure When psa_mac_start(), psa_encrypt_setup() or psa_cipher_setup() failed, depending on when the failure happened, it was possible that psa_mac_abort() or psa_cipher_abort() would crash because it would try to call a free() function uninitialized data in the operation structure. Refactor the functions so that they initialize the operation structure before doing anything else. Add non-regression tests and a few more positive and negative unit tests for psa_mac_start() and psa_cipher_setup() (the latter via psa_encrypt_setip()).
2018-06-20 16:05:20 +02:00
Put handle parameter last: psa_import_key In psa_import_key, change the order of parameters to pass the pointer where the newly created handle will be stored last. This is consistent with most other library functions that put inputs before outputs.
2019-05-15 20:15:10 +02:00
PSA_ASSERT( psa_import_key( &attributes, key, sizeof( key ), &handle ) );
Fix potential memory corruption on MAC/cipher setup failure When psa_mac_start(), psa_encrypt_setup() or psa_cipher_setup() failed, depending on when the failure happened, it was possible that psa_mac_abort() or psa_cipher_abort() would crash because it would try to call a free() function uninitialized data in the operation structure. Refactor the functions so that they initialize the operation structure before doing anything else. Add non-regression tests and a few more positive and negative unit tests for psa_mac_start() and psa_cipher_setup() (the latter via psa_encrypt_setip()).
2018-06-20 16:05:20 +02:00
psa: Disallow use of invalid MAC contexts Ensure that when doing MAC operations out of order, PSA_ERROR_BAD_STATE is returned as documented in crypto.h and the PSA Crypto specification.
2019-02-15 15:05:35 +01:00
/* Call update without calling setup beforehand. */
TEST_EQUAL( psa_mac_update( &operation, input, sizeof( input ) ),
PSA_ERROR_BAD_STATE );
PSA_ASSERT( psa_mac_abort( &operation ) );
/* Call sign finish without calling setup beforehand. */
TEST_EQUAL( psa_mac_sign_finish( &operation, sign_mac, sizeof( sign_mac ),
&sign_mac_length),
PSA_ERROR_BAD_STATE );
PSA_ASSERT( psa_mac_abort( &operation ) );
/* Call verify finish without calling setup beforehand. */
TEST_EQUAL( psa_mac_verify_finish( &operation,
verify_mac, sizeof( verify_mac ) ),
PSA_ERROR_BAD_STATE );
PSA_ASSERT( psa_mac_abort( &operation ) );
psa: Disallow repeated setup Calling psa_*_setup() twice on a MAC, cipher, or hash context should result in a PSA_ERROR_BAD_STATE error because the operation has already been set up. Fixes #10
2019-02-19 10:25:10 +01:00
/* Call setup twice in a row. */
PSA_ASSERT( psa_mac_sign_setup( &operation,
handle, alg ) );
TEST_EQUAL( psa_mac_sign_setup( &operation,
handle, alg ),
PSA_ERROR_BAD_STATE );
PSA_ASSERT( psa_mac_abort( &operation ) );
psa: Disallow use of invalid MAC contexts Ensure that when doing MAC operations out of order, PSA_ERROR_BAD_STATE is returned as documented in crypto.h and the PSA Crypto specification.
2019-02-15 15:05:35 +01:00
/* Call update after sign finish. */
PSA_ASSERT( psa_mac_sign_setup( &operation,
handle, alg ) );
PSA_ASSERT( psa_mac_update( &operation, input, sizeof( input ) ) );
PSA_ASSERT( psa_mac_sign_finish( &operation,
sign_mac, sizeof( sign_mac ),
&sign_mac_length ) );
TEST_EQUAL( psa_mac_update( &operation, input, sizeof( input ) ),
PSA_ERROR_BAD_STATE );
PSA_ASSERT( psa_mac_abort( &operation ) );
/* Call update after verify finish. */
PSA_ASSERT( psa_mac_verify_setup( &operation,
Wrap and reindent some lines After renaming several identifiers, re-wrap and re-indent some lines to make the code prettier.
2019-05-16 17:53:40 +02:00
handle, alg ) );
psa: Disallow use of invalid MAC contexts Ensure that when doing MAC operations out of order, PSA_ERROR_BAD_STATE is returned as documented in crypto.h and the PSA Crypto specification.
2019-02-15 15:05:35 +01:00
PSA_ASSERT( psa_mac_update( &operation, input, sizeof( input ) ) );
PSA_ASSERT( psa_mac_verify_finish( &operation,
verify_mac, sizeof( verify_mac ) ) );
TEST_EQUAL( psa_mac_update( &operation, input, sizeof( input ) ),
PSA_ERROR_BAD_STATE );
PSA_ASSERT( psa_mac_abort( &operation ) );
/* Call sign finish twice in a row. */
PSA_ASSERT( psa_mac_sign_setup( &operation,
handle, alg ) );
PSA_ASSERT( psa_mac_update( &operation, input, sizeof( input ) ) );
PSA_ASSERT( psa_mac_sign_finish( &operation,
sign_mac, sizeof( sign_mac ),
&sign_mac_length ) );
TEST_EQUAL( psa_mac_sign_finish( &operation,
sign_mac, sizeof( sign_mac ),
&sign_mac_length ),
PSA_ERROR_BAD_STATE );
PSA_ASSERT( psa_mac_abort( &operation ) );
/* Call verify finish twice in a row. */
PSA_ASSERT( psa_mac_verify_setup( &operation,
Wrap and reindent some lines After renaming several identifiers, re-wrap and re-indent some lines to make the code prettier.
2019-05-16 17:53:40 +02:00
handle, alg ) );
psa: Disallow use of invalid MAC contexts Ensure that when doing MAC operations out of order, PSA_ERROR_BAD_STATE is returned as documented in crypto.h and the PSA Crypto specification.
2019-02-15 15:05:35 +01:00
PSA_ASSERT( psa_mac_update( &operation, input, sizeof( input ) ) );
PSA_ASSERT( psa_mac_verify_finish( &operation,
verify_mac, sizeof( verify_mac ) ) );
TEST_EQUAL( psa_mac_verify_finish( &operation,
verify_mac, sizeof( verify_mac ) ),
PSA_ERROR_BAD_STATE );
PSA_ASSERT( psa_mac_abort( &operation ) );
/* Setup sign but try verify. */
PSA_ASSERT( psa_mac_sign_setup( &operation,
handle, alg ) );
PSA_ASSERT( psa_mac_update( &operation, input, sizeof( input ) ) );
TEST_EQUAL( psa_mac_verify_finish( &operation,
verify_mac, sizeof( verify_mac ) ),
PSA_ERROR_BAD_STATE );
PSA_ASSERT( psa_mac_abort( &operation ) );
/* Setup verify but try sign. */
PSA_ASSERT( psa_mac_verify_setup( &operation,
Wrap and reindent some lines After renaming several identifiers, re-wrap and re-indent some lines to make the code prettier.
2019-05-16 17:53:40 +02:00
handle, alg ) );
psa: Disallow use of invalid MAC contexts Ensure that when doing MAC operations out of order, PSA_ERROR_BAD_STATE is returned as documented in crypto.h and the PSA Crypto specification.
2019-02-15 15:05:35 +01:00
PSA_ASSERT( psa_mac_update( &operation, input, sizeof( input ) ) );
TEST_EQUAL( psa_mac_sign_finish( &operation,
sign_mac, sizeof( sign_mac ),
&sign_mac_length ),
PSA_ERROR_BAD_STATE );
PSA_ASSERT( psa_mac_abort( &operation ) );
Fix potential memory corruption on MAC/cipher setup failure When psa_mac_start(), psa_encrypt_setup() or psa_cipher_setup() failed, depending on when the failure happened, it was possible that psa_mac_abort() or psa_cipher_abort() would crash because it would try to call a free() function uninitialized data in the operation structure. Refactor the functions so that they initialize the operation structure before doing anything else. Add non-regression tests and a few more positive and negative unit tests for psa_mac_start() and psa_cipher_setup() (the latter via psa_encrypt_setip()).
2018-06-20 16:05:20 +02:00
Close or destroy keys explicitly in tests
2019-05-28 14:08:50 +02:00
PSA_ASSERT( psa_destroy_key( handle ) );
Add tests for mac_sign
2018-08-14 15:17:54 +02:00
exit:
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
Add tests for mac_sign
2018-08-14 15:17:54 +02:00
}
/* END_CASE */
/* BEGIN_CASE */
void mac_sign( int key_type_arg,
data_t *key,
int alg_arg,
data_t *input,
data_t *expected_mac )
{
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t handle = 0;
Add tests for mac_sign
2018-08-14 15:17:54 +02:00
psa_key_type_t key_type = key_type_arg;
psa_algorithm_t alg = alg_arg;
psa: Add initializers for MAC operation objects Add new initializers for MAC operation objects and use them in our tests and library code. Prefer using the macro initializers due to their straightforwardness.
2019-01-04 12:48:03 +01:00
psa_mac_operation_t operation = PSA_MAC_OPERATION_INIT;
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
Add tests for mac_sign
2018-08-14 15:17:54 +02:00
/* Leave a little extra room in the output buffer. At the end of the
* test, we'll check that the implementation didn't overwrite onto
* this extra room. */
uint8_t actual_mac[PSA_MAC_MAX_SIZE + 10];
size_t mac_buffer_size =
PSA_MAC_FINAL_SIZE( key_type, PSA_BYTES_TO_BITS( key->len ), alg );
size_t mac_length = 0;
memset( actual_mac, '+', sizeof( actual_mac ) );
TEST_ASSERT( mac_buffer_size <= PSA_MAC_MAX_SIZE );
TEST_ASSERT( expected_mac->len <= mac_buffer_size );
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
Add tests for mac_sign
2018-08-14 15:17:54 +02:00
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_set_key_usage_flags( &attributes, PSA_KEY_USAGE_SIGN );
psa_set_key_algorithm( &attributes, alg );
psa_set_key_type( &attributes, key_type );
Add tests for mac_sign
2018-08-14 15:17:54 +02:00
Put handle parameter last: psa_import_key In psa_import_key, change the order of parameters to pass the pointer where the newly created handle will be stored last. This is consistent with most other library functions that put inputs before outputs.
2019-05-15 20:15:10 +02:00
PSA_ASSERT( psa_import_key( &attributes, key->x, key->len, &handle ) );
Add tests for mac_sign
2018-08-14 15:17:54 +02:00
/* Calculate the MAC. */
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_mac_sign_setup( &operation,
handle, alg ) );
PSA_ASSERT( psa_mac_update( &operation,
input->x, input->len ) );
PSA_ASSERT( psa_mac_sign_finish( &operation,
actual_mac, mac_buffer_size,
&mac_length ) );
Add tests for mac_sign
2018-08-14 15:17:54 +02:00
/* Compare with the expected value. */
Use ASSERT_COMPARE in preference to memcmp in PSA tests
2018-12-18 00:40:50 +01:00
ASSERT_COMPARE( expected_mac->x, expected_mac->len,
actual_mac, mac_length );
Add tests for mac_sign
2018-08-14 15:17:54 +02:00
/* Verify that the end of the buffer is untouched. */
TEST_ASSERT( mem_is_char( actual_mac + mac_length, '+',
sizeof( actual_mac ) - mac_length ) );
Fix potential memory corruption on MAC/cipher setup failure When psa_mac_start(), psa_encrypt_setup() or psa_cipher_setup() failed, depending on when the failure happened, it was possible that psa_mac_abort() or psa_cipher_abort() would crash because it would try to call a free() function uninitialized data in the operation structure. Refactor the functions so that they initialize the operation structure before doing anything else. Add non-regression tests and a few more positive and negative unit tests for psa_mac_start() and psa_cipher_setup() (the latter via psa_encrypt_setip()).
2018-06-20 16:05:20 +02:00
exit:
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_destroy_key( handle );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
Fix potential memory corruption on MAC/cipher setup failure When psa_mac_start(), psa_encrypt_setup() or psa_cipher_setup() failed, depending on when the failure happened, it was possible that psa_mac_abort() or psa_cipher_abort() would crash because it would try to call a free() function uninitialized data in the operation structure. Refactor the functions so that they initialize the operation structure before doing anything else. Add non-regression tests and a few more positive and negative unit tests for psa_mac_start() and psa_cipher_setup() (the latter via psa_encrypt_setip()).
2018-06-20 16:05:20 +02:00
}
/* END_CASE */
PSA: Implement MAC functions Implement psa_mac_start, psa_mac_update and psa_mac_final. Implement HMAC anc CMAC. Smoke tests.
2018-02-08 10:02:12 +01:00
/* BEGIN_CASE */
mac_verify: remove unused IV argument We aren't going to have MAC with IV in the API any time soon, if at all, so remove the embryonic support for it in the tests.
2018-06-18 17:03:37 +02:00
void mac_verify( int key_type_arg,
data_t *key,
int alg_arg,
data_t *input,
data_t *expected_mac )
PSA: Implement MAC functions Implement psa_mac_start, psa_mac_update and psa_mac_final. Implement HMAC anc CMAC. Smoke tests.
2018-02-08 10:02:12 +01:00
{
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t handle = 0;
PSA: Implement MAC functions Implement psa_mac_start, psa_mac_update and psa_mac_final. Implement HMAC anc CMAC. Smoke tests.
2018-02-08 10:02:12 +01:00
psa_key_type_t key_type = key_type_arg;
psa_algorithm_t alg = alg_arg;
psa: Add initializers for MAC operation objects Add new initializers for MAC operation objects and use them in our tests and library code. Prefer using the macro initializers due to their straightforwardness.
2019-01-04 12:48:03 +01:00
psa_mac_operation_t operation = PSA_MAC_OPERATION_INIT;
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
PSA: Implement MAC functions Implement psa_mac_start, psa_mac_update and psa_mac_final. Implement HMAC anc CMAC. Smoke tests.
2018-02-08 10:02:12 +01:00
Use PSA_xxx_MAX_SIZE for hash/MAC/signature size in tests In tests that had a hard-coded buffer size, use PSA_MAC_MAX_SIZE or PSA_ASYMMETRIC_SIGNATURE_MAX_SIZE as appropriate. Test that PSA_xxx_MAX_SIZE is larger than the size used in tests that expect a specific output.
2018-06-28 00:07:19 +02:00
TEST_ASSERT( expected_mac->len <= PSA_MAC_MAX_SIZE );
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
PSA: Implement MAC functions Implement psa_mac_start, psa_mac_update and psa_mac_final. Implement HMAC anc CMAC. Smoke tests.
2018-02-08 10:02:12 +01:00
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_set_key_usage_flags( &attributes, PSA_KEY_USAGE_VERIFY );
psa_set_key_algorithm( &attributes, alg );
psa_set_key_type( &attributes, key_type );
Add static internal MAC finish function add new psa_mac_finish_internal() to be called by psa_mac_finish() and psa_mac_verify() in order to be able to check key usage separatly.
2018-04-02 17:34:15 +02:00
Put handle parameter last: psa_import_key In psa_import_key, change the order of parameters to pass the pointer where the newly created handle will be stored last. This is consistent with most other library functions that put inputs before outputs.
2019-05-15 20:15:10 +02:00
PSA_ASSERT( psa_import_key( &attributes, key->x, key->len, &handle ) );
mac_verify: remove unused IV argument We aren't going to have MAC with IV in the API any time soon, if at all, so remove the embryonic support for it in the tests.
2018-06-18 17:03:37 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_mac_verify_setup( &operation,
handle, alg ) );
PSA_ASSERT( psa_destroy_key( handle ) );
PSA_ASSERT( psa_mac_update( &operation,
input->x, input->len ) );
PSA_ASSERT( psa_mac_verify_finish( &operation,
expected_mac->x,
expected_mac->len ) );
PSA: Implement MAC functions Implement psa_mac_start, psa_mac_update and psa_mac_final. Implement HMAC anc CMAC. Smoke tests.
2018-02-08 10:02:12 +01:00
exit:
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_destroy_key( handle );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
PSA: Implement MAC functions Implement psa_mac_start, psa_mac_update and psa_mac_final. Implement HMAC anc CMAC. Smoke tests.
2018-02-08 10:02:12 +01:00
}
/* END_CASE */
psa: Add initializers for cipher operation objects Add new initializers for cipher operation objects and use them in our tests and library code. Prefer using the macro initializers due to their straightforwardness.
2019-01-04 12:48:27 +01:00
/* BEGIN_CASE */
void cipher_operation_init( )
{
psa: Disallow use of invalid cipher contexts Ensure that when doing cipher operations out of order, PSA_ERROR_BAD_STATE is returned as documented in crypto.h and the PSA Crypto specification.
2019-02-15 15:12:05 +01:00
const uint8_t input[1] = { 0 };
unsigned char output[1] = { 0 };
size_t output_length;
psa: Add initializers for cipher operation objects Add new initializers for cipher operation objects and use them in our tests and library code. Prefer using the macro initializers due to their straightforwardness.
2019-01-04 12:48:27 +01:00
/* Test each valid way of initializing the object, except for `= {0}`, as
* Clang 5 complains when `-Wmissing-field-initializers` is used, even
* though it's OK by the C standard. We could test for this, but we'd need
* to supress the Clang warning for the test. */
psa_cipher_operation_t func = psa_cipher_operation_init( );
psa_cipher_operation_t init = PSA_CIPHER_OPERATION_INIT;
psa_cipher_operation_t zero;
memset( &zero, 0, sizeof( zero ) );
psa: Disallow use of invalid cipher contexts Ensure that when doing cipher operations out of order, PSA_ERROR_BAD_STATE is returned as documented in crypto.h and the PSA Crypto specification.
2019-02-15 15:12:05 +01:00
/* A freshly-initialized cipher operation should not be usable. */
TEST_EQUAL( psa_cipher_update( &func,
input, sizeof( input ),
output, sizeof( output ),
&output_length ),
PSA_ERROR_BAD_STATE );
TEST_EQUAL( psa_cipher_update( &init,
input, sizeof( input ),
output, sizeof( output ),
&output_length ),
PSA_ERROR_BAD_STATE );
TEST_EQUAL( psa_cipher_update( &zero,
input, sizeof( input ),
output, sizeof( output ),
&output_length ),
PSA_ERROR_BAD_STATE );
psa: Test fresh contexts have default behavior Test that freshly-initialized contexts exhibit default behavior through the API. Do this without depending on the internal representation of the contexts. This provides better portability of our tests on compilers like MSVC.
2019-02-07 17:33:37 +01:00
/* A default cipher operation should be abortable without error. */
PSA_ASSERT( psa_cipher_abort( &func ) );
PSA_ASSERT( psa_cipher_abort( &init ) );
PSA_ASSERT( psa_cipher_abort( &zero ) );
psa: Add initializers for cipher operation objects Add new initializers for cipher operation objects and use them in our tests and library code. Prefer using the macro initializers due to their straightforwardness.
2019-01-04 12:48:27 +01:00
}
/* END_CASE */
Fix potential memory corruption on MAC/cipher setup failure When psa_mac_start(), psa_encrypt_setup() or psa_cipher_setup() failed, depending on when the failure happened, it was possible that psa_mac_abort() or psa_cipher_abort() would crash because it would try to call a free() function uninitialized data in the operation structure. Refactor the functions so that they initialize the operation structure before doing anything else. Add non-regression tests and a few more positive and negative unit tests for psa_mac_start() and psa_cipher_setup() (the latter via psa_encrypt_setip()).
2018-06-20 16:05:20 +02:00
/* BEGIN_CASE */
void cipher_setup( int key_type_arg,
data_t *key,
int alg_arg,
int expected_status_arg )
{
psa_key_type_t key_type = key_type_arg;
psa_algorithm_t alg = alg_arg;
Get rid of some casts in test_suite_psa_crypto Use more auxiliary variables to unmarshall int values.
2018-06-21 09:25:10 +02:00
psa_status_t expected_status = expected_status_arg;
psa: Add initializers for cipher operation objects Add new initializers for cipher operation objects and use them in our tests and library code. Prefer using the macro initializers due to their straightforwardness.
2019-01-04 12:48:27 +01:00
psa_cipher_operation_t operation = PSA_CIPHER_OPERATION_INIT;
Fix potential memory corruption on MAC/cipher setup failure When psa_mac_start(), psa_encrypt_setup() or psa_cipher_setup() failed, depending on when the failure happened, it was possible that psa_mac_abort() or psa_cipher_abort() would crash because it would try to call a free() function uninitialized data in the operation structure. Refactor the functions so that they initialize the operation structure before doing anything else. Add non-regression tests and a few more positive and negative unit tests for psa_mac_start() and psa_cipher_setup() (the latter via psa_encrypt_setip()).
2018-06-20 16:05:20 +02:00
psa_status_t status;
Smoke-test operation contexts after setup+abort After a successful setup followed by abort, or after a failed setup from an inactive state, a context must be usable. Test this for hash, MAC and cipher contexts.
2019-02-25 17:42:03 +01:00
#if defined(KNOWN_SUPPORTED_MAC_ALG)
const uint8_t smoke_test_key_data[16] = "kkkkkkkkkkkkkkkk";
#endif
Fix potential memory corruption on MAC/cipher setup failure When psa_mac_start(), psa_encrypt_setup() or psa_cipher_setup() failed, depending on when the failure happened, it was possible that psa_mac_abort() or psa_cipher_abort() would crash because it would try to call a free() function uninitialized data in the operation structure. Refactor the functions so that they initialize the operation structure before doing anything else. Add non-regression tests and a few more positive and negative unit tests for psa_mac_start() and psa_cipher_setup() (the latter via psa_encrypt_setip()).
2018-06-20 16:05:20 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
Fix potential memory corruption on MAC/cipher setup failure When psa_mac_start(), psa_encrypt_setup() or psa_cipher_setup() failed, depending on when the failure happened, it was possible that psa_mac_abort() or psa_cipher_abort() would crash because it would try to call a free() function uninitialized data in the operation structure. Refactor the functions so that they initialize the operation structure before doing anything else. Add non-regression tests and a few more positive and negative unit tests for psa_mac_start() and psa_cipher_setup() (the latter via psa_encrypt_setip()).
2018-06-20 16:05:20 +02:00
Smoke-test operation contexts after setup+abort After a successful setup followed by abort, or after a failed setup from an inactive state, a context must be usable. Test this for hash, MAC and cipher contexts.
2019-02-25 17:42:03 +01:00
if( ! exercise_cipher_setup( key_type, key->x, key->len, alg,
&operation, &status ) )
goto exit;
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( status, expected_status );
Fix potential memory corruption on MAC/cipher setup failure When psa_mac_start(), psa_encrypt_setup() or psa_cipher_setup() failed, depending on when the failure happened, it was possible that psa_mac_abort() or psa_cipher_abort() would crash because it would try to call a free() function uninitialized data in the operation structure. Refactor the functions so that they initialize the operation structure before doing anything else. Add non-regression tests and a few more positive and negative unit tests for psa_mac_start() and psa_cipher_setup() (the latter via psa_encrypt_setip()).
2018-06-20 16:05:20 +02:00
Smoke-test operation contexts after setup+abort After a successful setup followed by abort, or after a failed setup from an inactive state, a context must be usable. Test this for hash, MAC and cipher contexts.
2019-02-25 17:42:03 +01:00
/* The operation object should be reusable. */
#if defined(KNOWN_SUPPORTED_CIPHER_ALG)
if( ! exercise_cipher_setup( KNOWN_SUPPORTED_CIPHER_KEY_TYPE,
smoke_test_key_data,
sizeof( smoke_test_key_data ),
KNOWN_SUPPORTED_CIPHER_ALG,
&operation, &status ) )
goto exit;
TEST_EQUAL( status, PSA_SUCCESS );
#endif
Fix potential memory corruption on MAC/cipher setup failure When psa_mac_start(), psa_encrypt_setup() or psa_cipher_setup() failed, depending on when the failure happened, it was possible that psa_mac_abort() or psa_cipher_abort() would crash because it would try to call a free() function uninitialized data in the operation structure. Refactor the functions so that they initialize the operation structure before doing anything else. Add non-regression tests and a few more positive and negative unit tests for psa_mac_start() and psa_cipher_setup() (the latter via psa_encrypt_setip()).
2018-06-20 16:05:20 +02:00
exit:
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
Fix potential memory corruption on MAC/cipher setup failure When psa_mac_start(), psa_encrypt_setup() or psa_cipher_setup() failed, depending on when the failure happened, it was possible that psa_mac_abort() or psa_cipher_abort() would crash because it would try to call a free() function uninitialized data in the operation structure. Refactor the functions so that they initialize the operation structure before doing anything else. Add non-regression tests and a few more positive and negative unit tests for psa_mac_start() and psa_cipher_setup() (the latter via psa_encrypt_setip()).
2018-06-20 16:05:20 +02:00
}
/* END_CASE */
psa: Disallow use of invalid cipher contexts Ensure that when doing cipher operations out of order, PSA_ERROR_BAD_STATE is returned as documented in crypto.h and the PSA Crypto specification.
2019-02-15 15:12:05 +01:00
/* BEGIN_CASE */
void cipher_bad_order( )
{
psa_key_handle_t handle = 0;
psa_key_type_t key_type = PSA_KEY_TYPE_AES;
psa_algorithm_t alg = PSA_ALG_CBC_PKCS7;
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
psa: Disallow use of invalid cipher contexts Ensure that when doing cipher operations out of order, PSA_ERROR_BAD_STATE is returned as documented in crypto.h and the PSA Crypto specification.
2019-02-15 15:12:05 +01:00
psa_cipher_operation_t operation = PSA_CIPHER_OPERATION_INIT;
unsigned char iv[PSA_BLOCK_CIPHER_BLOCK_SIZE(PSA_KEY_TYPE_AES)] = { 0 };
const uint8_t key[] = {
0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa, 0xaa,
0xaa, 0xaa, 0xaa, 0xaa };
const uint8_t text[] = {
0xbb, 0xbb, 0xbb, 0xbb, 0xbb, 0xbb, 0xbb, 0xbb, 0xbb, 0xbb, 0xbb, 0xbb,
0xbb, 0xbb, 0xbb, 0xbb };
uint8_t buffer[PSA_BLOCK_CIPHER_BLOCK_SIZE(PSA_KEY_TYPE_AES)] = { 0 };
size_t length = 0;
PSA_ASSERT( psa_crypto_init( ) );
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_set_key_usage_flags( &attributes, PSA_KEY_USAGE_ENCRYPT | PSA_KEY_USAGE_DECRYPT );
psa_set_key_algorithm( &attributes, alg );
psa_set_key_type( &attributes, key_type );
Put handle parameter last: psa_import_key In psa_import_key, change the order of parameters to pass the pointer where the newly created handle will be stored last. This is consistent with most other library functions that put inputs before outputs.
2019-05-15 20:15:10 +02:00
PSA_ASSERT( psa_import_key( &attributes, key, sizeof( key ), &handle ) );
Fix potential memory corruption on MAC/cipher setup failure When psa_mac_start(), psa_encrypt_setup() or psa_cipher_setup() failed, depending on when the failure happened, it was possible that psa_mac_abort() or psa_cipher_abort() would crash because it would try to call a free() function uninitialized data in the operation structure. Refactor the functions so that they initialize the operation structure before doing anything else. Add non-regression tests and a few more positive and negative unit tests for psa_mac_start() and psa_cipher_setup() (the latter via psa_encrypt_setip()).
2018-06-20 16:05:20 +02:00
psa: Disallow use of invalid cipher contexts Ensure that when doing cipher operations out of order, PSA_ERROR_BAD_STATE is returned as documented in crypto.h and the PSA Crypto specification.
2019-02-15 15:12:05 +01:00
psa: Disallow repeated setup Calling psa_*_setup() twice on a MAC, cipher, or hash context should result in a PSA_ERROR_BAD_STATE error because the operation has already been set up. Fixes #10
2019-02-19 10:25:10 +01:00
/* Call encrypt setup twice in a row. */
PSA_ASSERT( psa_cipher_encrypt_setup( &operation, handle, alg ) );
TEST_EQUAL( psa_cipher_encrypt_setup( &operation, handle, alg ),
PSA_ERROR_BAD_STATE );
PSA_ASSERT( psa_cipher_abort( &operation ) );
/* Call decrypt setup twice in a row. */
PSA_ASSERT( psa_cipher_decrypt_setup( &operation, handle, alg ) );
TEST_EQUAL( psa_cipher_decrypt_setup( &operation, handle, alg ),
PSA_ERROR_BAD_STATE );
PSA_ASSERT( psa_cipher_abort( &operation ) );
psa: Disallow use of invalid cipher contexts Ensure that when doing cipher operations out of order, PSA_ERROR_BAD_STATE is returned as documented in crypto.h and the PSA Crypto specification.
2019-02-15 15:12:05 +01:00
/* Generate an IV without calling setup beforehand. */
TEST_EQUAL( psa_cipher_generate_iv( &operation,
buffer, sizeof( buffer ),
&length ),
PSA_ERROR_BAD_STATE );
PSA_ASSERT( psa_cipher_abort( &operation ) );
/* Generate an IV twice in a row. */
PSA_ASSERT( psa_cipher_encrypt_setup( &operation, handle, alg ) );
PSA_ASSERT( psa_cipher_generate_iv( &operation,
buffer, sizeof( buffer ),
&length ) );
TEST_EQUAL( psa_cipher_generate_iv( &operation,
buffer, sizeof( buffer ),
&length ),
PSA_ERROR_BAD_STATE );
PSA_ASSERT( psa_cipher_abort( &operation ) );
/* Generate an IV after it's already set. */
PSA_ASSERT( psa_cipher_encrypt_setup( &operation, handle, alg ) );
PSA_ASSERT( psa_cipher_set_iv( &operation,
iv, sizeof( iv ) ) );
TEST_EQUAL( psa_cipher_generate_iv( &operation,
buffer, sizeof( buffer ),
&length ),
PSA_ERROR_BAD_STATE );
PSA_ASSERT( psa_cipher_abort( &operation ) );
/* Set an IV without calling setup beforehand. */
TEST_EQUAL( psa_cipher_set_iv( &operation,
iv, sizeof( iv ) ),
PSA_ERROR_BAD_STATE );
PSA_ASSERT( psa_cipher_abort( &operation ) );
/* Set an IV after it's already set. */
PSA_ASSERT( psa_cipher_encrypt_setup( &operation, handle, alg ) );
PSA_ASSERT( psa_cipher_set_iv( &operation,
iv, sizeof( iv ) ) );
TEST_EQUAL( psa_cipher_set_iv( &operation,
iv, sizeof( iv ) ),
PSA_ERROR_BAD_STATE );
PSA_ASSERT( psa_cipher_abort( &operation ) );
/* Set an IV after it's already generated. */
PSA_ASSERT( psa_cipher_encrypt_setup( &operation, handle, alg ) );
PSA_ASSERT( psa_cipher_generate_iv( &operation,
buffer, sizeof( buffer ),
&length ) );
TEST_EQUAL( psa_cipher_set_iv( &operation,
iv, sizeof( iv ) ),
PSA_ERROR_BAD_STATE );
PSA_ASSERT( psa_cipher_abort( &operation ) );
/* Call update without calling setup beforehand. */
TEST_EQUAL( psa_cipher_update( &operation,
text, sizeof( text ),
buffer, sizeof( buffer ),
&length ),
PSA_ERROR_BAD_STATE );
PSA_ASSERT( psa_cipher_abort( &operation ) );
/* Call update without an IV where an IV is required. */
TEST_EQUAL( psa_cipher_update( &operation,
text, sizeof( text ),
buffer, sizeof( buffer ),
&length ),
PSA_ERROR_BAD_STATE );
PSA_ASSERT( psa_cipher_abort( &operation ) );
/* Call update after finish. */
PSA_ASSERT( psa_cipher_encrypt_setup( &operation, handle, alg ) );
PSA_ASSERT( psa_cipher_set_iv( &operation,
iv, sizeof( iv ) ) );
PSA_ASSERT( psa_cipher_finish( &operation,
buffer, sizeof( buffer ), &length ) );
TEST_EQUAL( psa_cipher_update( &operation,
text, sizeof( text ),
buffer, sizeof( buffer ),
&length ),
PSA_ERROR_BAD_STATE );
PSA_ASSERT( psa_cipher_abort( &operation ) );
/* Call finish without calling setup beforehand. */
TEST_EQUAL( psa_cipher_finish( &operation,
buffer, sizeof( buffer ), &length ),
PSA_ERROR_BAD_STATE );
PSA_ASSERT( psa_cipher_abort( &operation ) );
/* Call finish without an IV where an IV is required. */
PSA_ASSERT( psa_cipher_encrypt_setup( &operation, handle, alg ) );
/* Not calling update means we are encrypting an empty buffer, which is OK
* for cipher modes with padding. */
TEST_EQUAL( psa_cipher_finish( &operation,
buffer, sizeof( buffer ), &length ),
PSA_ERROR_BAD_STATE );
PSA_ASSERT( psa_cipher_abort( &operation ) );
/* Call finish twice in a row. */
PSA_ASSERT( psa_cipher_encrypt_setup( &operation, handle, alg ) );
PSA_ASSERT( psa_cipher_set_iv( &operation,
iv, sizeof( iv ) ) );
PSA_ASSERT( psa_cipher_finish( &operation,
buffer, sizeof( buffer ), &length ) );
TEST_EQUAL( psa_cipher_finish( &operation,
buffer, sizeof( buffer ), &length ),
PSA_ERROR_BAD_STATE );
PSA_ASSERT( psa_cipher_abort( &operation ) );
Fix potential memory corruption on MAC/cipher setup failure When psa_mac_start(), psa_encrypt_setup() or psa_cipher_setup() failed, depending on when the failure happened, it was possible that psa_mac_abort() or psa_cipher_abort() would crash because it would try to call a free() function uninitialized data in the operation structure. Refactor the functions so that they initialize the operation structure before doing anything else. Add non-regression tests and a few more positive and negative unit tests for psa_mac_start() and psa_cipher_setup() (the latter via psa_encrypt_setip()).
2018-06-20 16:05:20 +02:00
Close or destroy keys explicitly in tests
2019-05-28 14:08:50 +02:00
PSA_ASSERT( psa_destroy_key( handle ) );
Fix potential memory corruption on MAC/cipher setup failure When psa_mac_start(), psa_encrypt_setup() or psa_cipher_setup() failed, depending on when the failure happened, it was possible that psa_mac_abort() or psa_cipher_abort() would crash because it would try to call a free() function uninitialized data in the operation structure. Refactor the functions so that they initialize the operation structure before doing anything else. Add non-regression tests and a few more positive and negative unit tests for psa_mac_start() and psa_cipher_setup() (the latter via psa_encrypt_setip()).
2018-06-20 16:05:20 +02:00
exit:
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
Fix potential memory corruption on MAC/cipher setup failure When psa_mac_start(), psa_encrypt_setup() or psa_cipher_setup() failed, depending on when the failure happened, it was possible that psa_mac_abort() or psa_cipher_abort() would crash because it would try to call a free() function uninitialized data in the operation structure. Refactor the functions so that they initialize the operation structure before doing anything else. Add non-regression tests and a few more positive and negative unit tests for psa_mac_start() and psa_cipher_setup() (the latter via psa_encrypt_setip()).
2018-06-20 16:05:20 +02:00
}
/* END_CASE */
PSA crypto: PSA_ASYMMETRIC_SIGN_OUTPUT_SIZE macro Test it for RSA.
2018-02-03 23:57:22 +01:00
/* BEGIN_CASE */
We don't need _test_ in test function names Also fix typo multpart -> multipart
2018-06-08 14:28:46 +02:00
void cipher_encrypt( int alg_arg, int key_type_arg,
Cipher tests: pass the IV from the test data Don't hard-code an IV in cipher test functions. It restricts what can be used as test data.
2019-05-06 15:22:57 +02:00
data_t *key, data_t *iv,
On target testing tests adaptation Updated all psa crypto tests to use the new test format
2018-06-12 16:06:52 +02:00
data_t *input, data_t *expected_output,
Get rid of some casts in test_suite_psa_crypto Use more auxiliary variables to unmarshall int values.
2018-06-21 09:25:10 +02:00
int expected_status_arg )
PSA crypto: asymmetric signature (RSA PKCS#1v1.5 only) Define hash algorithms and RSA signature algorithms. New function psa_asymmetric_sign. Implement psa_asymmetric_sign for RSA PKCS#1 v1.5.
2018-02-03 22:44:14 +01:00
{
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t handle = 0;
fix and add tests case + fix for padding mode
2018-06-06 15:36:50 +02:00
psa_status_t status;
PSA crypto: asymmetric signature (RSA PKCS#1v1.5 only) Define hash algorithms and RSA signature algorithms. New function psa_asymmetric_sign. Implement psa_asymmetric_sign for RSA PKCS#1 v1.5.
2018-02-03 22:44:14 +01:00
psa_key_type_t key_type = key_type_arg;
psa_algorithm_t alg = alg_arg;
Get rid of some casts in test_suite_psa_crypto Use more auxiliary variables to unmarshall int values.
2018-06-21 09:25:10 +02:00
psa_status_t expected_status = expected_status_arg;
On target testing tests adaptation Updated all psa crypto tests to use the new test format
2018-06-12 16:06:52 +02:00
unsigned char *output = NULL;
Rename some variables to make the code easier to read In cipher_test_verify_output_multpart, tweak the ways chunk sizes are added in order to get rid of the variable temp. In other functions, this commit does not change the logic at all.
2018-06-08 14:20:49 +02:00
size_t output_buffer_size = 0;
size_t function_output_length = 0;
Cipher tests: calculate and verify the actual output size
2018-06-08 14:40:59 +02:00
size_t total_output_length = 0;
psa: Add initializers for cipher operation objects Add new initializers for cipher operation objects and use them in our tests and library code. Prefer using the macro initializers due to their straightforwardness.
2019-01-04 12:48:27 +01:00
psa_cipher_operation_t operation = PSA_CIPHER_OPERATION_INIT;
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
PSA crypto: asymmetric signature (RSA PKCS#1v1.5 only) Define hash algorithms and RSA signature algorithms. New function psa_asymmetric_sign. Implement psa_asymmetric_sign for RSA PKCS#1 v1.5.
2018-02-03 22:44:14 +01:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
PSA crypto: asymmetric signature (RSA PKCS#1v1.5 only) Define hash algorithms and RSA signature algorithms. New function psa_asymmetric_sign. Implement psa_asymmetric_sign for RSA PKCS#1 v1.5.
2018-02-03 22:44:14 +01:00
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_set_key_usage_flags( &attributes, PSA_KEY_USAGE_ENCRYPT );
psa_set_key_algorithm( &attributes, alg );
psa_set_key_type( &attributes, key_type );
Fix tests in test_suite_psa_crypto to set policy usage
2018-07-05 14:22:45 +02:00
Put handle parameter last: psa_import_key In psa_import_key, change the order of parameters to pass the pointer where the newly created handle will be stored last. This is consistent with most other library functions that put inputs before outputs.
2019-05-15 20:15:10 +02:00
PSA_ASSERT( psa_import_key( &attributes, key->x, key->len, &handle ) );
Add calls for set policy in export/sign tests Add calls for set policy in export/sign tests
2018-03-28 12:46:26 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_cipher_encrypt_setup( &operation,
handle, alg ) );
Add calls for set policy in export/sign tests Add calls for set policy in export/sign tests
2018-03-28 12:46:26 +02:00
Cipher tests: pass the IV from the test data Don't hard-code an IV in cipher test functions. It restricts what can be used as test data.
2019-05-06 15:22:57 +02:00
PSA_ASSERT( psa_cipher_set_iv( &operation, iv->x, iv->len ) );
Wrap some multiline expressions in parentheses This guarantees that they'll be indented as desired under most indentation rules.
2018-12-17 23:34:57 +01:00
output_buffer_size = ( (size_t) input->len +
PSA_BLOCK_CIPHER_BLOCK_SIZE( key_type ) );
Use ASSERT_ALLOC instead of mbedtls_calloc in PSA tests This commit resolves a bug whereby some test cases failed on systems where mbedtls_calloc returns NULL when the size of 0, because the test case asserted `pointer != NULL` regardless of the size.
2018-09-27 13:54:18 +02:00
ASSERT_ALLOC( output, output_buffer_size );
PSA crypto: PSA_ASYMMETRIC_SIGN_OUTPUT_SIZE macro Test it for RSA.
2018-02-03 23:57:22 +01:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_cipher_update( &operation,
input->x, input->len,
output, output_buffer_size,
&function_output_length ) );
Cipher tests: calculate and verify the actual output size
2018-06-08 14:40:59 +02:00
total_output_length += function_output_length;
Rename some variables to make the code easier to read In cipher_test_verify_output_multpart, tweak the ways chunk sizes are added in order to get rid of the variable temp. In other functions, this commit does not change the logic at all.
2018-06-08 14:20:49 +02:00
status = psa_cipher_finish( &operation,
Fix output size calculations in cipher tests Some calls to psa_cipher_finish or psa_cipher_update append to a buffer. Several of these calls were not calculating the offset into the buffer or the remaining buffer size correctly. This did not lead to buffer overflows before because the buffer sizes were sufficiently large for our test inputs. This did not lead to incorrect output when the test was designed to append but actually wrote too early because all the existing test cases either have no output from finish (stream cipher) or have no output from update (CBC, with less than one block of input).
2019-02-19 19:05:33 +01:00
output + total_output_length,
output_buffer_size - total_output_length,
Rename some variables to make the code easier to read In cipher_test_verify_output_multpart, tweak the ways chunk sizes are added in order to get rid of the variable temp. In other functions, this commit does not change the logic at all.
2018-06-08 14:20:49 +02:00
&function_output_length );
Cipher tests: calculate and verify the actual output size
2018-06-08 14:40:59 +02:00
total_output_length += function_output_length;
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( status, expected_status );
fix and add tests case + fix for padding mode
2018-06-06 15:36:50 +02:00
if( expected_status == PSA_SUCCESS )
{
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_cipher_abort( &operation ) );
Use ASSERT_COMPARE instead of memcmp in PSA tests This commit fixes some missing size comparison. In aead_encrypt_decrypt, aead_encrypt and aead_decrypt, the test code would not have noticed if the library function had reported an output length that was not the expected length.
2018-09-27 13:57:19 +02:00
ASSERT_COMPARE( expected_output->x, expected_output->len,
output, total_output_length );
fix and add tests case + fix for padding mode
2018-06-06 15:36:50 +02:00
}
Cipher tests: calculate and verify the actual output size
2018-06-08 14:40:59 +02:00
fix and add tests case + fix for padding mode
2018-06-06 15:36:50 +02:00
exit:
On target testing tests adaptation Updated all psa crypto tests to use the new test format
2018-06-12 16:06:52 +02:00
mbedtls_free( output );
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_destroy_key( handle );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
fix and add tests case + fix for padding mode
2018-06-06 15:36:50 +02:00
}
/* END_CASE */
/* BEGIN_CASE */
We don't need _test_ in test function names Also fix typo multpart -> multipart
2018-06-08 14:28:46 +02:00
void cipher_encrypt_multipart( int alg_arg, int key_type_arg,
Cipher tests: pass the IV from the test data Don't hard-code an IV in cipher test functions. It restricts what can be used as test data.
2019-05-06 15:22:57 +02:00
data_t *key, data_t *iv,
On target testing tests adaptation Updated all psa crypto tests to use the new test format
2018-06-12 16:06:52 +02:00
data_t *input,
Test the length of cipher_update output In multipart cipher tests, test that each step of psa_cipher_update produces output of the expected length. The length is hard-coded in the test data since it depends on the mode. The length of the output of psa_cipher_finish is effectively tested because it's the total output length minus the length produced by the update steps.
2019-02-19 19:44:00 +01:00
int first_part_size_arg,
int output1_length_arg, int output2_length_arg,
On target testing tests adaptation Updated all psa crypto tests to use the new test format
2018-06-12 16:06:52 +02:00
data_t *expected_output )
fix and add tests case + fix for padding mode
2018-06-06 15:36:50 +02:00
{
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t handle = 0;
fix and add tests case + fix for padding mode
2018-06-06 15:36:50 +02:00
psa_key_type_t key_type = key_type_arg;
psa_algorithm_t alg = alg_arg;
Test the length of cipher_update output In multipart cipher tests, test that each step of psa_cipher_update produces output of the expected length. The length is hard-coded in the test data since it depends on the mode. The length of the output of psa_cipher_finish is effectively tested because it's the total output length minus the length produced by the update steps.
2019-02-19 19:44:00 +01:00
size_t first_part_size = first_part_size_arg;
size_t output1_length = output1_length_arg;
size_t output2_length = output2_length_arg;
On target testing tests adaptation Updated all psa crypto tests to use the new test format
2018-06-12 16:06:52 +02:00
unsigned char *output = NULL;
Rename some variables to make the code easier to read In cipher_test_verify_output_multpart, tweak the ways chunk sizes are added in order to get rid of the variable temp. In other functions, this commit does not change the logic at all.
2018-06-08 14:20:49 +02:00
size_t output_buffer_size = 0;
size_t function_output_length = 0;
Cipher tests: calculate and verify the actual output size
2018-06-08 14:40:59 +02:00
size_t total_output_length = 0;
psa: Add initializers for cipher operation objects Add new initializers for cipher operation objects and use them in our tests and library code. Prefer using the macro initializers due to their straightforwardness.
2019-01-04 12:48:27 +01:00
psa_cipher_operation_t operation = PSA_CIPHER_OPERATION_INIT;
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
fix and add tests case + fix for padding mode
2018-06-06 15:36:50 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
fix and add tests case + fix for padding mode
2018-06-06 15:36:50 +02:00
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_set_key_usage_flags( &attributes, PSA_KEY_USAGE_ENCRYPT );
psa_set_key_algorithm( &attributes, alg );
psa_set_key_type( &attributes, key_type );
Fix tests in test_suite_psa_crypto to set policy usage
2018-07-05 14:22:45 +02:00
Put handle parameter last: psa_import_key In psa_import_key, change the order of parameters to pass the pointer where the newly created handle will be stored last. This is consistent with most other library functions that put inputs before outputs.
2019-05-15 20:15:10 +02:00
PSA_ASSERT( psa_import_key( &attributes, key->x, key->len, &handle ) );
fix and add tests case + fix for padding mode
2018-06-06 15:36:50 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_cipher_encrypt_setup( &operation,
handle, alg ) );
fix and add tests case + fix for padding mode
2018-06-06 15:36:50 +02:00
Cipher tests: pass the IV from the test data Don't hard-code an IV in cipher test functions. It restricts what can be used as test data.
2019-05-06 15:22:57 +02:00
PSA_ASSERT( psa_cipher_set_iv( &operation, iv->x, iv->len ) );
Wrap some multiline expressions in parentheses This guarantees that they'll be indented as desired under most indentation rules.
2018-12-17 23:34:57 +01:00
output_buffer_size = ( (size_t) input->len +
PSA_BLOCK_CIPHER_BLOCK_SIZE( key_type ) );
Use ASSERT_ALLOC instead of mbedtls_calloc in PSA tests This commit resolves a bug whereby some test cases failed on systems where mbedtls_calloc returns NULL when the size of 0, because the test case asserted `pointer != NULL` regardless of the size.
2018-09-27 13:54:18 +02:00
ASSERT_ALLOC( output, output_buffer_size );
fix and add tests case + fix for padding mode
2018-06-06 15:36:50 +02:00
Test the length of cipher_update output In multipart cipher tests, test that each step of psa_cipher_update produces output of the expected length. The length is hard-coded in the test data since it depends on the mode. The length of the output of psa_cipher_finish is effectively tested because it's the total output length minus the length produced by the update steps.
2019-02-19 19:44:00 +01:00
TEST_ASSERT( first_part_size <= input->len );
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_cipher_update( &operation, input->x, first_part_size,
output, output_buffer_size,
&function_output_length ) );
Test the length of cipher_update output In multipart cipher tests, test that each step of psa_cipher_update produces output of the expected length. The length is hard-coded in the test data since it depends on the mode. The length of the output of psa_cipher_finish is effectively tested because it's the total output length minus the length produced by the update steps.
2019-02-19 19:44:00 +01:00
TEST_ASSERT( function_output_length == output1_length );
Cipher tests: calculate and verify the actual output size
2018-06-08 14:40:59 +02:00
total_output_length += function_output_length;
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_cipher_update( &operation,
input->x + first_part_size,
input->len - first_part_size,
Fix output size calculations in cipher tests Some calls to psa_cipher_finish or psa_cipher_update append to a buffer. Several of these calls were not calculating the offset into the buffer or the remaining buffer size correctly. This did not lead to buffer overflows before because the buffer sizes were sufficiently large for our test inputs. This did not lead to incorrect output when the test was designed to append but actually wrote too early because all the existing test cases either have no output from finish (stream cipher) or have no output from update (CBC, with less than one block of input).
2019-02-19 19:05:33 +01:00
output + total_output_length,
output_buffer_size - total_output_length,
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
&function_output_length ) );
Test the length of cipher_update output In multipart cipher tests, test that each step of psa_cipher_update produces output of the expected length. The length is hard-coded in the test data since it depends on the mode. The length of the output of psa_cipher_finish is effectively tested because it's the total output length minus the length produced by the update steps.
2019-02-19 19:44:00 +01:00
TEST_ASSERT( function_output_length == output2_length );
Cipher tests: calculate and verify the actual output size
2018-06-08 14:40:59 +02:00
total_output_length += function_output_length;
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_cipher_finish( &operation,
Fix output size calculations in cipher tests Some calls to psa_cipher_finish or psa_cipher_update append to a buffer. Several of these calls were not calculating the offset into the buffer or the remaining buffer size correctly. This did not lead to buffer overflows before because the buffer sizes were sufficiently large for our test inputs. This did not lead to incorrect output when the test was designed to append but actually wrote too early because all the existing test cases either have no output from finish (stream cipher) or have no output from update (CBC, with less than one block of input).
2019-02-19 19:05:33 +01:00
output + total_output_length,
output_buffer_size - total_output_length,
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
&function_output_length ) );
Cipher tests: calculate and verify the actual output size
2018-06-08 14:40:59 +02:00
total_output_length += function_output_length;
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_cipher_abort( &operation ) );
Reordered cipher tests to be just after MAC tests
2018-06-06 15:19:24 +02:00
Use ASSERT_COMPARE instead of memcmp in PSA tests This commit fixes some missing size comparison. In aead_encrypt_decrypt, aead_encrypt and aead_decrypt, the test code would not have noticed if the library function had reported an output length that was not the expected length.
2018-09-27 13:57:19 +02:00
ASSERT_COMPARE( expected_output->x, expected_output->len,
output, total_output_length );
PSA crypto: asymmetric signature (RSA PKCS#1v1.5 only) Define hash algorithms and RSA signature algorithms. New function psa_asymmetric_sign. Implement psa_asymmetric_sign for RSA PKCS#1 v1.5.
2018-02-03 22:44:14 +01:00
exit:
On target testing tests adaptation Updated all psa crypto tests to use the new test format
2018-06-12 16:06:52 +02:00
mbedtls_free( output );
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_destroy_key( handle );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
PSA crypto: asymmetric signature (RSA PKCS#1v1.5 only) Define hash algorithms and RSA signature algorithms. New function psa_asymmetric_sign. Implement psa_asymmetric_sign for RSA PKCS#1 v1.5.
2018-02-03 22:44:14 +01:00
}
/* END_CASE */
/* BEGIN_CASE */
We don't need _test_ in test function names Also fix typo multpart -> multipart
2018-06-08 14:28:46 +02:00
void cipher_decrypt_multipart( int alg_arg, int key_type_arg,
Cipher tests: pass the IV from the test data Don't hard-code an IV in cipher test functions. It restricts what can be used as test data.
2019-05-06 15:22:57 +02:00
data_t *key, data_t *iv,
On target testing tests adaptation Updated all psa crypto tests to use the new test format
2018-06-12 16:06:52 +02:00
data_t *input,
Test the length of cipher_update output In multipart cipher tests, test that each step of psa_cipher_update produces output of the expected length. The length is hard-coded in the test data since it depends on the mode. The length of the output of psa_cipher_finish is effectively tested because it's the total output length minus the length produced by the update steps.
2019-02-19 19:44:00 +01:00
int first_part_size_arg,
int output1_length_arg, int output2_length_arg,
On target testing tests adaptation Updated all psa crypto tests to use the new test format
2018-06-12 16:06:52 +02:00
data_t *expected_output )
PSA crypto: asymmetric signature (RSA PKCS#1v1.5 only) Define hash algorithms and RSA signature algorithms. New function psa_asymmetric_sign. Implement psa_asymmetric_sign for RSA PKCS#1 v1.5.
2018-02-03 22:44:14 +01:00
{
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t handle = 0;
Fix whitespace issues Only whitespace changes. * Remove tabs. * Remove trailing whitespace. * Correct some misindented lines. * Normalize whitespace around some punctuation. * Split some lines to avoid going over 80 columns.
2018-06-06 18:44:09 +02:00
PSA crypto: asymmetric signature (RSA PKCS#1v1.5 only) Define hash algorithms and RSA signature algorithms. New function psa_asymmetric_sign. Implement psa_asymmetric_sign for RSA PKCS#1 v1.5.
2018-02-03 22:44:14 +01:00
psa_key_type_t key_type = key_type_arg;
psa_algorithm_t alg = alg_arg;
Test the length of cipher_update output In multipart cipher tests, test that each step of psa_cipher_update produces output of the expected length. The length is hard-coded in the test data since it depends on the mode. The length of the output of psa_cipher_finish is effectively tested because it's the total output length minus the length produced by the update steps.
2019-02-19 19:44:00 +01:00
size_t first_part_size = first_part_size_arg;
size_t output1_length = output1_length_arg;
size_t output2_length = output2_length_arg;
On target testing tests adaptation Updated all psa crypto tests to use the new test format
2018-06-12 16:06:52 +02:00
unsigned char *output = NULL;
Rename some variables to make the code easier to read In cipher_test_verify_output_multpart, tweak the ways chunk sizes are added in order to get rid of the variable temp. In other functions, this commit does not change the logic at all.
2018-06-08 14:20:49 +02:00
size_t output_buffer_size = 0;
size_t function_output_length = 0;
Cipher tests: calculate and verify the actual output size
2018-06-08 14:40:59 +02:00
size_t total_output_length = 0;
psa: Add initializers for cipher operation objects Add new initializers for cipher operation objects and use them in our tests and library code. Prefer using the macro initializers due to their straightforwardness.
2019-01-04 12:48:27 +01:00
psa_cipher_operation_t operation = PSA_CIPHER_OPERATION_INIT;
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
PSA crypto: asymmetric signature (RSA PKCS#1v1.5 only) Define hash algorithms and RSA signature algorithms. New function psa_asymmetric_sign. Implement psa_asymmetric_sign for RSA PKCS#1 v1.5.
2018-02-03 22:44:14 +01:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
PSA crypto: asymmetric signature (RSA PKCS#1v1.5 only) Define hash algorithms and RSA signature algorithms. New function psa_asymmetric_sign. Implement psa_asymmetric_sign for RSA PKCS#1 v1.5.
2018-02-03 22:44:14 +01:00
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_set_key_usage_flags( &attributes, PSA_KEY_USAGE_DECRYPT );
psa_set_key_algorithm( &attributes, alg );
psa_set_key_type( &attributes, key_type );
Fix tests in test_suite_psa_crypto to set policy usage
2018-07-05 14:22:45 +02:00
Put handle parameter last: psa_import_key In psa_import_key, change the order of parameters to pass the pointer where the newly created handle will be stored last. This is consistent with most other library functions that put inputs before outputs.
2019-05-15 20:15:10 +02:00
PSA_ASSERT( psa_import_key( &attributes, key->x, key->len, &handle ) );
Add calls for set policy in export/sign tests Add calls for set policy in export/sign tests
2018-03-28 12:46:26 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_cipher_decrypt_setup( &operation,
handle, alg ) );
Add calls for set policy in export/sign tests Add calls for set policy in export/sign tests
2018-03-28 12:46:26 +02:00
Cipher tests: pass the IV from the test data Don't hard-code an IV in cipher test functions. It restricts what can be used as test data.
2019-05-06 15:22:57 +02:00
PSA_ASSERT( psa_cipher_set_iv( &operation, iv->x, iv->len ) );
Add calls for set policy in export/sign tests Add calls for set policy in export/sign tests
2018-03-28 12:46:26 +02:00
Wrap some multiline expressions in parentheses This guarantees that they'll be indented as desired under most indentation rules.
2018-12-17 23:34:57 +01:00
output_buffer_size = ( (size_t) input->len +
PSA_BLOCK_CIPHER_BLOCK_SIZE( key_type ) );
Use ASSERT_ALLOC instead of mbedtls_calloc in PSA tests This commit resolves a bug whereby some test cases failed on systems where mbedtls_calloc returns NULL when the size of 0, because the test case asserted `pointer != NULL` regardless of the size.
2018-09-27 13:54:18 +02:00
ASSERT_ALLOC( output, output_buffer_size );
PSA crypto: asymmetric signature (RSA PKCS#1v1.5 only) Define hash algorithms and RSA signature algorithms. New function psa_asymmetric_sign. Implement psa_asymmetric_sign for RSA PKCS#1 v1.5.
2018-02-03 22:44:14 +01:00
Test the length of cipher_update output In multipart cipher tests, test that each step of psa_cipher_update produces output of the expected length. The length is hard-coded in the test data since it depends on the mode. The length of the output of psa_cipher_finish is effectively tested because it's the total output length minus the length produced by the update steps.
2019-02-19 19:44:00 +01:00
TEST_ASSERT( first_part_size <= input->len );
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_cipher_update( &operation,
input->x, first_part_size,
output, output_buffer_size,
&function_output_length ) );
Test the length of cipher_update output In multipart cipher tests, test that each step of psa_cipher_update produces output of the expected length. The length is hard-coded in the test data since it depends on the mode. The length of the output of psa_cipher_finish is effectively tested because it's the total output length minus the length produced by the update steps.
2019-02-19 19:44:00 +01:00
TEST_ASSERT( function_output_length == output1_length );
Cipher tests: calculate and verify the actual output size
2018-06-08 14:40:59 +02:00
total_output_length += function_output_length;
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_cipher_update( &operation,
input->x + first_part_size,
input->len - first_part_size,
Fix output size calculations in cipher tests Some calls to psa_cipher_finish or psa_cipher_update append to a buffer. Several of these calls were not calculating the offset into the buffer or the remaining buffer size correctly. This did not lead to buffer overflows before because the buffer sizes were sufficiently large for our test inputs. This did not lead to incorrect output when the test was designed to append but actually wrote too early because all the existing test cases either have no output from finish (stream cipher) or have no output from update (CBC, with less than one block of input).
2019-02-19 19:05:33 +01:00
output + total_output_length,
output_buffer_size - total_output_length,
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
&function_output_length ) );
Test the length of cipher_update output In multipart cipher tests, test that each step of psa_cipher_update produces output of the expected length. The length is hard-coded in the test data since it depends on the mode. The length of the output of psa_cipher_finish is effectively tested because it's the total output length minus the length produced by the update steps.
2019-02-19 19:44:00 +01:00
TEST_ASSERT( function_output_length == output2_length );
Cipher tests: calculate and verify the actual output size
2018-06-08 14:40:59 +02:00
total_output_length += function_output_length;
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_cipher_finish( &operation,
Fix output size calculations in cipher tests Some calls to psa_cipher_finish or psa_cipher_update append to a buffer. Several of these calls were not calculating the offset into the buffer or the remaining buffer size correctly. This did not lead to buffer overflows before because the buffer sizes were sufficiently large for our test inputs. This did not lead to incorrect output when the test was designed to append but actually wrote too early because all the existing test cases either have no output from finish (stream cipher) or have no output from update (CBC, with less than one block of input).
2019-02-19 19:05:33 +01:00
output + total_output_length,
output_buffer_size - total_output_length,
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
&function_output_length ) );
Cipher tests: calculate and verify the actual output size
2018-06-08 14:40:59 +02:00
total_output_length += function_output_length;
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_cipher_abort( &operation ) );
Reordered cipher tests to be just after MAC tests
2018-06-06 15:19:24 +02:00
Use ASSERT_COMPARE instead of memcmp in PSA tests This commit fixes some missing size comparison. In aead_encrypt_decrypt, aead_encrypt and aead_decrypt, the test code would not have noticed if the library function had reported an output length that was not the expected length.
2018-09-27 13:57:19 +02:00
ASSERT_COMPARE( expected_output->x, expected_output->len,
output, total_output_length );
PSA crypto: asymmetric signature (RSA PKCS#1v1.5 only) Define hash algorithms and RSA signature algorithms. New function psa_asymmetric_sign. Implement psa_asymmetric_sign for RSA PKCS#1 v1.5.
2018-02-03 22:44:14 +01:00
exit:
On target testing tests adaptation Updated all psa crypto tests to use the new test format
2018-06-12 16:06:52 +02:00
mbedtls_free( output );
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_destroy_key( handle );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
PSA crypto: asymmetric signature (RSA PKCS#1v1.5 only) Define hash algorithms and RSA signature algorithms. New function psa_asymmetric_sign. Implement psa_asymmetric_sign for RSA PKCS#1 v1.5.
2018-02-03 22:44:14 +01:00
}
/* END_CASE */
Key Policy APIs implementation
2018-03-28 00:21:33 +02:00
/* BEGIN_CASE */
We don't need _test_ in test function names Also fix typo multpart -> multipart
2018-06-08 14:28:46 +02:00
void cipher_decrypt( int alg_arg, int key_type_arg,
Cipher tests: pass the IV from the test data Don't hard-code an IV in cipher test functions. It restricts what can be used as test data.
2019-05-06 15:22:57 +02:00
data_t *key, data_t *iv,
On target testing tests adaptation Updated all psa crypto tests to use the new test format
2018-06-12 16:06:52 +02:00
data_t *input, data_t *expected_output,
Get rid of some casts in test_suite_psa_crypto Use more auxiliary variables to unmarshall int values.
2018-06-21 09:25:10 +02:00
int expected_status_arg )
Key Policy APIs implementation
2018-03-28 00:21:33 +02:00
{
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t handle = 0;
fix and add tests case + fix for padding mode
2018-06-06 15:36:50 +02:00
psa_status_t status;
Reordered cipher tests to be just after MAC tests
2018-06-06 15:19:24 +02:00
psa_key_type_t key_type = key_type_arg;
psa_algorithm_t alg = alg_arg;
Get rid of some casts in test_suite_psa_crypto Use more auxiliary variables to unmarshall int values.
2018-06-21 09:25:10 +02:00
psa_status_t expected_status = expected_status_arg;
On target testing tests adaptation Updated all psa crypto tests to use the new test format
2018-06-12 16:06:52 +02:00
unsigned char *output = NULL;
Rename some variables to make the code easier to read In cipher_test_verify_output_multpart, tweak the ways chunk sizes are added in order to get rid of the variable temp. In other functions, this commit does not change the logic at all.
2018-06-08 14:20:49 +02:00
size_t output_buffer_size = 0;
size_t function_output_length = 0;
Cipher tests: calculate and verify the actual output size
2018-06-08 14:40:59 +02:00
size_t total_output_length = 0;
psa: Add initializers for cipher operation objects Add new initializers for cipher operation objects and use them in our tests and library code. Prefer using the macro initializers due to their straightforwardness.
2019-01-04 12:48:27 +01:00
psa_cipher_operation_t operation = PSA_CIPHER_OPERATION_INIT;
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
Merge remote-tracking branch 'psa/pr/13' into feature-psa Conflicts: library/psa_crypto.c tests/suites/test_suite_psa_crypto.data tests/suites/test_suite_psa_crypto.function All the conflicts are concurrent additions where the order doesn't matter. I put the code from feature-psa (key policy) before the code from PR #13 (key lifetime).
2018-04-30 17:06:50 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
Key Policy APIs implementation
2018-03-28 00:21:33 +02:00
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_set_key_usage_flags( &attributes, PSA_KEY_USAGE_DECRYPT );
psa_set_key_algorithm( &attributes, alg );
psa_set_key_type( &attributes, key_type );
Fix tests in test_suite_psa_crypto to set policy usage
2018-07-05 14:22:45 +02:00
Put handle parameter last: psa_import_key In psa_import_key, change the order of parameters to pass the pointer where the newly created handle will be stored last. This is consistent with most other library functions that put inputs before outputs.
2019-05-15 20:15:10 +02:00
PSA_ASSERT( psa_import_key( &attributes, key->x, key->len, &handle ) );
Key Policy APIs implementation
2018-03-28 00:21:33 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_cipher_decrypt_setup( &operation,
handle, alg ) );
Key Policy APIs implementation
2018-03-28 00:21:33 +02:00
Cipher tests: pass the IV from the test data Don't hard-code an IV in cipher test functions. It restricts what can be used as test data.
2019-05-06 15:22:57 +02:00
PSA_ASSERT( psa_cipher_set_iv( &operation, iv->x, iv->len ) );
Key Policy APIs implementation
2018-03-28 00:21:33 +02:00
Wrap some multiline expressions in parentheses This guarantees that they'll be indented as desired under most indentation rules.
2018-12-17 23:34:57 +01:00
output_buffer_size = ( (size_t) input->len +
PSA_BLOCK_CIPHER_BLOCK_SIZE( key_type ) );
Use ASSERT_ALLOC instead of mbedtls_calloc in PSA tests This commit resolves a bug whereby some test cases failed on systems where mbedtls_calloc returns NULL when the size of 0, because the test case asserted `pointer != NULL` regardless of the size.
2018-09-27 13:54:18 +02:00
ASSERT_ALLOC( output, output_buffer_size );
Reordered cipher tests to be just after MAC tests
2018-06-06 15:19:24 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_cipher_update( &operation,
input->x, input->len,
output, output_buffer_size,
&function_output_length ) );
Cipher tests: calculate and verify the actual output size
2018-06-08 14:40:59 +02:00
total_output_length += function_output_length;
Rename some variables to make the code easier to read In cipher_test_verify_output_multpart, tweak the ways chunk sizes are added in order to get rid of the variable temp. In other functions, this commit does not change the logic at all.
2018-06-08 14:20:49 +02:00
status = psa_cipher_finish( &operation,
Fix output size calculations in cipher tests Some calls to psa_cipher_finish or psa_cipher_update append to a buffer. Several of these calls were not calculating the offset into the buffer or the remaining buffer size correctly. This did not lead to buffer overflows before because the buffer sizes were sufficiently large for our test inputs. This did not lead to incorrect output when the test was designed to append but actually wrote too early because all the existing test cases either have no output from finish (stream cipher) or have no output from update (CBC, with less than one block of input).
2019-02-19 19:05:33 +01:00
output + total_output_length,
output_buffer_size - total_output_length,
Rename some variables to make the code easier to read In cipher_test_verify_output_multpart, tweak the ways chunk sizes are added in order to get rid of the variable temp. In other functions, this commit does not change the logic at all.
2018-06-08 14:20:49 +02:00
&function_output_length );
Cipher tests: calculate and verify the actual output size
2018-06-08 14:40:59 +02:00
total_output_length += function_output_length;
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( status, expected_status );
Reordered cipher tests to be just after MAC tests
2018-06-06 15:19:24 +02:00
fix and add tests case + fix for padding mode
2018-06-06 15:36:50 +02:00
if( expected_status == PSA_SUCCESS )
{
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_cipher_abort( &operation ) );
Use ASSERT_COMPARE instead of memcmp in PSA tests This commit fixes some missing size comparison. In aead_encrypt_decrypt, aead_encrypt and aead_decrypt, the test code would not have noticed if the library function had reported an output length that was not the expected length.
2018-09-27 13:57:19 +02:00
ASSERT_COMPARE( expected_output->x, expected_output->len,
output, total_output_length );
fix and add tests case + fix for padding mode
2018-06-06 15:36:50 +02:00
}
Merge remote-tracking branch 'psa/pr/13' into feature-psa Conflicts: library/psa_crypto.c tests/suites/test_suite_psa_crypto.data tests/suites/test_suite_psa_crypto.function All the conflicts are concurrent additions where the order doesn't matter. I put the code from feature-psa (key policy) before the code from PR #13 (key lifetime).
2018-04-30 17:06:50 +02:00
Key Policy APIs implementation
2018-03-28 00:21:33 +02:00
exit:
On target testing tests adaptation Updated all psa crypto tests to use the new test format
2018-06-12 16:06:52 +02:00
mbedtls_free( output );
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_destroy_key( handle );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
Key Policy APIs implementation
2018-03-28 00:21:33 +02:00
}
/* END_CASE */
add new test scenarios
2018-03-28 14:14:59 +02:00
/* BEGIN_CASE */
We don't need _test_ in test function names Also fix typo multpart -> multipart
2018-06-08 14:28:46 +02:00
void cipher_verify_output( int alg_arg, int key_type_arg,
On target testing tests adaptation Updated all psa crypto tests to use the new test format
2018-06-12 16:06:52 +02:00
data_t *key,
data_t *input )
add new test scenarios
2018-03-28 14:14:59 +02:00
{
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t handle = 0;
Reordered cipher tests to be just after MAC tests
2018-06-06 15:19:24 +02:00
psa_key_type_t key_type = key_type_arg;
psa_algorithm_t alg = alg_arg;
unsigned char iv[16] = {0};
size_t iv_size = 16;
size_t iv_length = 0;
On target testing tests adaptation Updated all psa crypto tests to use the new test format
2018-06-12 16:06:52 +02:00
unsigned char *output1 = NULL;
Reordered cipher tests to be just after MAC tests
2018-06-06 15:19:24 +02:00
size_t output1_size = 0;
size_t output1_length = 0;
On target testing tests adaptation Updated all psa crypto tests to use the new test format
2018-06-12 16:06:52 +02:00
unsigned char *output2 = NULL;
Reordered cipher tests to be just after MAC tests
2018-06-06 15:19:24 +02:00
size_t output2_size = 0;
size_t output2_length = 0;
Rename some variables to make the code easier to read In cipher_test_verify_output_multpart, tweak the ways chunk sizes are added in order to get rid of the variable temp. In other functions, this commit does not change the logic at all.
2018-06-08 14:20:49 +02:00
size_t function_output_length = 0;
psa: Add initializers for cipher operation objects Add new initializers for cipher operation objects and use them in our tests and library code. Prefer using the macro initializers due to their straightforwardness.
2019-01-04 12:48:27 +01:00
psa_cipher_operation_t operation1 = PSA_CIPHER_OPERATION_INIT;
psa_cipher_operation_t operation2 = PSA_CIPHER_OPERATION_INIT;
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
add new test scenarios
2018-03-28 14:14:59 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
Reordered cipher tests to be just after MAC tests
2018-06-06 15:19:24 +02:00
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_set_key_usage_flags( &attributes, PSA_KEY_USAGE_ENCRYPT | PSA_KEY_USAGE_DECRYPT );
psa_set_key_algorithm( &attributes, alg );
psa_set_key_type( &attributes, key_type );
Fix tests in test_suite_psa_crypto to set policy usage
2018-07-05 14:22:45 +02:00
Put handle parameter last: psa_import_key In psa_import_key, change the order of parameters to pass the pointer where the newly created handle will be stored last. This is consistent with most other library functions that put inputs before outputs.
2019-05-15 20:15:10 +02:00
PSA_ASSERT( psa_import_key( &attributes, key->x, key->len, &handle ) );
Reordered cipher tests to be just after MAC tests
2018-06-06 15:19:24 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_cipher_encrypt_setup( &operation1,
handle, alg ) );
PSA_ASSERT( psa_cipher_decrypt_setup( &operation2,
handle, alg ) );
Reordered cipher tests to be just after MAC tests
2018-06-06 15:19:24 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_cipher_generate_iv( &operation1,
iv, iv_size,
&iv_length ) );
Wrap some multiline expressions in parentheses This guarantees that they'll be indented as desired under most indentation rules.
2018-12-17 23:34:57 +01:00
output1_size = ( (size_t) input->len +
PSA_BLOCK_CIPHER_BLOCK_SIZE( key_type ) );
Use ASSERT_ALLOC instead of mbedtls_calloc in PSA tests This commit resolves a bug whereby some test cases failed on systems where mbedtls_calloc returns NULL when the size of 0, because the test case asserted `pointer != NULL` regardless of the size.
2018-09-27 13:54:18 +02:00
ASSERT_ALLOC( output1, output1_size );
fix and add tests case + fix for padding mode
2018-06-06 15:36:50 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_cipher_update( &operation1, input->x, input->len,
output1, output1_size,
&output1_length ) );
PSA_ASSERT( psa_cipher_finish( &operation1,
Fix output size calculations in cipher tests Some calls to psa_cipher_finish or psa_cipher_update append to a buffer. Several of these calls were not calculating the offset into the buffer or the remaining buffer size correctly. This did not lead to buffer overflows before because the buffer sizes were sufficiently large for our test inputs. This did not lead to incorrect output when the test was designed to append but actually wrote too early because all the existing test cases either have no output from finish (stream cipher) or have no output from update (CBC, with less than one block of input).
2019-02-19 19:05:33 +01:00
output1 + output1_length,
output1_size - output1_length,
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
&function_output_length ) );
Fix whitespace issues Only whitespace changes. * Remove tabs. * Remove trailing whitespace. * Correct some misindented lines. * Normalize whitespace around some punctuation. * Split some lines to avoid going over 80 columns.
2018-06-06 18:44:09 +02:00
Rename some variables to make the code easier to read In cipher_test_verify_output_multpart, tweak the ways chunk sizes are added in order to get rid of the variable temp. In other functions, this commit does not change the logic at all.
2018-06-08 14:20:49 +02:00
output1_length += function_output_length;
Reordered cipher tests to be just after MAC tests
2018-06-06 15:19:24 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_cipher_abort( &operation1 ) );
Reordered cipher tests to be just after MAC tests
2018-06-06 15:19:24 +02:00
output2_size = output1_length;
Use ASSERT_ALLOC instead of mbedtls_calloc in PSA tests This commit resolves a bug whereby some test cases failed on systems where mbedtls_calloc returns NULL when the size of 0, because the test case asserted `pointer != NULL` regardless of the size.
2018-09-27 13:54:18 +02:00
ASSERT_ALLOC( output2, output2_size );
Reordered cipher tests to be just after MAC tests
2018-06-06 15:19:24 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_cipher_set_iv( &operation2,
iv, iv_length ) );
PSA_ASSERT( psa_cipher_update( &operation2, output1, output1_length,
output2, output2_size,
&output2_length ) );
Rename some variables to make the code easier to read In cipher_test_verify_output_multpart, tweak the ways chunk sizes are added in order to get rid of the variable temp. In other functions, this commit does not change the logic at all.
2018-06-08 14:20:49 +02:00
function_output_length = 0;
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_cipher_finish( &operation2,
output2 + output2_length,
Fix output size calculations in cipher tests Some calls to psa_cipher_finish or psa_cipher_update append to a buffer. Several of these calls were not calculating the offset into the buffer or the remaining buffer size correctly. This did not lead to buffer overflows before because the buffer sizes were sufficiently large for our test inputs. This did not lead to incorrect output when the test was designed to append but actually wrote too early because all the existing test cases either have no output from finish (stream cipher) or have no output from update (CBC, with less than one block of input).
2019-02-19 19:05:33 +01:00
output2_size - output2_length,
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
&function_output_length ) );
Reordered cipher tests to be just after MAC tests
2018-06-06 15:19:24 +02:00
Rename some variables to make the code easier to read In cipher_test_verify_output_multpart, tweak the ways chunk sizes are added in order to get rid of the variable temp. In other functions, this commit does not change the logic at all.
2018-06-08 14:20:49 +02:00
output2_length += function_output_length;
Fix whitespace issues Only whitespace changes. * Remove tabs. * Remove trailing whitespace. * Correct some misindented lines. * Normalize whitespace around some punctuation. * Split some lines to avoid going over 80 columns.
2018-06-06 18:44:09 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_cipher_abort( &operation2 ) );
fix and add tests case + fix for padding mode
2018-06-06 15:36:50 +02:00
Use ASSERT_COMPARE instead of memcmp in PSA tests This commit fixes some missing size comparison. In aead_encrypt_decrypt, aead_encrypt and aead_decrypt, the test code would not have noticed if the library function had reported an output length that was not the expected length.
2018-09-27 13:57:19 +02:00
ASSERT_COMPARE( input->x, input->len, output2, output2_length );
fix and add tests case + fix for padding mode
2018-06-06 15:36:50 +02:00
exit:
On target testing tests adaptation Updated all psa crypto tests to use the new test format
2018-06-12 16:06:52 +02:00
mbedtls_free( output1 );
mbedtls_free( output2 );
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_destroy_key( handle );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
fix and add tests case + fix for padding mode
2018-06-06 15:36:50 +02:00
}
/* END_CASE */
/* BEGIN_CASE */
We don't need _test_ in test function names Also fix typo multpart -> multipart
2018-06-08 14:28:46 +02:00
void cipher_verify_output_multipart( int alg_arg,
int key_type_arg,
On target testing tests adaptation Updated all psa crypto tests to use the new test format
2018-06-12 16:06:52 +02:00
data_t *key,
data_t *input,
Test the length of cipher_update output In multipart cipher tests, test that each step of psa_cipher_update produces output of the expected length. The length is hard-coded in the test data since it depends on the mode. The length of the output of psa_cipher_finish is effectively tested because it's the total output length minus the length produced by the update steps.
2019-02-19 19:44:00 +01:00
int first_part_size_arg )
fix and add tests case + fix for padding mode
2018-06-06 15:36:50 +02:00
{
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t handle = 0;
fix and add tests case + fix for padding mode
2018-06-06 15:36:50 +02:00
psa_key_type_t key_type = key_type_arg;
psa_algorithm_t alg = alg_arg;
Test the length of cipher_update output In multipart cipher tests, test that each step of psa_cipher_update produces output of the expected length. The length is hard-coded in the test data since it depends on the mode. The length of the output of psa_cipher_finish is effectively tested because it's the total output length minus the length produced by the update steps.
2019-02-19 19:44:00 +01:00
size_t first_part_size = first_part_size_arg;
fix and add tests case + fix for padding mode
2018-06-06 15:36:50 +02:00
unsigned char iv[16] = {0};
size_t iv_size = 16;
size_t iv_length = 0;
On target testing tests adaptation Updated all psa crypto tests to use the new test format
2018-06-12 16:06:52 +02:00
unsigned char *output1 = NULL;
Rename some variables to make the code easier to read In cipher_test_verify_output_multpart, tweak the ways chunk sizes are added in order to get rid of the variable temp. In other functions, this commit does not change the logic at all.
2018-06-08 14:20:49 +02:00
size_t output1_buffer_size = 0;
fix and add tests case + fix for padding mode
2018-06-06 15:36:50 +02:00
size_t output1_length = 0;
On target testing tests adaptation Updated all psa crypto tests to use the new test format
2018-06-12 16:06:52 +02:00
unsigned char *output2 = NULL;
Rename some variables to make the code easier to read In cipher_test_verify_output_multpart, tweak the ways chunk sizes are added in order to get rid of the variable temp. In other functions, this commit does not change the logic at all.
2018-06-08 14:20:49 +02:00
size_t output2_buffer_size = 0;
fix and add tests case + fix for padding mode
2018-06-06 15:36:50 +02:00
size_t output2_length = 0;
Rename some variables to make the code easier to read In cipher_test_verify_output_multpart, tweak the ways chunk sizes are added in order to get rid of the variable temp. In other functions, this commit does not change the logic at all.
2018-06-08 14:20:49 +02:00
size_t function_output_length;
psa: Add initializers for cipher operation objects Add new initializers for cipher operation objects and use them in our tests and library code. Prefer using the macro initializers due to their straightforwardness.
2019-01-04 12:48:27 +01:00
psa_cipher_operation_t operation1 = PSA_CIPHER_OPERATION_INIT;
psa_cipher_operation_t operation2 = PSA_CIPHER_OPERATION_INIT;
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
fix and add tests case + fix for padding mode
2018-06-06 15:36:50 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
fix and add tests case + fix for padding mode
2018-06-06 15:36:50 +02:00
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_set_key_usage_flags( &attributes, PSA_KEY_USAGE_ENCRYPT | PSA_KEY_USAGE_DECRYPT );
psa_set_key_algorithm( &attributes, alg );
psa_set_key_type( &attributes, key_type );
Fix tests in test_suite_psa_crypto to set policy usage
2018-07-05 14:22:45 +02:00
Put handle parameter last: psa_import_key In psa_import_key, change the order of parameters to pass the pointer where the newly created handle will be stored last. This is consistent with most other library functions that put inputs before outputs.
2019-05-15 20:15:10 +02:00
PSA_ASSERT( psa_import_key( &attributes, key->x, key->len, &handle ) );
fix and add tests case + fix for padding mode
2018-06-06 15:36:50 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_cipher_encrypt_setup( &operation1,
handle, alg ) );
PSA_ASSERT( psa_cipher_decrypt_setup( &operation2,
handle, alg ) );
fix and add tests case + fix for padding mode
2018-06-06 15:36:50 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_cipher_generate_iv( &operation1,
iv, iv_size,
&iv_length ) );
Wrap some multiline expressions in parentheses This guarantees that they'll be indented as desired under most indentation rules.
2018-12-17 23:34:57 +01:00
output1_buffer_size = ( (size_t) input->len +
PSA_BLOCK_CIPHER_BLOCK_SIZE( key_type ) );
Use ASSERT_ALLOC instead of mbedtls_calloc in PSA tests This commit resolves a bug whereby some test cases failed on systems where mbedtls_calloc returns NULL when the size of 0, because the test case asserted `pointer != NULL` regardless of the size.
2018-09-27 13:54:18 +02:00
ASSERT_ALLOC( output1, output1_buffer_size );
fix and add tests case + fix for padding mode
2018-06-06 15:36:50 +02:00
Test the length of cipher_update output In multipart cipher tests, test that each step of psa_cipher_update produces output of the expected length. The length is hard-coded in the test data since it depends on the mode. The length of the output of psa_cipher_finish is effectively tested because it's the total output length minus the length produced by the update steps.
2019-02-19 19:44:00 +01:00
TEST_ASSERT( first_part_size <= input->len );
fix and add tests case + fix for padding mode
2018-06-06 15:36:50 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_cipher_update( &operation1, input->x, first_part_size,
output1, output1_buffer_size,
&function_output_length ) );
Rename some variables to make the code easier to read In cipher_test_verify_output_multpart, tweak the ways chunk sizes are added in order to get rid of the variable temp. In other functions, this commit does not change the logic at all.
2018-06-08 14:20:49 +02:00
output1_length += function_output_length;
Fix whitespace issues Only whitespace changes. * Remove tabs. * Remove trailing whitespace. * Correct some misindented lines. * Normalize whitespace around some punctuation. * Split some lines to avoid going over 80 columns.
2018-06-06 18:44:09 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_cipher_update( &operation1,
input->x + first_part_size,
input->len - first_part_size,
output1, output1_buffer_size,
&function_output_length ) );
Rename some variables to make the code easier to read In cipher_test_verify_output_multpart, tweak the ways chunk sizes are added in order to get rid of the variable temp. In other functions, this commit does not change the logic at all.
2018-06-08 14:20:49 +02:00
output1_length += function_output_length;
fix and add tests case + fix for padding mode
2018-06-06 15:36:50 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_cipher_finish( &operation1,
output1 + output1_length,
output1_buffer_size - output1_length,
&function_output_length ) );
Rename some variables to make the code easier to read In cipher_test_verify_output_multpart, tweak the ways chunk sizes are added in order to get rid of the variable temp. In other functions, this commit does not change the logic at all.
2018-06-08 14:20:49 +02:00
output1_length += function_output_length;
Reordered cipher tests to be just after MAC tests
2018-06-06 15:19:24 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_cipher_abort( &operation1 ) );
Reordered cipher tests to be just after MAC tests
2018-06-06 15:19:24 +02:00
Rename some variables to make the code easier to read In cipher_test_verify_output_multpart, tweak the ways chunk sizes are added in order to get rid of the variable temp. In other functions, this commit does not change the logic at all.
2018-06-08 14:20:49 +02:00
output2_buffer_size = output1_length;
Use ASSERT_ALLOC instead of mbedtls_calloc in PSA tests This commit resolves a bug whereby some test cases failed on systems where mbedtls_calloc returns NULL when the size of 0, because the test case asserted `pointer != NULL` regardless of the size.
2018-09-27 13:54:18 +02:00
ASSERT_ALLOC( output2, output2_buffer_size );
fix and add tests case + fix for padding mode
2018-06-06 15:36:50 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_cipher_set_iv( &operation2,
iv, iv_length ) );
fix and add tests case + fix for padding mode
2018-06-06 15:36:50 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_cipher_update( &operation2, output1, first_part_size,
output2, output2_buffer_size,
&function_output_length ) );
Rename some variables to make the code easier to read In cipher_test_verify_output_multpart, tweak the ways chunk sizes are added in order to get rid of the variable temp. In other functions, this commit does not change the logic at all.
2018-06-08 14:20:49 +02:00
output2_length += function_output_length;
fix and add tests case + fix for padding mode
2018-06-06 15:36:50 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_cipher_update( &operation2,
output1 + first_part_size,
output1_length - first_part_size,
output2, output2_buffer_size,
&function_output_length ) );
Rename some variables to make the code easier to read In cipher_test_verify_output_multpart, tweak the ways chunk sizes are added in order to get rid of the variable temp. In other functions, this commit does not change the logic at all.
2018-06-08 14:20:49 +02:00
output2_length += function_output_length;
fix and add tests case + fix for padding mode
2018-06-06 15:36:50 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_cipher_finish( &operation2,
output2 + output2_length,
output2_buffer_size - output2_length,
&function_output_length ) );
Rename some variables to make the code easier to read In cipher_test_verify_output_multpart, tweak the ways chunk sizes are added in order to get rid of the variable temp. In other functions, this commit does not change the logic at all.
2018-06-08 14:20:49 +02:00
output2_length += function_output_length;
Fix whitespace issues Only whitespace changes. * Remove tabs. * Remove trailing whitespace. * Correct some misindented lines. * Normalize whitespace around some punctuation. * Split some lines to avoid going over 80 columns.
2018-06-06 18:44:09 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_cipher_abort( &operation2 ) );
fix and add tests case + fix for padding mode
2018-06-06 15:36:50 +02:00
Use ASSERT_COMPARE instead of memcmp in PSA tests This commit fixes some missing size comparison. In aead_encrypt_decrypt, aead_encrypt and aead_decrypt, the test code would not have noticed if the library function had reported an output length that was not the expected length.
2018-09-27 13:57:19 +02:00
ASSERT_COMPARE( input->x, input->len, output2, output2_length );
add new test scenarios
2018-03-28 14:14:59 +02:00
exit:
On target testing tests adaptation Updated all psa crypto tests to use the new test format
2018-06-12 16:06:52 +02:00
mbedtls_free( output1 );
mbedtls_free( output2 );
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_destroy_key( handle );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
add new test scenarios
2018-03-28 14:14:59 +02:00
}
/* END_CASE */
Merge remote-tracking branch 'psa/pr/13' into feature-psa Conflicts: library/psa_crypto.c tests/suites/test_suite_psa_crypto.data tests/suites/test_suite_psa_crypto.function All the conflicts are concurrent additions where the order doesn't matter. I put the code from feature-psa (key policy) before the code from PR #13 (key lifetime).
2018-04-30 17:06:50 +02:00
Move AEAD tests just after cipher Always adding things at the end tends to create merge conflicts. Adding in the middle in this way makes the order more logical in addition to avoiding conflicts.
2018-06-11 19:33:02 +02:00
/* BEGIN_CASE */
Reorder parameters of AEAD unit tests to be more logical Pass the nonce first, then the AD, then the input. This is the order in which the data is processed and it's the order of the parameters to the API functions.
2018-08-17 18:45:42 +02:00
void aead_encrypt_decrypt( int key_type_arg, data_t *key_data,
Formatting improvements Avoid lines longer than 80 columns. Remove some redundant parentheses, e.g. change if( ( a == b ) && ( c == d ) ) to if( a == b && c == d ) which makes lines less long and makes the remaining parentheses more relevant. Add missing parentheses around return statements. There should be no semantic change in this commit.
2018-06-18 16:04:39 +02:00
int alg_arg,
Reorder parameters of AEAD unit tests to be more logical Pass the nonce first, then the AD, then the input. This is the order in which the data is processed and it's the order of the parameters to the API functions.
2018-08-17 18:45:42 +02:00
data_t *nonce,
data_t *additional_data,
data_t *input_data,
Formatting improvements Avoid lines longer than 80 columns. Remove some redundant parentheses, e.g. change if( ( a == b ) && ( c == d ) ) to if( a == b && c == d ) which makes lines less long and makes the remaining parentheses more relevant. Add missing parentheses around return statements. There should be no semantic change in this commit.
2018-06-18 16:04:39 +02:00
int expected_result_arg )
Move AEAD tests just after cipher Always adding things at the end tends to create merge conflicts. Adding in the middle in this way makes the order more logical in addition to avoiding conflicts.
2018-06-11 19:33:02 +02:00
{
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t handle = 0;
Move AEAD tests just after cipher Always adding things at the end tends to create merge conflicts. Adding in the middle in this way makes the order more logical in addition to avoiding conflicts.
2018-06-11 19:33:02 +02:00
psa_key_type_t key_type = key_type_arg;
psa_algorithm_t alg = alg_arg;
unsigned char *output_data = NULL;
size_t output_size = 0;
size_t output_length = 0;
unsigned char *output_data2 = NULL;
size_t output_length2 = 0;
Use PSA_AEAD_{ENCRYPT,DECRYPT}_OUTPUT_SIZE in tests
2019-05-14 16:09:40 +02:00
size_t tag_length = PSA_AEAD_TAG_LENGTH( alg );
Hygiene improvements in PSA crypto test code Get rid of many redundant casts. In particular, it is not useful to cast uint32_t values to size_t before performing arithmetic or comparisons on them. Rewrap a number of function calls, many of which now have narrower arguments thanks to the removed casts. When a function call doesn't fit on a single line, avoid grouping unrelated parameters together, but do try to group a buffer pointer and the associated size. Define more auxiliary variables xxx of a particular integer type (psa_algorithm_t, psa_key_usage_t, etc.) corresponding to a test function xxx_arg which has the type int. This avoids the need to cast xxx_arg to an unsigned type sometimes in the code.
2018-06-18 16:35:34 +02:00
psa_status_t expected_result = expected_result_arg;
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
Move AEAD tests just after cipher Always adding things at the end tends to create merge conflicts. Adding in the middle in this way makes the order more logical in addition to avoiding conflicts.
2018-06-11 19:33:02 +02:00
Hygiene improvements in PSA crypto test code Get rid of many redundant casts. In particular, it is not useful to cast uint32_t values to size_t before performing arithmetic or comparisons on them. Rewrap a number of function calls, many of which now have narrower arguments thanks to the removed casts. When a function call doesn't fit on a single line, avoid grouping unrelated parameters together, but do try to group a buffer pointer and the associated size. Define more auxiliary variables xxx of a particular integer type (psa_algorithm_t, psa_key_usage_t, etc.) corresponding to a test function xxx_arg which has the type int. This avoids the need to cast xxx_arg to an unsigned type sometimes in the code.
2018-06-18 16:35:34 +02:00
output_size = input_data->len + tag_length;
Use PSA_AEAD_{ENCRYPT,DECRYPT}_OUTPUT_SIZE in tests
2019-05-14 16:09:40 +02:00
/* For all currently defined algorithms, PSA_AEAD_ENCRYPT_OUTPUT_SIZE
* should be exact. */
if( expected_result != PSA_ERROR_INVALID_ARGUMENT )
TEST_EQUAL( output_size,
PSA_AEAD_ENCRYPT_OUTPUT_SIZE( alg, input_data->len ) );
Use ASSERT_ALLOC instead of mbedtls_calloc in PSA tests This commit resolves a bug whereby some test cases failed on systems where mbedtls_calloc returns NULL when the size of 0, because the test case asserted `pointer != NULL` regardless of the size.
2018-09-27 13:54:18 +02:00
ASSERT_ALLOC( output_data, output_size );
Move AEAD tests just after cipher Always adding things at the end tends to create merge conflicts. Adding in the middle in this way makes the order more logical in addition to avoiding conflicts.
2018-06-11 19:33:02 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
Move AEAD tests just after cipher Always adding things at the end tends to create merge conflicts. Adding in the middle in this way makes the order more logical in addition to avoiding conflicts.
2018-06-11 19:33:02 +02:00
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_set_key_usage_flags( &attributes, PSA_KEY_USAGE_ENCRYPT | PSA_KEY_USAGE_DECRYPT );
psa_set_key_algorithm( &attributes, alg );
psa_set_key_type( &attributes, key_type );
Move AEAD tests just after cipher Always adding things at the end tends to create merge conflicts. Adding in the middle in this way makes the order more logical in addition to avoiding conflicts.
2018-06-11 19:33:02 +02:00
Split long lines after psa_import_key refactoring
2019-05-15 20:22:09 +02:00
PSA_ASSERT( psa_import_key( &attributes, key_data->x, key_data->len,
&handle ) );
Move AEAD tests just after cipher Always adding things at the end tends to create merge conflicts. Adding in the middle in this way makes the order more logical in addition to avoiding conflicts.
2018-06-11 19:33:02 +02:00
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( psa_aead_encrypt( handle, alg,
nonce->x, nonce->len,
additional_data->x,
additional_data->len,
input_data->x, input_data->len,
output_data, output_size,
Rewrap some lines after the macro changes Change the way some lines are wrapped to cut at a more logical place. This commit mainly rewrites multi-line calls to TEST_EQUAL, and also a few calls to PSA_ASSERT.
2018-12-18 00:33:25 +01:00
&output_length ),
expected_result );
Move AEAD tests just after cipher Always adding things at the end tends to create merge conflicts. Adding in the middle in this way makes the order more logical in addition to avoiding conflicts.
2018-06-11 19:33:02 +02:00
if( PSA_SUCCESS == expected_result )
{
Use ASSERT_ALLOC instead of mbedtls_calloc in PSA tests This commit resolves a bug whereby some test cases failed on systems where mbedtls_calloc returns NULL when the size of 0, because the test case asserted `pointer != NULL` regardless of the size.
2018-09-27 13:54:18 +02:00
ASSERT_ALLOC( output_data2, output_length );
Move AEAD tests just after cipher Always adding things at the end tends to create merge conflicts. Adding in the middle in this way makes the order more logical in addition to avoiding conflicts.
2018-06-11 19:33:02 +02:00
Use PSA_AEAD_{ENCRYPT,DECRYPT}_OUTPUT_SIZE in tests
2019-05-14 16:09:40 +02:00
/* For all currently defined algorithms, PSA_AEAD_DECRYPT_OUTPUT_SIZE
* should be exact. */
TEST_EQUAL( input_data->len,
PSA_AEAD_DECRYPT_OUTPUT_SIZE( alg, output_length ) );
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( psa_aead_decrypt( handle, alg,
nonce->x, nonce->len,
additional_data->x,
additional_data->len,
output_data, output_length,
output_data2, output_length,
Rewrap some lines after the macro changes Change the way some lines are wrapped to cut at a more logical place. This commit mainly rewrites multi-line calls to TEST_EQUAL, and also a few calls to PSA_ASSERT.
2018-12-18 00:33:25 +01:00
&output_length2 ),
expected_result );
Normalize whitespace Normalize whitespace to Mbed TLS standards. There are only whitespace changes in this commit.
2018-06-18 15:41:12 +02:00
Use ASSERT_COMPARE instead of memcmp in PSA tests This commit fixes some missing size comparison. In aead_encrypt_decrypt, aead_encrypt and aead_decrypt, the test code would not have noticed if the library function had reported an output length that was not the expected length.
2018-09-27 13:57:19 +02:00
ASSERT_COMPARE( input_data->x, input_data->len,
output_data2, output_length2 );
Move AEAD tests just after cipher Always adding things at the end tends to create merge conflicts. Adding in the middle in this way makes the order more logical in addition to avoiding conflicts.
2018-06-11 19:33:02 +02:00
}
Normalize whitespace Normalize whitespace to Mbed TLS standards. There are only whitespace changes in this commit.
2018-06-18 15:41:12 +02:00
Move AEAD tests just after cipher Always adding things at the end tends to create merge conflicts. Adding in the middle in this way makes the order more logical in addition to avoiding conflicts.
2018-06-11 19:33:02 +02:00
exit:
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_destroy_key( handle );
Move AEAD tests just after cipher Always adding things at the end tends to create merge conflicts. Adding in the middle in this way makes the order more logical in addition to avoiding conflicts.
2018-06-11 19:33:02 +02:00
mbedtls_free( output_data );
mbedtls_free( output_data2 );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
Move AEAD tests just after cipher Always adding things at the end tends to create merge conflicts. Adding in the middle in this way makes the order more logical in addition to avoiding conflicts.
2018-06-11 19:33:02 +02:00
}
/* END_CASE */
/* BEGIN_CASE */
Reorder parameters of AEAD unit tests to be more logical Pass the nonce first, then the AD, then the input. This is the order in which the data is processed and it's the order of the parameters to the API functions.
2018-08-17 18:45:42 +02:00
void aead_encrypt( int key_type_arg, data_t *key_data,
int alg_arg,
data_t *nonce,
data_t *additional_data,
data_t *input_data,
data_t *expected_result )
Move AEAD tests just after cipher Always adding things at the end tends to create merge conflicts. Adding in the middle in this way makes the order more logical in addition to avoiding conflicts.
2018-06-11 19:33:02 +02:00
{
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t handle = 0;
Move AEAD tests just after cipher Always adding things at the end tends to create merge conflicts. Adding in the middle in this way makes the order more logical in addition to avoiding conflicts.
2018-06-11 19:33:02 +02:00
psa_key_type_t key_type = key_type_arg;
psa_algorithm_t alg = alg_arg;
unsigned char *output_data = NULL;
size_t output_size = 0;
size_t output_length = 0;
Use PSA_AEAD_{ENCRYPT,DECRYPT}_OUTPUT_SIZE in tests
2019-05-14 16:09:40 +02:00
size_t tag_length = PSA_AEAD_TAG_LENGTH( alg );
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
Move AEAD tests just after cipher Always adding things at the end tends to create merge conflicts. Adding in the middle in this way makes the order more logical in addition to avoiding conflicts.
2018-06-11 19:33:02 +02:00
Hygiene improvements in PSA crypto test code Get rid of many redundant casts. In particular, it is not useful to cast uint32_t values to size_t before performing arithmetic or comparisons on them. Rewrap a number of function calls, many of which now have narrower arguments thanks to the removed casts. When a function call doesn't fit on a single line, avoid grouping unrelated parameters together, but do try to group a buffer pointer and the associated size. Define more auxiliary variables xxx of a particular integer type (psa_algorithm_t, psa_key_usage_t, etc.) corresponding to a test function xxx_arg which has the type int. This avoids the need to cast xxx_arg to an unsigned type sometimes in the code.
2018-06-18 16:35:34 +02:00
output_size = input_data->len + tag_length;
Use PSA_AEAD_{ENCRYPT,DECRYPT}_OUTPUT_SIZE in tests
2019-05-14 16:09:40 +02:00
/* For all currently defined algorithms, PSA_AEAD_ENCRYPT_OUTPUT_SIZE
* should be exact. */
TEST_EQUAL( output_size,
PSA_AEAD_ENCRYPT_OUTPUT_SIZE( alg, input_data->len ) );
Use ASSERT_ALLOC instead of mbedtls_calloc in PSA tests This commit resolves a bug whereby some test cases failed on systems where mbedtls_calloc returns NULL when the size of 0, because the test case asserted `pointer != NULL` regardless of the size.
2018-09-27 13:54:18 +02:00
ASSERT_ALLOC( output_data, output_size );
Normalize whitespace Normalize whitespace to Mbed TLS standards. There are only whitespace changes in this commit.
2018-06-18 15:41:12 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
Move AEAD tests just after cipher Always adding things at the end tends to create merge conflicts. Adding in the middle in this way makes the order more logical in addition to avoiding conflicts.
2018-06-11 19:33:02 +02:00
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_set_key_usage_flags( &attributes, PSA_KEY_USAGE_ENCRYPT );
psa_set_key_algorithm( &attributes, alg );
psa_set_key_type( &attributes, key_type );
Move AEAD tests just after cipher Always adding things at the end tends to create merge conflicts. Adding in the middle in this way makes the order more logical in addition to avoiding conflicts.
2018-06-11 19:33:02 +02:00
Split long lines after psa_import_key refactoring
2019-05-15 20:22:09 +02:00
PSA_ASSERT( psa_import_key( &attributes, key_data->x, key_data->len,
&handle ) );
Move AEAD tests just after cipher Always adding things at the end tends to create merge conflicts. Adding in the middle in this way makes the order more logical in addition to avoiding conflicts.
2018-06-11 19:33:02 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_aead_encrypt( handle, alg,
nonce->x, nonce->len,
additional_data->x, additional_data->len,
input_data->x, input_data->len,
output_data, output_size,
&output_length ) );
Move AEAD tests just after cipher Always adding things at the end tends to create merge conflicts. Adding in the middle in this way makes the order more logical in addition to avoiding conflicts.
2018-06-11 19:33:02 +02:00
Use ASSERT_COMPARE instead of memcmp in PSA tests This commit fixes some missing size comparison. In aead_encrypt_decrypt, aead_encrypt and aead_decrypt, the test code would not have noticed if the library function had reported an output length that was not the expected length.
2018-09-27 13:57:19 +02:00
ASSERT_COMPARE( expected_result->x, expected_result->len,
output_data, output_length );
Normalize whitespace Normalize whitespace to Mbed TLS standards. There are only whitespace changes in this commit.
2018-06-18 15:41:12 +02:00
Move AEAD tests just after cipher Always adding things at the end tends to create merge conflicts. Adding in the middle in this way makes the order more logical in addition to avoiding conflicts.
2018-06-11 19:33:02 +02:00
exit:
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_destroy_key( handle );
Move AEAD tests just after cipher Always adding things at the end tends to create merge conflicts. Adding in the middle in this way makes the order more logical in addition to avoiding conflicts.
2018-06-11 19:33:02 +02:00
mbedtls_free( output_data );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
Move AEAD tests just after cipher Always adding things at the end tends to create merge conflicts. Adding in the middle in this way makes the order more logical in addition to avoiding conflicts.
2018-06-11 19:33:02 +02:00
}
/* END_CASE */
/* BEGIN_CASE */
Reorder parameters of AEAD unit tests to be more logical Pass the nonce first, then the AD, then the input. This is the order in which the data is processed and it's the order of the parameters to the API functions.
2018-08-17 18:45:42 +02:00
void aead_decrypt( int key_type_arg, data_t *key_data,
int alg_arg,
data_t *nonce,
data_t *additional_data,
data_t *input_data,
data_t *expected_data,
int expected_result_arg )
Move AEAD tests just after cipher Always adding things at the end tends to create merge conflicts. Adding in the middle in this way makes the order more logical in addition to avoiding conflicts.
2018-06-11 19:33:02 +02:00
{
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t handle = 0;
Move AEAD tests just after cipher Always adding things at the end tends to create merge conflicts. Adding in the middle in this way makes the order more logical in addition to avoiding conflicts.
2018-06-11 19:33:02 +02:00
psa_key_type_t key_type = key_type_arg;
psa_algorithm_t alg = alg_arg;
unsigned char *output_data = NULL;
size_t output_size = 0;
size_t output_length = 0;
Use PSA_AEAD_{ENCRYPT,DECRYPT}_OUTPUT_SIZE in tests
2019-05-14 16:09:40 +02:00
size_t tag_length = PSA_AEAD_TAG_LENGTH( alg );
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
Hygiene improvements in PSA crypto test code Get rid of many redundant casts. In particular, it is not useful to cast uint32_t values to size_t before performing arithmetic or comparisons on them. Rewrap a number of function calls, many of which now have narrower arguments thanks to the removed casts. When a function call doesn't fit on a single line, avoid grouping unrelated parameters together, but do try to group a buffer pointer and the associated size. Define more auxiliary variables xxx of a particular integer type (psa_algorithm_t, psa_key_usage_t, etc.) corresponding to a test function xxx_arg which has the type int. This avoids the need to cast xxx_arg to an unsigned type sometimes in the code.
2018-06-18 16:35:34 +02:00
psa_status_t expected_result = expected_result_arg;
Move AEAD tests just after cipher Always adding things at the end tends to create merge conflicts. Adding in the middle in this way makes the order more logical in addition to avoiding conflicts.
2018-06-11 19:33:02 +02:00
Use PSA_AEAD_{ENCRYPT,DECRYPT}_OUTPUT_SIZE in tests
2019-05-14 16:09:40 +02:00
output_size = input_data->len - tag_length;
/* For all currently defined algorithms, PSA_AEAD_DECRYPT_OUTPUT_SIZE
* should be exact. */
if( expected_result != PSA_ERROR_INVALID_ARGUMENT )
TEST_EQUAL( output_size,
PSA_AEAD_DECRYPT_OUTPUT_SIZE( alg, input_data->len ) );
Use ASSERT_ALLOC instead of mbedtls_calloc in PSA tests This commit resolves a bug whereby some test cases failed on systems where mbedtls_calloc returns NULL when the size of 0, because the test case asserted `pointer != NULL` regardless of the size.
2018-09-27 13:54:18 +02:00
ASSERT_ALLOC( output_data, output_size );
Normalize whitespace Normalize whitespace to Mbed TLS standards. There are only whitespace changes in this commit.
2018-06-18 15:41:12 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
Move AEAD tests just after cipher Always adding things at the end tends to create merge conflicts. Adding in the middle in this way makes the order more logical in addition to avoiding conflicts.
2018-06-11 19:33:02 +02:00
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_set_key_usage_flags( &attributes, PSA_KEY_USAGE_DECRYPT );
psa_set_key_algorithm( &attributes, alg );
psa_set_key_type( &attributes, key_type );
Move AEAD tests just after cipher Always adding things at the end tends to create merge conflicts. Adding in the middle in this way makes the order more logical in addition to avoiding conflicts.
2018-06-11 19:33:02 +02:00
Split long lines after psa_import_key refactoring
2019-05-15 20:22:09 +02:00
PSA_ASSERT( psa_import_key( &attributes, key_data->x, key_data->len,
&handle ) );
Move AEAD tests just after cipher Always adding things at the end tends to create merge conflicts. Adding in the middle in this way makes the order more logical in addition to avoiding conflicts.
2018-06-11 19:33:02 +02:00
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( psa_aead_decrypt( handle, alg,
nonce->x, nonce->len,
additional_data->x,
additional_data->len,
input_data->x, input_data->len,
output_data, output_size,
Rewrap some lines after the macro changes Change the way some lines are wrapped to cut at a more logical place. This commit mainly rewrites multi-line calls to TEST_EQUAL, and also a few calls to PSA_ASSERT.
2018-12-18 00:33:25 +01:00
&output_length ),
expected_result );
Move AEAD tests just after cipher Always adding things at the end tends to create merge conflicts. Adding in the middle in this way makes the order more logical in addition to avoiding conflicts.
2018-06-11 19:33:02 +02:00
Normalize whitespace Normalize whitespace to Mbed TLS standards. There are only whitespace changes in this commit.
2018-06-18 15:41:12 +02:00
if( expected_result == PSA_SUCCESS )
Use ASSERT_COMPARE instead of memcmp in PSA tests This commit fixes some missing size comparison. In aead_encrypt_decrypt, aead_encrypt and aead_decrypt, the test code would not have noticed if the library function had reported an output length that was not the expected length.
2018-09-27 13:57:19 +02:00
ASSERT_COMPARE( expected_data->x, expected_data->len,
output_data, output_length );
Move AEAD tests just after cipher Always adding things at the end tends to create merge conflicts. Adding in the middle in this way makes the order more logical in addition to avoiding conflicts.
2018-06-11 19:33:02 +02:00
exit:
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_destroy_key( handle );
Move AEAD tests just after cipher Always adding things at the end tends to create merge conflicts. Adding in the middle in this way makes the order more logical in addition to avoiding conflicts.
2018-06-11 19:33:02 +02:00
mbedtls_free( output_data );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
Move AEAD tests just after cipher Always adding things at the end tends to create merge conflicts. Adding in the middle in this way makes the order more logical in addition to avoiding conflicts.
2018-06-11 19:33:02 +02:00
}
/* END_CASE */
PSA crypto: PSA_ASYMMETRIC_SIGN_OUTPUT_SIZE macro Test it for RSA.
2018-02-03 23:57:22 +01:00
/* BEGIN_CASE */
Formatting improvements Avoid lines longer than 80 columns. Remove some redundant parentheses, e.g. change if( ( a == b ) && ( c == d ) ) to if( a == b && c == d ) which makes lines less long and makes the remaining parentheses more relevant. Add missing parentheses around return statements. There should be no semantic change in this commit.
2018-06-18 16:04:39 +02:00
void signature_size( int type_arg,
int bits,
int alg_arg,
int expected_size_arg )
PSA crypto: PSA_ASYMMETRIC_SIGN_OUTPUT_SIZE macro Test it for RSA.
2018-02-03 23:57:22 +01:00
{
psa_key_type_t type = type_arg;
psa_algorithm_t alg = alg_arg;
Normalize whitespace Normalize whitespace to Mbed TLS standards. There are only whitespace changes in this commit.
2018-06-18 15:41:12 +02:00
size_t actual_size = PSA_ASYMMETRIC_SIGN_OUTPUT_SIZE( type, bits, alg );
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( actual_size, (size_t) expected_size_arg );
PSA crypto: PSA_ASYMMETRIC_SIGN_OUTPUT_SIZE macro Test it for RSA.
2018-02-03 23:57:22 +01:00
exit:
;
}
/* END_CASE */
PSA crypto: asymmetric signature (RSA PKCS#1v1.5 only) Define hash algorithms and RSA signature algorithms. New function psa_asymmetric_sign. Implement psa_asymmetric_sign for RSA PKCS#1 v1.5.
2018-02-03 22:44:14 +01:00
/* BEGIN_CASE */
On target testing tests adaptation Updated all psa crypto tests to use the new test format
2018-06-12 16:06:52 +02:00
void sign_deterministic( int key_type_arg, data_t *key_data,
int alg_arg, data_t *input_data,
data_t *output_data )
PSA crypto: asymmetric signature (RSA PKCS#1v1.5 only) Define hash algorithms and RSA signature algorithms. New function psa_asymmetric_sign. Implement psa_asymmetric_sign for RSA PKCS#1 v1.5.
2018-02-03 22:44:14 +01:00
{
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t handle = 0;
PSA crypto: asymmetric signature (RSA PKCS#1v1.5 only) Define hash algorithms and RSA signature algorithms. New function psa_asymmetric_sign. Implement psa_asymmetric_sign for RSA PKCS#1 v1.5.
2018-02-03 22:44:14 +01:00
psa_key_type_t key_type = key_type_arg;
psa_algorithm_t alg = alg_arg;
PSA crypto: PSA_ASYMMETRIC_SIGN_OUTPUT_SIZE macro Test it for RSA.
2018-02-03 23:57:22 +01:00
size_t key_bits;
unsigned char *signature = NULL;
size_t signature_size;
PSA asymmetric signature: set *signature_length = 0 on failure
2018-02-03 23:58:03 +01:00
size_t signature_length = 0xdeadbeef;
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
PSA crypto: asymmetric signature (RSA PKCS#1v1.5 only) Define hash algorithms and RSA signature algorithms. New function psa_asymmetric_sign. Implement psa_asymmetric_sign for RSA PKCS#1 v1.5.
2018-02-03 22:44:14 +01:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
PSA crypto: asymmetric signature (RSA PKCS#1v1.5 only) Define hash algorithms and RSA signature algorithms. New function psa_asymmetric_sign. Implement psa_asymmetric_sign for RSA PKCS#1 v1.5.
2018-02-03 22:44:14 +01:00
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_set_key_usage_flags( &attributes, PSA_KEY_USAGE_SIGN );
psa_set_key_algorithm( &attributes, alg );
psa_set_key_type( &attributes, key_type );
Add calls for set policy in export/sign tests Add calls for set policy in export/sign tests
2018-03-28 12:46:26 +02:00
Split long lines after psa_import_key refactoring
2019-05-15 20:22:09 +02:00
PSA_ASSERT( psa_import_key( &attributes, key_data->x, key_data->len,
&handle ) );
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
PSA_ASSERT( psa_get_key_attributes( handle, &attributes ) );
key_bits = psa_get_key_bits( &attributes );
PSA crypto: PSA_ASYMMETRIC_SIGN_OUTPUT_SIZE macro Test it for RSA.
2018-02-03 23:57:22 +01:00
Document what the signature tests are doing a bit better Add a check that the purported output length is less than the buffer size in sign_fail.
2018-06-28 12:23:00 +02:00
/* Allocate a buffer which has the size advertized by the
* library. */
Formatting improvements Avoid lines longer than 80 columns. Remove some redundant parentheses, e.g. change if( ( a == b ) && ( c == d ) ) to if( a == b && c == d ) which makes lines less long and makes the remaining parentheses more relevant. Add missing parentheses around return statements. There should be no semantic change in this commit.
2018-06-18 16:04:39 +02:00
signature_size = PSA_ASYMMETRIC_SIGN_OUTPUT_SIZE( key_type,
key_bits, alg );
PSA crypto: PSA_ASYMMETRIC_SIGN_OUTPUT_SIZE macro Test it for RSA.
2018-02-03 23:57:22 +01:00
TEST_ASSERT( signature_size != 0 );
Use PSA_xxx_MAX_SIZE for hash/MAC/signature size in tests In tests that had a hard-coded buffer size, use PSA_MAC_MAX_SIZE or PSA_ASYMMETRIC_SIGNATURE_MAX_SIZE as appropriate. Test that PSA_xxx_MAX_SIZE is larger than the size used in tests that expect a specific output.
2018-06-28 00:07:19 +02:00
TEST_ASSERT( signature_size <= PSA_ASYMMETRIC_SIGNATURE_MAX_SIZE );
Use ASSERT_ALLOC instead of mbedtls_calloc in PSA tests This commit resolves a bug whereby some test cases failed on systems where mbedtls_calloc returns NULL when the size of 0, because the test case asserted `pointer != NULL` regardless of the size.
2018-09-27 13:54:18 +02:00
ASSERT_ALLOC( signature, signature_size );
PSA crypto: asymmetric signature (RSA PKCS#1v1.5 only) Define hash algorithms and RSA signature algorithms. New function psa_asymmetric_sign. Implement psa_asymmetric_sign for RSA PKCS#1 v1.5.
2018-02-03 22:44:14 +01:00
Document what the signature tests are doing a bit better Add a check that the purported output length is less than the buffer size in sign_fail.
2018-06-28 12:23:00 +02:00
/* Perform the signature. */
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_asymmetric_sign( handle, alg,
input_data->x, input_data->len,
signature, signature_size,
&signature_length ) );
Add sign_verify test and use it to smoke-test PSS
2018-06-29 17:30:48 +02:00
/* Verify that the signature is what is expected. */
Use ASSERT_COMPARE instead of memcmp in PSA tests This commit fixes some missing size comparison. In aead_encrypt_decrypt, aead_encrypt and aead_decrypt, the test code would not have noticed if the library function had reported an output length that was not the expected length.
2018-09-27 13:57:19 +02:00
ASSERT_COMPARE( output_data->x, output_data->len,
signature, signature_length );
PSA crypto: asymmetric signature (RSA PKCS#1v1.5 only) Define hash algorithms and RSA signature algorithms. New function psa_asymmetric_sign. Implement psa_asymmetric_sign for RSA PKCS#1 v1.5.
2018-02-03 22:44:14 +01:00
exit:
Call psa_reset_key_attributes after psa_get_key_attributes After calling psa_get_key_attributes(), call psa_reset_key_attributes() if the key may have domain parameters, because that's the way to free the domain parameter substructure in the attribute structure. Keep not calling reset() in some places where the key can only be a symmetric key which doesn't have domain parameters.
2019-04-26 16:03:33 +02:00
psa_reset_key_attributes( &attributes );
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_destroy_key( handle );
PSA crypto: PSA_ASYMMETRIC_SIGN_OUTPUT_SIZE macro Test it for RSA.
2018-02-03 23:57:22 +01:00
mbedtls_free( signature );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
PSA crypto: asymmetric signature (RSA PKCS#1v1.5 only) Define hash algorithms and RSA signature algorithms. New function psa_asymmetric_sign. Implement psa_asymmetric_sign for RSA PKCS#1 v1.5.
2018-02-03 22:44:14 +01:00
}
/* END_CASE */
/* BEGIN_CASE */
On target testing tests adaptation Updated all psa crypto tests to use the new test format
2018-06-12 16:06:52 +02:00
void sign_fail( int key_type_arg, data_t *key_data,
int alg_arg, data_t *input_data,
Document what the signature tests are doing a bit better Add a check that the purported output length is less than the buffer size in sign_fail.
2018-06-28 12:23:00 +02:00
int signature_size_arg, int expected_status_arg )
PSA crypto: asymmetric signature (RSA PKCS#1v1.5 only) Define hash algorithms and RSA signature algorithms. New function psa_asymmetric_sign. Implement psa_asymmetric_sign for RSA PKCS#1 v1.5.
2018-02-03 22:44:14 +01:00
{
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t handle = 0;
PSA crypto: asymmetric signature (RSA PKCS#1v1.5 only) Define hash algorithms and RSA signature algorithms. New function psa_asymmetric_sign. Implement psa_asymmetric_sign for RSA PKCS#1 v1.5.
2018-02-03 22:44:14 +01:00
psa_key_type_t key_type = key_type_arg;
psa_algorithm_t alg = alg_arg;
Document what the signature tests are doing a bit better Add a check that the purported output length is less than the buffer size in sign_fail.
2018-06-28 12:23:00 +02:00
size_t signature_size = signature_size_arg;
PSA crypto: asymmetric signature (RSA PKCS#1v1.5 only) Define hash algorithms and RSA signature algorithms. New function psa_asymmetric_sign. Implement psa_asymmetric_sign for RSA PKCS#1 v1.5.
2018-02-03 22:44:14 +01:00
psa_status_t actual_status;
psa_status_t expected_status = expected_status_arg;
Use unhexify_alloc where applicable
2018-03-07 16:43:36 +01:00
unsigned char *signature = NULL;
PSA asymmetric signature: set *signature_length = 0 on failure
2018-02-03 23:58:03 +01:00
size_t signature_length = 0xdeadbeef;
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
PSA crypto: asymmetric signature (RSA PKCS#1v1.5 only) Define hash algorithms and RSA signature algorithms. New function psa_asymmetric_sign. Implement psa_asymmetric_sign for RSA PKCS#1 v1.5.
2018-02-03 22:44:14 +01:00
Use ASSERT_ALLOC instead of mbedtls_calloc in PSA tests This commit resolves a bug whereby some test cases failed on systems where mbedtls_calloc returns NULL when the size of 0, because the test case asserted `pointer != NULL` regardless of the size.
2018-09-27 13:54:18 +02:00
ASSERT_ALLOC( signature, signature_size );
PSA crypto: asymmetric signature (RSA PKCS#1v1.5 only) Define hash algorithms and RSA signature algorithms. New function psa_asymmetric_sign. Implement psa_asymmetric_sign for RSA PKCS#1 v1.5.
2018-02-03 22:44:14 +01:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
PSA crypto: asymmetric signature (RSA PKCS#1v1.5 only) Define hash algorithms and RSA signature algorithms. New function psa_asymmetric_sign. Implement psa_asymmetric_sign for RSA PKCS#1 v1.5.
2018-02-03 22:44:14 +01:00
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_set_key_usage_flags( &attributes, PSA_KEY_USAGE_SIGN );
psa_set_key_algorithm( &attributes, alg );
psa_set_key_type( &attributes, key_type );
Add calls for set policy in export/sign tests Add calls for set policy in export/sign tests
2018-03-28 12:46:26 +02:00
Split long lines after psa_import_key refactoring
2019-05-15 20:22:09 +02:00
PSA_ASSERT( psa_import_key( &attributes, key_data->x, key_data->len,
&handle ) );
PSA crypto: asymmetric signature (RSA PKCS#1v1.5 only) Define hash algorithms and RSA signature algorithms. New function psa_asymmetric_sign. Implement psa_asymmetric_sign for RSA PKCS#1 v1.5.
2018-02-03 22:44:14 +01:00
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
actual_status = psa_asymmetric_sign( handle, alg,
Hygiene improvements in PSA crypto test code Get rid of many redundant casts. In particular, it is not useful to cast uint32_t values to size_t before performing arithmetic or comparisons on them. Rewrap a number of function calls, many of which now have narrower arguments thanks to the removed casts. When a function call doesn't fit on a single line, avoid grouping unrelated parameters together, but do try to group a buffer pointer and the associated size. Define more auxiliary variables xxx of a particular integer type (psa_algorithm_t, psa_key_usage_t, etc.) corresponding to a test function xxx_arg which has the type int. This avoids the need to cast xxx_arg to an unsigned type sometimes in the code.
2018-06-18 16:35:34 +02:00
input_data->x, input_data->len,
PSA crypto: asymmetric signature (RSA PKCS#1v1.5 only) Define hash algorithms and RSA signature algorithms. New function psa_asymmetric_sign. Implement psa_asymmetric_sign for RSA PKCS#1 v1.5.
2018-02-03 22:44:14 +01:00
signature, signature_size,
&signature_length );
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( actual_status, expected_status );
Document what the signature tests are doing a bit better Add a check that the purported output length is less than the buffer size in sign_fail.
2018-06-28 12:23:00 +02:00
/* The value of *signature_length is unspecified on error, but
* whatever it is, it should be less than signature_size, so that
* if the caller tries to read *signature_length bytes without
* checking the error code then they don't overflow a buffer. */
TEST_ASSERT( signature_length <= signature_size );
PSA crypto: asymmetric signature (RSA PKCS#1v1.5 only) Define hash algorithms and RSA signature algorithms. New function psa_asymmetric_sign. Implement psa_asymmetric_sign for RSA PKCS#1 v1.5.
2018-02-03 22:44:14 +01:00
exit:
Call psa_reset_key_attributes after psa_get_key_attributes After calling psa_get_key_attributes(), call psa_reset_key_attributes() if the key may have domain parameters, because that's the way to free the domain parameter substructure in the attribute structure. Keep not calling reset() in some places where the key can only be a symmetric key which doesn't have domain parameters.
2019-04-26 16:03:33 +02:00
psa_reset_key_attributes( &attributes );
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_destroy_key( handle );
PSA crypto: asymmetric signature (RSA PKCS#1v1.5 only) Define hash algorithms and RSA signature algorithms. New function psa_asymmetric_sign. Implement psa_asymmetric_sign for RSA PKCS#1 v1.5.
2018-02-03 22:44:14 +01:00
mbedtls_free( signature );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
Add sign_verify test and use it to smoke-test PSS
2018-06-29 17:30:48 +02:00
}
/* END_CASE */
/* BEGIN_CASE */
void sign_verify( int key_type_arg, data_t *key_data,
int alg_arg, data_t *input_data )
{
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t handle = 0;
Add sign_verify test and use it to smoke-test PSS
2018-06-29 17:30:48 +02:00
psa_key_type_t key_type = key_type_arg;
psa_algorithm_t alg = alg_arg;
size_t key_bits;
unsigned char *signature = NULL;
size_t signature_size;
size_t signature_length = 0xdeadbeef;
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
Add sign_verify test and use it to smoke-test PSS
2018-06-29 17:30:48 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
Add sign_verify test and use it to smoke-test PSS
2018-06-29 17:30:48 +02:00
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_set_key_usage_flags( &attributes, PSA_KEY_USAGE_SIGN | PSA_KEY_USAGE_VERIFY );
psa_set_key_algorithm( &attributes, alg );
psa_set_key_type( &attributes, key_type );
Add sign_verify test and use it to smoke-test PSS
2018-06-29 17:30:48 +02:00
Split long lines after psa_import_key refactoring
2019-05-15 20:22:09 +02:00
PSA_ASSERT( psa_import_key( &attributes, key_data->x, key_data->len,
&handle ) );
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
PSA_ASSERT( psa_get_key_attributes( handle, &attributes ) );
key_bits = psa_get_key_bits( &attributes );
Add sign_verify test and use it to smoke-test PSS
2018-06-29 17:30:48 +02:00
/* Allocate a buffer which has the size advertized by the
* library. */
signature_size = PSA_ASYMMETRIC_SIGN_OUTPUT_SIZE( key_type,
key_bits, alg );
TEST_ASSERT( signature_size != 0 );
TEST_ASSERT( signature_size <= PSA_ASYMMETRIC_SIGNATURE_MAX_SIZE );
Use ASSERT_ALLOC instead of mbedtls_calloc in PSA tests This commit resolves a bug whereby some test cases failed on systems where mbedtls_calloc returns NULL when the size of 0, because the test case asserted `pointer != NULL` regardless of the size.
2018-09-27 13:54:18 +02:00
ASSERT_ALLOC( signature, signature_size );
Add sign_verify test and use it to smoke-test PSS
2018-06-29 17:30:48 +02:00
/* Perform the signature. */
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_asymmetric_sign( handle, alg,
input_data->x, input_data->len,
signature, signature_size,
&signature_length ) );
Add sign_verify test and use it to smoke-test PSS
2018-06-29 17:30:48 +02:00
/* Check that the signature length looks sensible. */
TEST_ASSERT( signature_length <= signature_size );
TEST_ASSERT( signature_length > 0 );
/* Use the library to verify that the signature is correct. */
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_asymmetric_verify(
handle, alg,
input_data->x, input_data->len,
signature, signature_length ) );
Add sign_verify test and use it to smoke-test PSS
2018-06-29 17:30:48 +02:00
if( input_data->len != 0 )
{
/* Flip a bit in the input and verify that the signature is now
* detected as invalid. Flip a bit at the beginning, not at the end,
* because ECDSA may ignore the last few bits of the input. */
input_data->x[0] ^= 1;
Rewrap some lines after the macro changes Change the way some lines are wrapped to cut at a more logical place. This commit mainly rewrites multi-line calls to TEST_EQUAL, and also a few calls to PSA_ASSERT.
2018-12-18 00:33:25 +01:00
TEST_EQUAL( psa_asymmetric_verify( handle, alg,
input_data->x, input_data->len,
signature, signature_length ),
PSA_ERROR_INVALID_SIGNATURE );
Add sign_verify test and use it to smoke-test PSS
2018-06-29 17:30:48 +02:00
}
exit:
Call psa_reset_key_attributes after psa_get_key_attributes After calling psa_get_key_attributes(), call psa_reset_key_attributes() if the key may have domain parameters, because that's the way to free the domain parameter substructure in the attribute structure. Keep not calling reset() in some places where the key can only be a symmetric key which doesn't have domain parameters.
2019-04-26 16:03:33 +02:00
psa_reset_key_attributes( &attributes );
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_destroy_key( handle );
Add sign_verify test and use it to smoke-test PSS
2018-06-29 17:30:48 +02:00
mbedtls_free( signature );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
PSA crypto: asymmetric signature (RSA PKCS#1v1.5 only) Define hash algorithms and RSA signature algorithms. New function psa_asymmetric_sign. Implement psa_asymmetric_sign for RSA PKCS#1 v1.5.
2018-02-03 22:44:14 +01:00
}
/* END_CASE */
Key Policy APIs implementation
2018-03-28 00:21:33 +02:00
ECDSA sign and verify implementation and tests ECDSA sign and verify implementation and tests
2018-05-08 10:18:38 +02:00
/* BEGIN_CASE */
On target testing tests adaptation Updated all psa crypto tests to use the new test format
2018-06-12 16:06:52 +02:00
void asymmetric_verify( int key_type_arg, data_t *key_data,
int alg_arg, data_t *hash_data,
data_t *signature_data )
ECDSA sign and verify implementation and tests ECDSA sign and verify implementation and tests
2018-05-08 10:18:38 +02:00
{
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t handle = 0;
ECDSA sign and verify implementation and tests ECDSA sign and verify implementation and tests
2018-05-08 10:18:38 +02:00
psa_key_type_t key_type = key_type_arg;
psa_algorithm_t alg = alg_arg;
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
ECDSA sign and verify implementation and tests ECDSA sign and verify implementation and tests
2018-05-08 10:18:38 +02:00
Use PSA_xxx_MAX_SIZE for hash/MAC/signature size in tests In tests that had a hard-coded buffer size, use PSA_MAC_MAX_SIZE or PSA_ASYMMETRIC_SIGNATURE_MAX_SIZE as appropriate. Test that PSA_xxx_MAX_SIZE is larger than the size used in tests that expect a specific output.
2018-06-28 00:07:19 +02:00
TEST_ASSERT( signature_data->len <= PSA_ASYMMETRIC_SIGNATURE_MAX_SIZE );
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
ECDSA sign and verify implementation and tests ECDSA sign and verify implementation and tests
2018-05-08 10:18:38 +02:00
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_set_key_usage_flags( &attributes, PSA_KEY_USAGE_VERIFY );
psa_set_key_algorithm( &attributes, alg );
psa_set_key_type( &attributes, key_type );
ECDSA sign and verify implementation and tests ECDSA sign and verify implementation and tests
2018-05-08 10:18:38 +02:00
Split long lines after psa_import_key refactoring
2019-05-15 20:22:09 +02:00
PSA_ASSERT( psa_import_key( &attributes, key_data->x, key_data->len,
&handle ) );
ECDSA sign and verify implementation and tests ECDSA sign and verify implementation and tests
2018-05-08 10:18:38 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_asymmetric_verify( handle, alg,
hash_data->x, hash_data->len,
signature_data->x,
signature_data->len ) );
ECDSA sign and verify implementation and tests ECDSA sign and verify implementation and tests
2018-05-08 10:18:38 +02:00
exit:
Call psa_reset_key_attributes after psa_get_key_attributes After calling psa_get_key_attributes(), call psa_reset_key_attributes() if the key may have domain parameters, because that's the way to free the domain parameter substructure in the attribute structure. Keep not calling reset() in some places where the key can only be a symmetric key which doesn't have domain parameters.
2019-04-26 16:03:33 +02:00
psa_reset_key_attributes( &attributes );
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_destroy_key( handle );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
ECDSA sign and verify implementation and tests ECDSA sign and verify implementation and tests
2018-05-08 10:18:38 +02:00
}
/* END_CASE */
added support for PKCSv1.5 signature verification and encryption/decryption and very basic tests.
2018-05-02 22:16:26 +02:00
/* BEGIN_CASE */
On target testing tests adaptation Updated all psa crypto tests to use the new test format
2018-06-12 16:06:52 +02:00
void asymmetric_verify_fail( int key_type_arg, data_t *key_data,
int alg_arg, data_t *hash_data,
data_t *signature_data,
int expected_status_arg )
added support for PKCSv1.5 signature verification and encryption/decryption and very basic tests.
2018-05-02 22:16:26 +02:00
{
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t handle = 0;
added support for PKCSv1.5 signature verification and encryption/decryption and very basic tests.
2018-05-02 22:16:26 +02:00
psa_key_type_t key_type = key_type_arg;
psa_algorithm_t alg = alg_arg;
psa_status_t actual_status;
psa_status_t expected_status = expected_status_arg;
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
added support for PKCSv1.5 signature verification and encryption/decryption and very basic tests.
2018-05-02 22:16:26 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
added support for PKCSv1.5 signature verification and encryption/decryption and very basic tests.
2018-05-02 22:16:26 +02:00
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_set_key_usage_flags( &attributes, PSA_KEY_USAGE_VERIFY );
psa_set_key_algorithm( &attributes, alg );
psa_set_key_type( &attributes, key_type );
add key policy enforcement implementation add checks that keys have been set for the correct usage for asymmetric functions.
2018-06-04 15:45:27 +02:00
Split long lines after psa_import_key refactoring
2019-05-15 20:22:09 +02:00
PSA_ASSERT( psa_import_key( &attributes, key_data->x, key_data->len,
&handle ) );
added support for PKCSv1.5 signature verification and encryption/decryption and very basic tests.
2018-05-02 22:16:26 +02:00
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
actual_status = psa_asymmetric_verify( handle, alg,
Hygiene improvements in PSA crypto test code Get rid of many redundant casts. In particular, it is not useful to cast uint32_t values to size_t before performing arithmetic or comparisons on them. Rewrap a number of function calls, many of which now have narrower arguments thanks to the removed casts. When a function call doesn't fit on a single line, avoid grouping unrelated parameters together, but do try to group a buffer pointer and the associated size. Define more auxiliary variables xxx of a particular integer type (psa_algorithm_t, psa_key_usage_t, etc.) corresponding to a test function xxx_arg which has the type int. This avoids the need to cast xxx_arg to an unsigned type sometimes in the code.
2018-06-18 16:35:34 +02:00
hash_data->x, hash_data->len,
Normalize whitespace Normalize whitespace to Mbed TLS standards. There are only whitespace changes in this commit.
2018-06-18 15:41:12 +02:00
signature_data->x,
Hygiene improvements in PSA crypto test code Get rid of many redundant casts. In particular, it is not useful to cast uint32_t values to size_t before performing arithmetic or comparisons on them. Rewrap a number of function calls, many of which now have narrower arguments thanks to the removed casts. When a function call doesn't fit on a single line, avoid grouping unrelated parameters together, but do try to group a buffer pointer and the associated size. Define more auxiliary variables xxx of a particular integer type (psa_algorithm_t, psa_key_usage_t, etc.) corresponding to a test function xxx_arg which has the type int. This avoids the need to cast xxx_arg to an unsigned type sometimes in the code.
2018-06-18 16:35:34 +02:00
signature_data->len );
added support for PKCSv1.5 signature verification and encryption/decryption and very basic tests.
2018-05-02 22:16:26 +02:00
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( actual_status, expected_status );
added support for PKCSv1.5 signature verification and encryption/decryption and very basic tests.
2018-05-02 22:16:26 +02:00
exit:
Call psa_reset_key_attributes after psa_get_key_attributes After calling psa_get_key_attributes(), call psa_reset_key_attributes() if the key may have domain parameters, because that's the way to free the domain parameter substructure in the attribute structure. Keep not calling reset() in some places where the key can only be a symmetric key which doesn't have domain parameters.
2019-04-26 16:03:33 +02:00
psa_reset_key_attributes( &attributes );
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_destroy_key( handle );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
added support for PKCSv1.5 signature verification and encryption/decryption and very basic tests.
2018-05-02 22:16:26 +02:00
}
/* END_CASE */
Add positive asymmetric encryption tests Revise the test function asymmetric_encrypt_fail into asymmetric_encrypt and use it for positive tests as well. Get the expected output length from PSA_ASYMMETRIC_ENCRYPT_OUTPUT_SIZE. Check the actual output length against test data. Add positive test cases for encryption: one with an RSA public key (this is the only test for encryption with a public key rather than a key pair) and one with a key pair.
2018-06-29 19:12:28 +02:00
/* BEGIN_CASE */
void asymmetric_encrypt( int key_type_arg,
data_t *key_data,
int alg_arg,
data_t *input_data,
Asymmetric encryption tests: allow label argument Add a label argument to all asymmetric encryption test functions (currently empty in all tests, but that will change soon). In asymmetric_encrypt and asymmetric_decrypt, with an empty label, test with both a null pointer and a non-null pointer.
2018-06-30 18:42:41 +02:00
data_t *label,
Add positive asymmetric encryption tests Revise the test function asymmetric_encrypt_fail into asymmetric_encrypt and use it for positive tests as well. Get the expected output length from PSA_ASYMMETRIC_ENCRYPT_OUTPUT_SIZE. Check the actual output length against test data. Add positive test cases for encryption: one with an RSA public key (this is the only test for encryption with a public key rather than a key pair) and one with a key pair.
2018-06-29 19:12:28 +02:00
int expected_output_length_arg,
int expected_status_arg )
{
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t handle = 0;
Add positive asymmetric encryption tests Revise the test function asymmetric_encrypt_fail into asymmetric_encrypt and use it for positive tests as well. Get the expected output length from PSA_ASYMMETRIC_ENCRYPT_OUTPUT_SIZE. Check the actual output length against test data. Add positive test cases for encryption: one with an RSA public key (this is the only test for encryption with a public key rather than a key pair) and one with a key pair.
2018-06-29 19:12:28 +02:00
psa_key_type_t key_type = key_type_arg;
psa_algorithm_t alg = alg_arg;
size_t expected_output_length = expected_output_length_arg;
size_t key_bits;
unsigned char *output = NULL;
size_t output_size;
size_t output_length = ~0;
psa_status_t actual_status;
psa_status_t expected_status = expected_status_arg;
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
Add positive asymmetric encryption tests Revise the test function asymmetric_encrypt_fail into asymmetric_encrypt and use it for positive tests as well. Get the expected output length from PSA_ASYMMETRIC_ENCRYPT_OUTPUT_SIZE. Check the actual output length against test data. Add positive test cases for encryption: one with an RSA public key (this is the only test for encryption with a public key rather than a key pair) and one with a key pair.
2018-06-29 19:12:28 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
Add positive asymmetric encryption tests Revise the test function asymmetric_encrypt_fail into asymmetric_encrypt and use it for positive tests as well. Get the expected output length from PSA_ASYMMETRIC_ENCRYPT_OUTPUT_SIZE. Check the actual output length against test data. Add positive test cases for encryption: one with an RSA public key (this is the only test for encryption with a public key rather than a key pair) and one with a key pair.
2018-06-29 19:12:28 +02:00
/* Import the key */
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_set_key_usage_flags( &attributes, PSA_KEY_USAGE_ENCRYPT );
psa_set_key_algorithm( &attributes, alg );
psa_set_key_type( &attributes, key_type );
Split long lines after psa_import_key refactoring
2019-05-15 20:22:09 +02:00
PSA_ASSERT( psa_import_key( &attributes, key_data->x, key_data->len,
&handle ) );
Add positive asymmetric encryption tests Revise the test function asymmetric_encrypt_fail into asymmetric_encrypt and use it for positive tests as well. Get the expected output length from PSA_ASYMMETRIC_ENCRYPT_OUTPUT_SIZE. Check the actual output length against test data. Add positive test cases for encryption: one with an RSA public key (this is the only test for encryption with a public key rather than a key pair) and one with a key pair.
2018-06-29 19:12:28 +02:00
/* Determine the maximum output length */
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
PSA_ASSERT( psa_get_key_attributes( handle, &attributes ) );
key_bits = psa_get_key_bits( &attributes );
Add positive asymmetric encryption tests Revise the test function asymmetric_encrypt_fail into asymmetric_encrypt and use it for positive tests as well. Get the expected output length from PSA_ASYMMETRIC_ENCRYPT_OUTPUT_SIZE. Check the actual output length against test data. Add positive test cases for encryption: one with an RSA public key (this is the only test for encryption with a public key rather than a key pair) and one with a key pair.
2018-06-29 19:12:28 +02:00
output_size = PSA_ASYMMETRIC_ENCRYPT_OUTPUT_SIZE( key_type, key_bits, alg );
Use ASSERT_ALLOC instead of mbedtls_calloc in PSA tests This commit resolves a bug whereby some test cases failed on systems where mbedtls_calloc returns NULL when the size of 0, because the test case asserted `pointer != NULL` regardless of the size.
2018-09-27 13:54:18 +02:00
ASSERT_ALLOC( output, output_size );
Add positive asymmetric encryption tests Revise the test function asymmetric_encrypt_fail into asymmetric_encrypt and use it for positive tests as well. Get the expected output length from PSA_ASYMMETRIC_ENCRYPT_OUTPUT_SIZE. Check the actual output length against test data. Add positive test cases for encryption: one with an RSA public key (this is the only test for encryption with a public key rather than a key pair) and one with a key pair.
2018-06-29 19:12:28 +02:00
/* Encrypt the input */
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
actual_status = psa_asymmetric_encrypt( handle, alg,
Add positive asymmetric encryption tests Revise the test function asymmetric_encrypt_fail into asymmetric_encrypt and use it for positive tests as well. Get the expected output length from PSA_ASYMMETRIC_ENCRYPT_OUTPUT_SIZE. Check the actual output length against test data. Add positive test cases for encryption: one with an RSA public key (this is the only test for encryption with a public key rather than a key pair) and one with a key pair.
2018-06-29 19:12:28 +02:00
input_data->x, input_data->len,
Asymmetric encryption tests: allow label argument Add a label argument to all asymmetric encryption test functions (currently empty in all tests, but that will change soon). In asymmetric_encrypt and asymmetric_decrypt, with an empty label, test with both a null pointer and a non-null pointer.
2018-06-30 18:42:41 +02:00
label->x, label->len,
Add positive asymmetric encryption tests Revise the test function asymmetric_encrypt_fail into asymmetric_encrypt and use it for positive tests as well. Get the expected output length from PSA_ASYMMETRIC_ENCRYPT_OUTPUT_SIZE. Check the actual output length against test data. Add positive test cases for encryption: one with an RSA public key (this is the only test for encryption with a public key rather than a key pair) and one with a key pair.
2018-06-29 19:12:28 +02:00
output, output_size,
&output_length );
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( actual_status, expected_status );
TEST_EQUAL( output_length, expected_output_length );
Add positive asymmetric encryption tests Revise the test function asymmetric_encrypt_fail into asymmetric_encrypt and use it for positive tests as well. Get the expected output length from PSA_ASYMMETRIC_ENCRYPT_OUTPUT_SIZE. Check the actual output length against test data. Add positive test cases for encryption: one with an RSA public key (this is the only test for encryption with a public key rather than a key pair) and one with a key pair.
2018-06-29 19:12:28 +02:00
Asymmetric encryption tests: allow label argument Add a label argument to all asymmetric encryption test functions (currently empty in all tests, but that will change soon). In asymmetric_encrypt and asymmetric_decrypt, with an empty label, test with both a null pointer and a non-null pointer.
2018-06-30 18:42:41 +02:00
/* If the label is empty, the test framework puts a non-null pointer
* in label->x. Test that a null pointer works as well. */
if( label->len == 0 )
{
output_length = ~0;
Skip calling memset when the size is 0 memset(NULL, c, 0) has undefined behavior, so don't do it. clang-asan complains.
2018-09-26 18:19:24 +02:00
if( output_size != 0 )
memset( output, 0, output_size );
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
actual_status = psa_asymmetric_encrypt( handle, alg,
Asymmetric encryption tests: allow label argument Add a label argument to all asymmetric encryption test functions (currently empty in all tests, but that will change soon). In asymmetric_encrypt and asymmetric_decrypt, with an empty label, test with both a null pointer and a non-null pointer.
2018-06-30 18:42:41 +02:00
input_data->x, input_data->len,
NULL, label->len,
output, output_size,
&output_length );
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( actual_status, expected_status );
TEST_EQUAL( output_length, expected_output_length );
Asymmetric encryption tests: allow label argument Add a label argument to all asymmetric encryption test functions (currently empty in all tests, but that will change soon). In asymmetric_encrypt and asymmetric_decrypt, with an empty label, test with both a null pointer and a non-null pointer.
2018-06-30 18:42:41 +02:00
}
Add positive asymmetric encryption tests Revise the test function asymmetric_encrypt_fail into asymmetric_encrypt and use it for positive tests as well. Get the expected output length from PSA_ASYMMETRIC_ENCRYPT_OUTPUT_SIZE. Check the actual output length against test data. Add positive test cases for encryption: one with an RSA public key (this is the only test for encryption with a public key rather than a key pair) and one with a key pair.
2018-06-29 19:12:28 +02:00
exit:
Call psa_reset_key_attributes after psa_get_key_attributes After calling psa_get_key_attributes(), call psa_reset_key_attributes() if the key may have domain parameters, because that's the way to free the domain parameter substructure in the attribute structure. Keep not calling reset() in some places where the key can only be a symmetric key which doesn't have domain parameters.
2019-04-26 16:03:33 +02:00
psa_reset_key_attributes( &attributes );
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_destroy_key( handle );
Add positive asymmetric encryption tests Revise the test function asymmetric_encrypt_fail into asymmetric_encrypt and use it for positive tests as well. Get the expected output length from PSA_ASYMMETRIC_ENCRYPT_OUTPUT_SIZE. Check the actual output length against test data. Add positive test cases for encryption: one with an RSA public key (this is the only test for encryption with a public key rather than a key pair) and one with a key pair.
2018-06-29 19:12:28 +02:00
mbedtls_free( output );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
Add positive asymmetric encryption tests Revise the test function asymmetric_encrypt_fail into asymmetric_encrypt and use it for positive tests as well. Get the expected output length from PSA_ASYMMETRIC_ENCRYPT_OUTPUT_SIZE. Check the actual output length against test data. Add positive test cases for encryption: one with an RSA public key (this is the only test for encryption with a public key rather than a key pair) and one with a key pair.
2018-06-29 19:12:28 +02:00
}
/* END_CASE */
added support for PKCSv1.5 signature verification and encryption/decryption and very basic tests.
2018-05-02 22:16:26 +02:00
/* BEGIN_CASE */
Asymmetric encryption tests: allow label argument Add a label argument to all asymmetric encryption test functions (currently empty in all tests, but that will change soon). In asymmetric_encrypt and asymmetric_decrypt, with an empty label, test with both a null pointer and a non-null pointer.
2018-06-30 18:42:41 +02:00
void asymmetric_encrypt_decrypt( int key_type_arg,
data_t *key_data,
int alg_arg,
data_t *input_data,
data_t *label )
added support for PKCSv1.5 signature verification and encryption/decryption and very basic tests.
2018-05-02 22:16:26 +02:00
{
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t handle = 0;
added support for PKCSv1.5 signature verification and encryption/decryption and very basic tests.
2018-05-02 22:16:26 +02:00
psa_key_type_t key_type = key_type_arg;
psa_algorithm_t alg = alg_arg;
Asymmetric encrypt/decrypt tests: check output length In asymmetric_encrypt_decrypt, use the buffer size advertized by the library for the ciphertext, and the length of the plaintext for the re-decrypted output. Test the output length if known. Require it to be 0 on error for encrypt/decrypt functions. If the output length is unknown, test at least that it's within the buffer limits.
2018-06-30 18:54:48 +02:00
size_t key_bits;
added support for PKCSv1.5 signature verification and encryption/decryption and very basic tests.
2018-05-02 22:16:26 +02:00
unsigned char *output = NULL;
Asymmetric encrypt/decrypt tests: check output length In asymmetric_encrypt_decrypt, use the buffer size advertized by the library for the ciphertext, and the length of the plaintext for the re-decrypted output. Test the output length if known. Require it to be 0 on error for encrypt/decrypt functions. If the output length is unknown, test at least that it's within the buffer limits.
2018-06-30 18:54:48 +02:00
size_t output_size;
size_t output_length = ~0;
change RSA encryption tests compensate for random component in encryption.
2018-05-02 22:56:12 +02:00
unsigned char *output2 = NULL;
Asymmetric encrypt/decrypt tests: check output length In asymmetric_encrypt_decrypt, use the buffer size advertized by the library for the ciphertext, and the length of the plaintext for the re-decrypted output. Test the output length if known. Require it to be 0 on error for encrypt/decrypt functions. If the output length is unknown, test at least that it's within the buffer limits.
2018-06-30 18:54:48 +02:00
size_t output2_size;
size_t output2_length = ~0;
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
added support for PKCSv1.5 signature verification and encryption/decryption and very basic tests.
2018-05-02 22:16:26 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
added support for PKCSv1.5 signature verification and encryption/decryption and very basic tests.
2018-05-02 22:16:26 +02:00
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_set_key_usage_flags( &attributes, PSA_KEY_USAGE_ENCRYPT | PSA_KEY_USAGE_DECRYPT );
psa_set_key_algorithm( &attributes, alg );
psa_set_key_type( &attributes, key_type );
added support for PKCSv1.5 signature verification and encryption/decryption and very basic tests.
2018-05-02 22:16:26 +02:00
Split long lines after psa_import_key refactoring
2019-05-15 20:22:09 +02:00
PSA_ASSERT( psa_import_key( &attributes, key_data->x, key_data->len,
&handle ) );
Asymmetric encrypt/decrypt tests: check output length In asymmetric_encrypt_decrypt, use the buffer size advertized by the library for the ciphertext, and the length of the plaintext for the re-decrypted output. Test the output length if known. Require it to be 0 on error for encrypt/decrypt functions. If the output length is unknown, test at least that it's within the buffer limits.
2018-06-30 18:54:48 +02:00
/* Determine the maximum ciphertext length */
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
PSA_ASSERT( psa_get_key_attributes( handle, &attributes ) );
key_bits = psa_get_key_bits( &attributes );
Asymmetric encrypt/decrypt tests: check output length In asymmetric_encrypt_decrypt, use the buffer size advertized by the library for the ciphertext, and the length of the plaintext for the re-decrypted output. Test the output length if known. Require it to be 0 on error for encrypt/decrypt functions. If the output length is unknown, test at least that it's within the buffer limits.
2018-06-30 18:54:48 +02:00
output_size = PSA_ASYMMETRIC_ENCRYPT_OUTPUT_SIZE( key_type, key_bits, alg );
Use ASSERT_ALLOC instead of mbedtls_calloc in PSA tests This commit resolves a bug whereby some test cases failed on systems where mbedtls_calloc returns NULL when the size of 0, because the test case asserted `pointer != NULL` regardless of the size.
2018-09-27 13:54:18 +02:00
ASSERT_ALLOC( output, output_size );
Asymmetric encrypt/decrypt tests: check output length In asymmetric_encrypt_decrypt, use the buffer size advertized by the library for the ciphertext, and the length of the plaintext for the re-decrypted output. Test the output length if known. Require it to be 0 on error for encrypt/decrypt functions. If the output length is unknown, test at least that it's within the buffer limits.
2018-06-30 18:54:48 +02:00
output2_size = input_data->len;
Use ASSERT_ALLOC instead of mbedtls_calloc in PSA tests This commit resolves a bug whereby some test cases failed on systems where mbedtls_calloc returns NULL when the size of 0, because the test case asserted `pointer != NULL` regardless of the size.
2018-09-27 13:54:18 +02:00
ASSERT_ALLOC( output2, output2_size );
Asymmetric encrypt/decrypt tests: check output length In asymmetric_encrypt_decrypt, use the buffer size advertized by the library for the ciphertext, and the length of the plaintext for the re-decrypted output. Test the output length if known. Require it to be 0 on error for encrypt/decrypt functions. If the output length is unknown, test at least that it's within the buffer limits.
2018-06-30 18:54:48 +02:00
Rename asymmetric_encrypt to clarify what it does Renamed to asymmetric_encrypt_decrypt
2018-06-08 18:11:54 +02:00
/* We test encryption by checking that encrypt-then-decrypt gives back
* the original plaintext because of the non-optional random
* part of encryption process which prevents using fixed vectors. */
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_asymmetric_encrypt( handle, alg,
input_data->x, input_data->len,
label->x, label->len,
output, output_size,
&output_length ) );
Asymmetric encrypt/decrypt tests: check output length In asymmetric_encrypt_decrypt, use the buffer size advertized by the library for the ciphertext, and the length of the plaintext for the re-decrypted output. Test the output length if known. Require it to be 0 on error for encrypt/decrypt functions. If the output length is unknown, test at least that it's within the buffer limits.
2018-06-30 18:54:48 +02:00
/* We don't know what ciphertext length to expect, but check that
* it looks sensible. */
TEST_ASSERT( output_length <= output_size );
Normalize whitespace Normalize whitespace to Mbed TLS standards. There are only whitespace changes in this commit.
2018-06-18 15:41:12 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_asymmetric_decrypt( handle, alg,
output, output_length,
label->x, label->len,
output2, output2_size,
&output2_length ) );
Use ASSERT_COMPARE instead of memcmp in PSA tests This commit fixes some missing size comparison. In aead_encrypt_decrypt, aead_encrypt and aead_decrypt, the test code would not have noticed if the library function had reported an output length that was not the expected length.
2018-09-27 13:57:19 +02:00
ASSERT_COMPARE( input_data->x, input_data->len,
output2, output2_length );
added support for PKCSv1.5 signature verification and encryption/decryption and very basic tests.
2018-05-02 22:16:26 +02:00
exit:
Call psa_reset_key_attributes after psa_get_key_attributes After calling psa_get_key_attributes(), call psa_reset_key_attributes() if the key may have domain parameters, because that's the way to free the domain parameter substructure in the attribute structure. Keep not calling reset() in some places where the key can only be a symmetric key which doesn't have domain parameters.
2019-04-26 16:03:33 +02:00
psa_reset_key_attributes( &attributes );
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_destroy_key( handle );
On target testing tests adaptation Updated all psa crypto tests to use the new test format
2018-06-12 16:06:52 +02:00
mbedtls_free( output );
mbedtls_free( output2 );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
added support for PKCSv1.5 signature verification and encryption/decryption and very basic tests.
2018-05-02 22:16:26 +02:00
}
/* END_CASE */
/* BEGIN_CASE */
Asymmetric encryption tests: allow label argument Add a label argument to all asymmetric encryption test functions (currently empty in all tests, but that will change soon). In asymmetric_encrypt and asymmetric_decrypt, with an empty label, test with both a null pointer and a non-null pointer.
2018-06-30 18:42:41 +02:00
void asymmetric_decrypt( int key_type_arg,
data_t *key_data,
int alg_arg,
data_t *input_data,
data_t *label,
asymmetric_decrypt test: remove redundant argument The expected output size is the size of the expected output, it doesn't need to be passed separately.
2018-06-29 21:54:10 +02:00
data_t *expected_data )
added support for PKCSv1.5 signature verification and encryption/decryption and very basic tests.
2018-05-02 22:16:26 +02:00
{
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t handle = 0;
added support for PKCSv1.5 signature verification and encryption/decryption and very basic tests.
2018-05-02 22:16:26 +02:00
psa_key_type_t key_type = key_type_arg;
psa_algorithm_t alg = alg_arg;
unsigned char *output = NULL;
Fix test output size 1. set output size to safe value 2. set output size correctly 3. check correct length of actual output
2018-06-04 15:31:13 +02:00
size_t output_size = 0;
Asymmetric encrypt/decrypt tests: check output length In asymmetric_encrypt_decrypt, use the buffer size advertized by the library for the ciphertext, and the length of the plaintext for the re-decrypted output. Test the output length if known. Require it to be 0 on error for encrypt/decrypt functions. If the output length is unknown, test at least that it's within the buffer limits.
2018-06-30 18:54:48 +02:00
size_t output_length = ~0;
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
added support for PKCSv1.5 signature verification and encryption/decryption and very basic tests.
2018-05-02 22:16:26 +02:00
psa: Expect output-buffer-sized RSA-decryption When RSA decrypting, unlike with RSA encrypting, we sometimes expect the output length will be less than the key size. For instance, in the case where the plaintext is zero-length we expect the output length of the decryption to be zero-length as well, not key size in length. For must-fail tests, we don't expect output-buffer-sized RSA-decryption, only that the output length is less than or equal to the output size, so these tests remain unchanged. Change the must-pass tests to expect that the actual output size is equal to the expected length of the output buffer instead of always being the key size.
2019-02-06 13:57:46 +01:00
output_size = expected_data->len;
Use ASSERT_ALLOC instead of mbedtls_calloc in PSA tests This commit resolves a bug whereby some test cases failed on systems where mbedtls_calloc returns NULL when the size of 0, because the test case asserted `pointer != NULL` regardless of the size.
2018-09-27 13:54:18 +02:00
ASSERT_ALLOC( output, output_size );
Remove trailing whitespace Only horizontal whitespace changes in this commit.
2018-05-31 13:25:48 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
added support for PKCSv1.5 signature verification and encryption/decryption and very basic tests.
2018-05-02 22:16:26 +02:00
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_set_key_usage_flags( &attributes, PSA_KEY_USAGE_DECRYPT );
psa_set_key_algorithm( &attributes, alg );
psa_set_key_type( &attributes, key_type );
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
Split long lines after psa_import_key refactoring
2019-05-15 20:22:09 +02:00
PSA_ASSERT( psa_import_key( &attributes, key_data->x, key_data->len,
&handle ) );
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_asymmetric_decrypt( handle, alg,
input_data->x, input_data->len,
label->x, label->len,
output,
output_size,
&output_length ) );
Use ASSERT_COMPARE instead of memcmp in PSA tests This commit fixes some missing size comparison. In aead_encrypt_decrypt, aead_encrypt and aead_decrypt, the test code would not have noticed if the library function had reported an output length that was not the expected length.
2018-09-27 13:57:19 +02:00
ASSERT_COMPARE( expected_data->x, expected_data->len,
output, output_length );
added support for PKCSv1.5 signature verification and encryption/decryption and very basic tests.
2018-05-02 22:16:26 +02:00
Asymmetric encryption tests: allow label argument Add a label argument to all asymmetric encryption test functions (currently empty in all tests, but that will change soon). In asymmetric_encrypt and asymmetric_decrypt, with an empty label, test with both a null pointer and a non-null pointer.
2018-06-30 18:42:41 +02:00
/* If the label is empty, the test framework puts a non-null pointer
* in label->x. Test that a null pointer works as well. */
if( label->len == 0 )
{
output_length = ~0;
Skip calling memset when the size is 0 memset(NULL, c, 0) has undefined behavior, so don't do it. clang-asan complains.
2018-09-26 18:19:24 +02:00
if( output_size != 0 )
memset( output, 0, output_size );
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_asymmetric_decrypt( handle, alg,
input_data->x, input_data->len,
NULL, label->len,
output,
output_size,
&output_length ) );
Use ASSERT_COMPARE instead of memcmp in PSA tests This commit fixes some missing size comparison. In aead_encrypt_decrypt, aead_encrypt and aead_decrypt, the test code would not have noticed if the library function had reported an output length that was not the expected length.
2018-09-27 13:57:19 +02:00
ASSERT_COMPARE( expected_data->x, expected_data->len,
output, output_length );
Asymmetric encryption tests: allow label argument Add a label argument to all asymmetric encryption test functions (currently empty in all tests, but that will change soon). In asymmetric_encrypt and asymmetric_decrypt, with an empty label, test with both a null pointer and a non-null pointer.
2018-06-30 18:42:41 +02:00
}
added support for PKCSv1.5 signature verification and encryption/decryption and very basic tests.
2018-05-02 22:16:26 +02:00
exit:
Call psa_reset_key_attributes after psa_get_key_attributes After calling psa_get_key_attributes(), call psa_reset_key_attributes() if the key may have domain parameters, because that's the way to free the domain parameter substructure in the attribute structure. Keep not calling reset() in some places where the key can only be a symmetric key which doesn't have domain parameters.
2019-04-26 16:03:33 +02:00
psa_reset_key_attributes( &attributes );
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_destroy_key( handle );
On target testing tests adaptation Updated all psa crypto tests to use the new test format
2018-06-12 16:06:52 +02:00
mbedtls_free( output );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
added support for PKCSv1.5 signature verification and encryption/decryption and very basic tests.
2018-05-02 22:16:26 +02:00
}
/* END_CASE */
/* BEGIN_CASE */
Asymmetric encryption tests: allow label argument Add a label argument to all asymmetric encryption test functions (currently empty in all tests, but that will change soon). In asymmetric_encrypt and asymmetric_decrypt, with an empty label, test with both a null pointer and a non-null pointer.
2018-06-30 18:42:41 +02:00
void asymmetric_decrypt_fail( int key_type_arg,
data_t *key_data,
int alg_arg,
data_t *input_data,
data_t *label,
psa: Enable testing RSA decryption with output sizes For must-fail asymmetric decryption tests, add an output size parameter so that tests can directly control what output buffer size they allocate and use independently from the key size used. This enables better testing of behavior with various output buffer sizes.
2019-02-06 13:57:46 +01:00
int output_size_arg,
Normalize whitespace Normalize whitespace to Mbed TLS standards. There are only whitespace changes in this commit.
2018-06-18 15:41:12 +02:00
int expected_status_arg )
added support for PKCSv1.5 signature verification and encryption/decryption and very basic tests.
2018-05-02 22:16:26 +02:00
{
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t handle = 0;
added support for PKCSv1.5 signature verification and encryption/decryption and very basic tests.
2018-05-02 22:16:26 +02:00
psa_key_type_t key_type = key_type_arg;
psa_algorithm_t alg = alg_arg;
unsigned char *output = NULL;
psa: Enable testing RSA decryption with output sizes For must-fail asymmetric decryption tests, add an output size parameter so that tests can directly control what output buffer size they allocate and use independently from the key size used. This enables better testing of behavior with various output buffer sizes.
2019-02-06 13:57:46 +01:00
size_t output_size = output_size_arg;
Asymmetric encrypt/decrypt tests: check output length In asymmetric_encrypt_decrypt, use the buffer size advertized by the library for the ciphertext, and the length of the plaintext for the re-decrypted output. Test the output length if known. Require it to be 0 on error for encrypt/decrypt functions. If the output length is unknown, test at least that it's within the buffer limits.
2018-06-30 18:54:48 +02:00
size_t output_length = ~0;
added support for PKCSv1.5 signature verification and encryption/decryption and very basic tests.
2018-05-02 22:16:26 +02:00
psa_status_t actual_status;
psa_status_t expected_status = expected_status_arg;
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
added support for PKCSv1.5 signature verification and encryption/decryption and very basic tests.
2018-05-02 22:16:26 +02:00
Use ASSERT_ALLOC instead of mbedtls_calloc in PSA tests This commit resolves a bug whereby some test cases failed on systems where mbedtls_calloc returns NULL when the size of 0, because the test case asserted `pointer != NULL` regardless of the size.
2018-09-27 13:54:18 +02:00
ASSERT_ALLOC( output, output_size );
Remove trailing whitespace Only horizontal whitespace changes in this commit.
2018-05-31 13:25:48 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
added support for PKCSv1.5 signature verification and encryption/decryption and very basic tests.
2018-05-02 22:16:26 +02:00
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_set_key_usage_flags( &attributes, PSA_KEY_USAGE_DECRYPT );
psa_set_key_algorithm( &attributes, alg );
psa_set_key_type( &attributes, key_type );
add key policy enforcement implementation add checks that keys have been set for the correct usage for asymmetric functions.
2018-06-04 15:45:27 +02:00
Split long lines after psa_import_key refactoring
2019-05-15 20:22:09 +02:00
PSA_ASSERT( psa_import_key( &attributes, key_data->x, key_data->len,
&handle ) );
Normalize whitespace Normalize whitespace to Mbed TLS standards. There are only whitespace changes in this commit.
2018-06-18 15:41:12 +02:00
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
actual_status = psa_asymmetric_decrypt( handle, alg,
Hygiene improvements in PSA crypto test code Get rid of many redundant casts. In particular, it is not useful to cast uint32_t values to size_t before performing arithmetic or comparisons on them. Rewrap a number of function calls, many of which now have narrower arguments thanks to the removed casts. When a function call doesn't fit on a single line, avoid grouping unrelated parameters together, but do try to group a buffer pointer and the associated size. Define more auxiliary variables xxx of a particular integer type (psa_algorithm_t, psa_key_usage_t, etc.) corresponding to a test function xxx_arg which has the type int. This avoids the need to cast xxx_arg to an unsigned type sometimes in the code.
2018-06-18 16:35:34 +02:00
input_data->x, input_data->len,
Asymmetric encryption tests: allow label argument Add a label argument to all asymmetric encryption test functions (currently empty in all tests, but that will change soon). In asymmetric_encrypt and asymmetric_decrypt, with an empty label, test with both a null pointer and a non-null pointer.
2018-06-30 18:42:41 +02:00
label->x, label->len,
Hygiene improvements in PSA crypto test code Get rid of many redundant casts. In particular, it is not useful to cast uint32_t values to size_t before performing arithmetic or comparisons on them. Rewrap a number of function calls, many of which now have narrower arguments thanks to the removed casts. When a function call doesn't fit on a single line, avoid grouping unrelated parameters together, but do try to group a buffer pointer and the associated size. Define more auxiliary variables xxx of a particular integer type (psa_algorithm_t, psa_key_usage_t, etc.) corresponding to a test function xxx_arg which has the type int. This avoids the need to cast xxx_arg to an unsigned type sometimes in the code.
2018-06-18 16:35:34 +02:00
output, output_size,
Normalize whitespace Normalize whitespace to Mbed TLS standards. There are only whitespace changes in this commit.
2018-06-18 15:41:12 +02:00
&output_length );
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( actual_status, expected_status );
Asymmetric encrypt/decrypt tests: check output length In asymmetric_encrypt_decrypt, use the buffer size advertized by the library for the ciphertext, and the length of the plaintext for the re-decrypted output. Test the output length if known. Require it to be 0 on error for encrypt/decrypt functions. If the output length is unknown, test at least that it's within the buffer limits.
2018-06-30 18:54:48 +02:00
TEST_ASSERT( output_length <= output_size );
added support for PKCSv1.5 signature verification and encryption/decryption and very basic tests.
2018-05-02 22:16:26 +02:00
Asymmetric encryption tests: allow label argument Add a label argument to all asymmetric encryption test functions (currently empty in all tests, but that will change soon). In asymmetric_encrypt and asymmetric_decrypt, with an empty label, test with both a null pointer and a non-null pointer.
2018-06-30 18:42:41 +02:00
/* If the label is empty, the test framework puts a non-null pointer
* in label->x. Test that a null pointer works as well. */
if( label->len == 0 )
{
output_length = ~0;
Skip calling memset when the size is 0 memset(NULL, c, 0) has undefined behavior, so don't do it. clang-asan complains.
2018-09-26 18:19:24 +02:00
if( output_size != 0 )
memset( output, 0, output_size );
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
actual_status = psa_asymmetric_decrypt( handle, alg,
Asymmetric encryption tests: allow label argument Add a label argument to all asymmetric encryption test functions (currently empty in all tests, but that will change soon). In asymmetric_encrypt and asymmetric_decrypt, with an empty label, test with both a null pointer and a non-null pointer.
2018-06-30 18:42:41 +02:00
input_data->x, input_data->len,
NULL, label->len,
output, output_size,
&output_length );
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( actual_status, expected_status );
Asymmetric encrypt/decrypt tests: check output length In asymmetric_encrypt_decrypt, use the buffer size advertized by the library for the ciphertext, and the length of the plaintext for the re-decrypted output. Test the output length if known. Require it to be 0 on error for encrypt/decrypt functions. If the output length is unknown, test at least that it's within the buffer limits.
2018-06-30 18:54:48 +02:00
TEST_ASSERT( output_length <= output_size );
Asymmetric encryption tests: allow label argument Add a label argument to all asymmetric encryption test functions (currently empty in all tests, but that will change soon). In asymmetric_encrypt and asymmetric_decrypt, with an empty label, test with both a null pointer and a non-null pointer.
2018-06-30 18:42:41 +02:00
}
added support for PKCSv1.5 signature verification and encryption/decryption and very basic tests.
2018-05-02 22:16:26 +02:00
exit:
Call psa_reset_key_attributes after psa_get_key_attributes After calling psa_get_key_attributes(), call psa_reset_key_attributes() if the key may have domain parameters, because that's the way to free the domain parameter substructure in the attribute structure. Keep not calling reset() in some places where the key can only be a symmetric key which doesn't have domain parameters.
2019-04-26 16:03:33 +02:00
psa_reset_key_attributes( &attributes );
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_destroy_key( handle );
Normalize whitespace Normalize whitespace to Mbed TLS standards. There are only whitespace changes in this commit.
2018-06-18 15:41:12 +02:00
mbedtls_free( output );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
added support for PKCSv1.5 signature verification and encryption/decryption and very basic tests.
2018-05-02 22:16:26 +02:00
}
Remove trailing whitespace Only horizontal whitespace changes in this commit.
2018-05-31 13:25:48 +02:00
/* END_CASE */
Implement psa_generate_random
2018-06-19 22:00:52 +02:00
psa: Test that generator initializers work
2019-01-04 15:11:48 +01:00
/* BEGIN_CASE */
Rename generator-related internal identifiers perl -pe 's/crypto_generator/key_derivation/gi' $(git ls-files) perl -pe 's/_generator/_key_derivation/gi' $(git ls-files)
2019-05-16 16:59:18 +02:00
void key_derivation_init( )
psa: Test that generator initializers work
2019-01-04 15:11:48 +01:00
{
/* Test each valid way of initializing the object, except for `= {0}`, as
* Clang 5 complains when `-Wmissing-field-initializers` is used, even
* though it's OK by the C standard. We could test for this, but we'd need
* to supress the Clang warning for the test. */
psa: Test fresh contexts have default behavior Test that freshly-initialized contexts exhibit default behavior through the API. Do this without depending on the internal representation of the contexts. This provides better portability of our tests on compilers like MSVC.
2019-02-07 17:33:37 +01:00
size_t capacity;
Rename generator functions to psa_key_derivation_xxx Generators are mostly about key derivation (currently: only about key derivation). "Generator" is not a commonly used term in cryptography. So favor "derivation" as terminology. Call a generator a key derivation operation structure, since it behaves like other multipart operation structures. Furthermore, the function names are not fully consistent. In this commit, I rename the functions to consistently have the prefix "psa_key_derivation_". I used the following command: perl -i -pe '%t = ( psa_crypto_generator_t => "psa_key_derivation_operation_t", psa_crypto_generator_init => "psa_key_derivation_init", psa_key_derivation_setup => "psa_key_derivation_setup", psa_key_derivation_input_key => "psa_key_derivation_input_key", psa_key_derivation_input_bytes => "psa_key_derivation_input_bytes", psa_key_agreement => "psa_key_derivation_key_agreement", psa_set_generator_capacity => "psa_key_derivation_set_capacity", psa_get_generator_capacity => "psa_key_derivation_get_capacity", psa_generator_read => "psa_key_derivation_output_bytes", psa_generate_derived_key => "psa_key_derivation_output_key", psa_generator_abort => "psa_key_derivation_abort", PSA_CRYPTO_GENERATOR_INIT => "PSA_KEY_DERIVATION_OPERATION_INIT", PSA_GENERATOR_UNBRIDLED_CAPACITY => "PSA_KEY_DERIVATION_UNLIMITED_CAPACITY", ); s/\b(@{[join("|", keys %t)]})\b/$t{$1}/ge' $(git ls-files)
2019-05-16 15:28:51 +02:00
psa_key_derivation_operation_t func = psa_key_derivation_operation_init( );
psa_key_derivation_operation_t init = PSA_KEY_DERIVATION_OPERATION_INIT;
psa_key_derivation_operation_t zero;
psa: Test that generator initializers work
2019-01-04 15:11:48 +01:00
memset( &zero, 0, sizeof( zero ) );
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
/* A default operation should not be able to report its capacity. */
Rename generator functions to psa_key_derivation_xxx Generators are mostly about key derivation (currently: only about key derivation). "Generator" is not a commonly used term in cryptography. So favor "derivation" as terminology. Call a generator a key derivation operation structure, since it behaves like other multipart operation structures. Furthermore, the function names are not fully consistent. In this commit, I rename the functions to consistently have the prefix "psa_key_derivation_". I used the following command: perl -i -pe '%t = ( psa_crypto_generator_t => "psa_key_derivation_operation_t", psa_crypto_generator_init => "psa_key_derivation_init", psa_key_derivation_setup => "psa_key_derivation_setup", psa_key_derivation_input_key => "psa_key_derivation_input_key", psa_key_derivation_input_bytes => "psa_key_derivation_input_bytes", psa_key_agreement => "psa_key_derivation_key_agreement", psa_set_generator_capacity => "psa_key_derivation_set_capacity", psa_get_generator_capacity => "psa_key_derivation_get_capacity", psa_generator_read => "psa_key_derivation_output_bytes", psa_generate_derived_key => "psa_key_derivation_output_key", psa_generator_abort => "psa_key_derivation_abort", PSA_CRYPTO_GENERATOR_INIT => "PSA_KEY_DERIVATION_OPERATION_INIT", PSA_GENERATOR_UNBRIDLED_CAPACITY => "PSA_KEY_DERIVATION_UNLIMITED_CAPACITY", ); s/\b(@{[join("|", keys %t)]})\b/$t{$1}/ge' $(git ls-files)
2019-05-16 15:28:51 +02:00
TEST_EQUAL( psa_key_derivation_get_capacity( &func, &capacity ),
psa: Check generator validity before read Check generator validity (i.e. that alg has been initialized) before allowing reads from the generator or allowing reads of the generator's capacity. This aligns our implementation with the documented error code behavior in our crypto.h and the PSA Crypto API.
2019-02-15 14:05:49 +01:00
PSA_ERROR_BAD_STATE );
Rename generator functions to psa_key_derivation_xxx Generators are mostly about key derivation (currently: only about key derivation). "Generator" is not a commonly used term in cryptography. So favor "derivation" as terminology. Call a generator a key derivation operation structure, since it behaves like other multipart operation structures. Furthermore, the function names are not fully consistent. In this commit, I rename the functions to consistently have the prefix "psa_key_derivation_". I used the following command: perl -i -pe '%t = ( psa_crypto_generator_t => "psa_key_derivation_operation_t", psa_crypto_generator_init => "psa_key_derivation_init", psa_key_derivation_setup => "psa_key_derivation_setup", psa_key_derivation_input_key => "psa_key_derivation_input_key", psa_key_derivation_input_bytes => "psa_key_derivation_input_bytes", psa_key_agreement => "psa_key_derivation_key_agreement", psa_set_generator_capacity => "psa_key_derivation_set_capacity", psa_get_generator_capacity => "psa_key_derivation_get_capacity", psa_generator_read => "psa_key_derivation_output_bytes", psa_generate_derived_key => "psa_key_derivation_output_key", psa_generator_abort => "psa_key_derivation_abort", PSA_CRYPTO_GENERATOR_INIT => "PSA_KEY_DERIVATION_OPERATION_INIT", PSA_GENERATOR_UNBRIDLED_CAPACITY => "PSA_KEY_DERIVATION_UNLIMITED_CAPACITY", ); s/\b(@{[join("|", keys %t)]})\b/$t{$1}/ge' $(git ls-files)
2019-05-16 15:28:51 +02:00
TEST_EQUAL( psa_key_derivation_get_capacity( &init, &capacity ),
psa: Check generator validity before read Check generator validity (i.e. that alg has been initialized) before allowing reads from the generator or allowing reads of the generator's capacity. This aligns our implementation with the documented error code behavior in our crypto.h and the PSA Crypto API.
2019-02-15 14:05:49 +01:00
PSA_ERROR_BAD_STATE );
Rename generator functions to psa_key_derivation_xxx Generators are mostly about key derivation (currently: only about key derivation). "Generator" is not a commonly used term in cryptography. So favor "derivation" as terminology. Call a generator a key derivation operation structure, since it behaves like other multipart operation structures. Furthermore, the function names are not fully consistent. In this commit, I rename the functions to consistently have the prefix "psa_key_derivation_". I used the following command: perl -i -pe '%t = ( psa_crypto_generator_t => "psa_key_derivation_operation_t", psa_crypto_generator_init => "psa_key_derivation_init", psa_key_derivation_setup => "psa_key_derivation_setup", psa_key_derivation_input_key => "psa_key_derivation_input_key", psa_key_derivation_input_bytes => "psa_key_derivation_input_bytes", psa_key_agreement => "psa_key_derivation_key_agreement", psa_set_generator_capacity => "psa_key_derivation_set_capacity", psa_get_generator_capacity => "psa_key_derivation_get_capacity", psa_generator_read => "psa_key_derivation_output_bytes", psa_generate_derived_key => "psa_key_derivation_output_key", psa_generator_abort => "psa_key_derivation_abort", PSA_CRYPTO_GENERATOR_INIT => "PSA_KEY_DERIVATION_OPERATION_INIT", PSA_GENERATOR_UNBRIDLED_CAPACITY => "PSA_KEY_DERIVATION_UNLIMITED_CAPACITY", ); s/\b(@{[join("|", keys %t)]})\b/$t{$1}/ge' $(git ls-files)
2019-05-16 15:28:51 +02:00
TEST_EQUAL( psa_key_derivation_get_capacity( &zero, &capacity ),
psa: Check generator validity before read Check generator validity (i.e. that alg has been initialized) before allowing reads from the generator or allowing reads of the generator's capacity. This aligns our implementation with the documented error code behavior in our crypto.h and the PSA Crypto API.
2019-02-15 14:05:49 +01:00
PSA_ERROR_BAD_STATE );
psa: Test fresh contexts have default behavior Test that freshly-initialized contexts exhibit default behavior through the API. Do this without depending on the internal representation of the contexts. This provides better portability of our tests on compilers like MSVC.
2019-02-07 17:33:37 +01:00
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
/* A default operation should be abortable without error. */
Rename generator functions to psa_key_derivation_xxx Generators are mostly about key derivation (currently: only about key derivation). "Generator" is not a commonly used term in cryptography. So favor "derivation" as terminology. Call a generator a key derivation operation structure, since it behaves like other multipart operation structures. Furthermore, the function names are not fully consistent. In this commit, I rename the functions to consistently have the prefix "psa_key_derivation_". I used the following command: perl -i -pe '%t = ( psa_crypto_generator_t => "psa_key_derivation_operation_t", psa_crypto_generator_init => "psa_key_derivation_init", psa_key_derivation_setup => "psa_key_derivation_setup", psa_key_derivation_input_key => "psa_key_derivation_input_key", psa_key_derivation_input_bytes => "psa_key_derivation_input_bytes", psa_key_agreement => "psa_key_derivation_key_agreement", psa_set_generator_capacity => "psa_key_derivation_set_capacity", psa_get_generator_capacity => "psa_key_derivation_get_capacity", psa_generator_read => "psa_key_derivation_output_bytes", psa_generate_derived_key => "psa_key_derivation_output_key", psa_generator_abort => "psa_key_derivation_abort", PSA_CRYPTO_GENERATOR_INIT => "PSA_KEY_DERIVATION_OPERATION_INIT", PSA_GENERATOR_UNBRIDLED_CAPACITY => "PSA_KEY_DERIVATION_UNLIMITED_CAPACITY", ); s/\b(@{[join("|", keys %t)]})\b/$t{$1}/ge' $(git ls-files)
2019-05-16 15:28:51 +02:00
PSA_ASSERT( psa_key_derivation_abort(&func) );
PSA_ASSERT( psa_key_derivation_abort(&init) );
PSA_ASSERT( psa_key_derivation_abort(&zero) );
psa: Test that generator initializers work
2019-01-04 15:11:48 +01:00
}
/* END_CASE */
Adapt the derive_setup tests to the new API Part of the tests are adapted in this commit, another part is already covered by the derive_input tests and some of them are not applicable to the new API (the new API does not request capacity at the setup stage). The test coverage temporarily drops with this commit, the two test cases conserning capacity will be re-added in a later commit.
2019-06-13 17:32:24 +02:00
/* BEGIN_CASE */
void derive_setup( int alg_arg, int expected_status_arg )
Add framework for simple key derivation New key type PSA_KEY_TYPE_DERIVE. New usage flag PSA_KEY_USAGE_DERIVE. New function psa_key_derivation. No key derivation algorithm is implemented yet. The code may not compile with -Wunused. Write some unit test code for psa_key_derivation. Most of it cannot be used yet due to the lack of a key derivation algorithm.
2018-07-12 17:17:20 +02:00
{
psa_algorithm_t alg = alg_arg;
psa_status_t expected_status = expected_status_arg;
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
psa_key_derivation_operation_t operation = PSA_KEY_DERIVATION_OPERATION_INIT;
Add framework for simple key derivation New key type PSA_KEY_TYPE_DERIVE. New usage flag PSA_KEY_USAGE_DERIVE. New function psa_key_derivation. No key derivation algorithm is implemented yet. The code may not compile with -Wunused. Write some unit test code for psa_key_derivation. Most of it cannot be used yet due to the lack of a key derivation algorithm.
2018-07-12 17:17:20 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
Add framework for simple key derivation New key type PSA_KEY_TYPE_DERIVE. New usage flag PSA_KEY_USAGE_DERIVE. New function psa_key_derivation. No key derivation algorithm is implemented yet. The code may not compile with -Wunused. Write some unit test code for psa_key_derivation. Most of it cannot be used yet due to the lack of a key derivation algorithm.
2018-07-12 17:17:20 +02:00
Adapt the derive_setup tests to the new API Part of the tests are adapted in this commit, another part is already covered by the derive_input tests and some of them are not applicable to the new API (the new API does not request capacity at the setup stage). The test coverage temporarily drops with this commit, the two test cases conserning capacity will be re-added in a later commit.
2019-06-13 17:32:24 +02:00
TEST_EQUAL( psa_key_derivation_setup( &operation, alg ),
Rewrap some lines after the macro changes Change the way some lines are wrapped to cut at a more logical place. This commit mainly rewrites multi-line calls to TEST_EQUAL, and also a few calls to PSA_ASSERT.
2018-12-18 00:33:25 +01:00
expected_status );
Add framework for simple key derivation New key type PSA_KEY_TYPE_DERIVE. New usage flag PSA_KEY_USAGE_DERIVE. New function psa_key_derivation. No key derivation algorithm is implemented yet. The code may not compile with -Wunused. Write some unit test code for psa_key_derivation. Most of it cannot be used yet due to the lack of a key derivation algorithm.
2018-07-12 17:17:20 +02:00
exit:
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
psa_key_derivation_abort( &operation );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
Add framework for simple key derivation New key type PSA_KEY_TYPE_DERIVE. New usage flag PSA_KEY_USAGE_DERIVE. New function psa_key_derivation. No key derivation algorithm is implemented yet. The code may not compile with -Wunused. Write some unit test code for psa_key_derivation. Most of it cannot be used yet due to the lack of a key derivation algorithm.
2018-07-12 17:17:20 +02:00
}
/* END_CASE */
Add test for psa_key_derivation_set_capacity This commit restores the test coverage to the level before adapting the derive_setup tests.
2019-06-14 10:59:36 +02:00
/* BEGIN_CASE */
void derive_set_capacity( int alg_arg, int capacity_arg,
int expected_status_arg )
{
psa_algorithm_t alg = alg_arg;
size_t capacity = capacity_arg;
psa_status_t expected_status = expected_status_arg;
psa_key_derivation_operation_t operation = PSA_KEY_DERIVATION_OPERATION_INIT;
PSA_ASSERT( psa_crypto_init( ) );
PSA_ASSERT( psa_key_derivation_setup( &operation, alg ) );
TEST_EQUAL( psa_key_derivation_set_capacity( &operation, capacity ),
expected_status );
exit:
psa_key_derivation_abort( &operation );
PSA_DONE( );
}
/* END_CASE */
Add a test for psa_key_derivation_input
2019-06-12 13:34:34 +02:00
/* BEGIN_CASE */
void derive_input( int alg_arg,
PSA crypto KDF: test bytes/key input independently of the step type This commit only makes derive_input more flexible so that the key derivation API can be tested with different key types and raw data for each input step. The behavior of the test cases remains the same.
2019-09-23 13:49:33 +02:00
int step_arg1, int key_type_arg1, data_t *input1,
Add a test for psa_key_derivation_input
2019-06-12 13:34:34 +02:00
int expected_status_arg1,
derive_input test function: More logical parameter order No behavior change.
2019-09-24 17:19:33 +02:00
int step_arg2, int key_type_arg2, data_t *input2,
Add a test for psa_key_derivation_input
2019-06-12 13:34:34 +02:00
int expected_status_arg2,
derive_input test function: More logical parameter order No behavior change.
2019-09-24 17:19:33 +02:00
int step_arg3, int key_type_arg3, data_t *input3,
derive_input test function: Try output afterwards After passing some inputs, try getting one byte of output, just to check that this succeeds (for a valid sequence of inputs) or fails with BAD_STATE (for an invalid sequence of inputs). Either output a 1-byte key or a 1-byte buffer depending on the test data. The test data was expanded as follows: * Output key type (or not a key): same as the SECRET input if success is expected, otherwise NONE. * Expected status: PSA_SUCCESS after valid inputs, BAD_STATE after any invalid input.
2019-09-24 17:45:07 +02:00
int expected_status_arg3,
int output_key_type_arg, int expected_output_status_arg )
Add a test for psa_key_derivation_input
2019-06-12 13:34:34 +02:00
{
psa_algorithm_t alg = alg_arg;
PSA crypto KDF: test bytes/key input independently of the step type This commit only makes derive_input more flexible so that the key derivation API can be tested with different key types and raw data for each input step. The behavior of the test cases remains the same.
2019-09-23 13:49:33 +02:00
psa_key_derivation_step_t steps[] = {step_arg1, step_arg2, step_arg3};
psa_key_type_t key_types[] = {key_type_arg1, key_type_arg2, key_type_arg3};
Add a test for psa_key_derivation_input
2019-06-12 13:34:34 +02:00
psa_status_t expected_statuses[] = {expected_status_arg1,
expected_status_arg2,
expected_status_arg3};
data_t *inputs[] = {input1, input2, input3};
psa_key_handle_t handles[] = {0, 0, 0};
psa_key_derivation_operation_t operation = PSA_KEY_DERIVATION_OPERATION_INIT;
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
size_t i;
derive_input test function: Try output afterwards After passing some inputs, try getting one byte of output, just to check that this succeeds (for a valid sequence of inputs) or fails with BAD_STATE (for an invalid sequence of inputs). Either output a 1-byte key or a 1-byte buffer depending on the test data. The test data was expanded as follows: * Output key type (or not a key): same as the SECRET input if success is expected, otherwise NONE. * Expected status: PSA_SUCCESS after valid inputs, BAD_STATE after any invalid input.
2019-09-24 17:45:07 +02:00
psa_key_type_t output_key_type = output_key_type_arg;
psa_key_handle_t output_handle = 0;
psa_status_t expected_output_status = expected_output_status_arg;
psa_status_t actual_output_status;
Add a test for psa_key_derivation_input
2019-06-12 13:34:34 +02:00
PSA_ASSERT( psa_crypto_init( ) );
psa_set_key_usage_flags( &attributes, PSA_KEY_USAGE_DERIVE );
psa_set_key_algorithm( &attributes, alg );
PSA_ASSERT( psa_key_derivation_setup( &operation, alg ) );
for( i = 0; i < ARRAY_LENGTH( steps ); i++ )
{
Use the constant PSA_KEY_TYPE_NONE rather than 0 No behavior change, just a readability improvement.
2019-09-24 16:21:10 +02:00
if( key_types[i] != PSA_KEY_TYPE_NONE )
PSA crypto KDF: test bytes/key input independently of the step type This commit only makes derive_input more flexible so that the key derivation API can be tested with different key types and raw data for each input step. The behavior of the test cases remains the same.
2019-09-23 13:49:33 +02:00
{
psa_set_key_type( &attributes, key_types[i] );
PSA_ASSERT( psa_import_key( &attributes,
inputs[i]->x, inputs[i]->len,
&handles[i] ) );
TEST_EQUAL( psa_key_derivation_input_key( &operation, steps[i],
handles[i] ),
expected_statuses[i] );
}
else
Add a test for psa_key_derivation_input
2019-06-12 13:34:34 +02:00
{
PSA crypto KDF: test bytes/key input independently of the step type This commit only makes derive_input more flexible so that the key derivation API can be tested with different key types and raw data for each input step. The behavior of the test cases remains the same.
2019-09-23 13:49:33 +02:00
TEST_EQUAL( psa_key_derivation_input_bytes(
&operation, steps[i],
inputs[i]->x, inputs[i]->len ),
expected_statuses[i] );
Add a test for psa_key_derivation_input
2019-06-12 13:34:34 +02:00
}
}
derive_input test function: Try output afterwards After passing some inputs, try getting one byte of output, just to check that this succeeds (for a valid sequence of inputs) or fails with BAD_STATE (for an invalid sequence of inputs). Either output a 1-byte key or a 1-byte buffer depending on the test data. The test data was expanded as follows: * Output key type (or not a key): same as the SECRET input if success is expected, otherwise NONE. * Expected status: PSA_SUCCESS after valid inputs, BAD_STATE after any invalid input.
2019-09-24 17:45:07 +02:00
if( output_key_type != PSA_KEY_TYPE_NONE )
{
psa_reset_key_attributes( &attributes );
psa_set_key_type( &attributes, PSA_KEY_TYPE_RAW_DATA );
psa_set_key_bits( &attributes, 8 );
actual_output_status =
psa_key_derivation_output_key( &attributes, &operation,
&output_handle );
}
else
{
uint8_t buffer[1];
actual_output_status =
psa_key_derivation_output_bytes( &operation,
buffer, sizeof( buffer ) );
}
TEST_EQUAL( actual_output_status, expected_output_status );
Add a test for psa_key_derivation_input
2019-06-12 13:34:34 +02:00
exit:
psa_key_derivation_abort( &operation );
for( i = 0; i < ARRAY_LENGTH( handles ); i++ )
psa_destroy_key( handles[i] );
derive_input test function: Try output afterwards After passing some inputs, try getting one byte of output, just to check that this succeeds (for a valid sequence of inputs) or fails with BAD_STATE (for an invalid sequence of inputs). Either output a 1-byte key or a 1-byte buffer depending on the test data. The test data was expanded as follows: * Output key type (or not a key): same as the SECRET input if success is expected, otherwise NONE. * Expected status: PSA_SUCCESS after valid inputs, BAD_STATE after any invalid input.
2019-09-24 17:45:07 +02:00
psa_destroy_key( output_handle );
Add a test for psa_key_derivation_input
2019-06-12 13:34:34 +02:00
PSA_DONE( );
}
/* END_CASE */
Convert invalid_key_derivation_state to new API Convert the test_derive_invalid_key_derivation_state() test to the new KDF API.
2019-07-03 16:02:16 +02:00
/* BEGIN_CASE */
void test_derive_invalid_key_derivation_state( int alg_arg )
add tests that increase key derivation code coverage slightly added tests that increase code coverage for the key derivation functions slightly by reaching error cases not covered before.
2018-10-22 16:24:55 +02:00
{
Convert invalid_key_derivation_state to new API Convert the test_derive_invalid_key_derivation_state() test to the new KDF API.
2019-07-03 16:02:16 +02:00
psa_algorithm_t alg = alg_arg;
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t handle = 0;
streamline test function API by removing parameter streamline the API for the test test_derive_invalid_generator_state by removing the key type paramter (it is assumed to always be PSA_KEY_TYPE_DERIVE)
2018-10-31 11:15:58 +01:00
size_t key_type = PSA_KEY_TYPE_DERIVE;
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
psa_key_derivation_operation_t operation = PSA_KEY_DERIVATION_OPERATION_INIT;
Convert invalid_key_derivation_state to new API Convert the test_derive_invalid_key_derivation_state() test to the new KDF API.
2019-07-03 16:02:16 +02:00
unsigned char input1[] = "Input 1";
size_t input1_length = sizeof( input1 );
unsigned char input2[] = "Input 2";
size_t input2_length = sizeof( input2 );
add additional generator tests and generalize key derivation test Key derivation test now uses an indirect way to test generator validity as the direct way previously used isn't compatible with the PSA IPC implementation. Additional bad path test for the generator added to check basic bad-path scenarios.
2018-10-25 13:46:09 +02:00
uint8_t buffer[42];
Fix code style and clarify issue comment * remove unneeded constants * clarify comment reference to issue 183 * add additional reference comment * fix brace spacing issues
2018-11-01 11:27:20 +01:00
size_t capacity = sizeof( buffer );
Streamline test function API by removing parameter streamline the API for the test test_derive_invalid_generator_state: by removing the key_data parameter. This parameter is not important for test flow and can be hard-coded.
2018-11-01 11:24:23 +01:00
const uint8_t key_data[22] = { 0x0b, 0x0b, 0x0b, 0x0b, 0x0b, 0x0b, 0x0b, 0x0b,
0x0b, 0x0b, 0x0b, 0x0b, 0x0b, 0x0b, 0x0b, 0x0b,
0x0b, 0x0b, 0x0b, 0x0b, 0x0b, 0x0b};
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
add additional generator tests and generalize key derivation test Key derivation test now uses an indirect way to test generator validity as the direct way previously used isn't compatible with the PSA IPC implementation. Additional bad path test for the generator added to check basic bad-path scenarios.
2018-10-25 13:46:09 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
add additional generator tests and generalize key derivation test Key derivation test now uses an indirect way to test generator validity as the direct way previously used isn't compatible with the PSA IPC implementation. Additional bad path test for the generator added to check basic bad-path scenarios.
2018-10-25 13:46:09 +02:00
Update remaining test cases to use key attributes Finish updating the tests to use psa_key_attributes_t and psa_import_key instead of psa_key_policy_t and psa_import_key_to_handle.
2019-04-19 19:58:20 +02:00
psa_set_key_usage_flags( &attributes, PSA_KEY_USAGE_DERIVE );
psa_set_key_algorithm( &attributes, alg );
psa_set_key_type( &attributes, key_type );
add tests that increase key derivation code coverage slightly added tests that increase code coverage for the key derivation functions slightly by reaching error cases not covered before.
2018-10-22 16:24:55 +02:00
Put handle parameter last: psa_import_key In psa_import_key, change the order of parameters to pass the pointer where the newly created handle will be stored last. This is consistent with most other library functions that put inputs before outputs.
2019-05-15 20:15:10 +02:00
PSA_ASSERT( psa_import_key( &attributes,
key_data, sizeof( key_data ),
&handle ) );
add additional generator tests and generalize key derivation test Key derivation test now uses an indirect way to test generator validity as the direct way previously used isn't compatible with the PSA IPC implementation. Additional bad path test for the generator added to check basic bad-path scenarios.
2018-10-25 13:46:09 +02:00
/* valid key derivation */
Convert invalid_key_derivation_state to new API Convert the test_derive_invalid_key_derivation_state() test to the new KDF API.
2019-07-03 16:02:16 +02:00
if( !setup_key_derivation_wrap( &operation, handle, alg,
input1, input1_length,
input2, input2_length,
capacity ) )
goto exit;
add additional generator tests and generalize key derivation test Key derivation test now uses an indirect way to test generator validity as the direct way previously used isn't compatible with the PSA IPC implementation. Additional bad path test for the generator added to check basic bad-path scenarios.
2018-10-25 13:46:09 +02:00
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
/* state of operation shouldn't allow additional generation */
Convert invalid_key_derivation_state to new API Convert the test_derive_invalid_key_derivation_state() test to the new KDF API.
2019-07-03 16:02:16 +02:00
TEST_EQUAL( psa_key_derivation_setup( &operation, alg ),
Rewrap some lines after the macro changes Change the way some lines are wrapped to cut at a more logical place. This commit mainly rewrites multi-line calls to TEST_EQUAL, and also a few calls to PSA_ASSERT.
2018-12-18 00:33:25 +01:00
PSA_ERROR_BAD_STATE );
add additional generator tests and generalize key derivation test Key derivation test now uses an indirect way to test generator validity as the direct way previously used isn't compatible with the PSA IPC implementation. Additional bad path test for the generator added to check basic bad-path scenarios.
2018-10-25 13:46:09 +02:00
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
PSA_ASSERT( psa_key_derivation_output_bytes( &operation, buffer, capacity ) );
add additional generator tests and generalize key derivation test Key derivation test now uses an indirect way to test generator validity as the direct way previously used isn't compatible with the PSA IPC implementation. Additional bad path test for the generator added to check basic bad-path scenarios.
2018-10-25 13:46:09 +02:00
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
TEST_EQUAL( psa_key_derivation_output_bytes( &operation, buffer, capacity ),
Replace PSA error code definitions with the ones defined in PSA spec
2019-02-14 12:48:10 +01:00
PSA_ERROR_INSUFFICIENT_DATA );
add additional generator tests and generalize key derivation test Key derivation test now uses an indirect way to test generator validity as the direct way previously used isn't compatible with the PSA IPC implementation. Additional bad path test for the generator added to check basic bad-path scenarios.
2018-10-25 13:46:09 +02:00
exit:
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
psa_key_derivation_abort( &operation );
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_destroy_key( handle );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
add additional generator tests and generalize key derivation test Key derivation test now uses an indirect way to test generator validity as the direct way previously used isn't compatible with the PSA IPC implementation. Additional bad path test for the generator added to check basic bad-path scenarios.
2018-10-25 13:46:09 +02:00
}
/* END_CASE */
/* BEGIN_CASE */
Rename generator-related internal identifiers perl -pe 's/crypto_generator/key_derivation/gi' $(git ls-files) perl -pe 's/_generator/_key_derivation/gi' $(git ls-files)
2019-05-16 16:59:18 +02:00
void test_derive_invalid_key_derivation_tests( )
add additional generator tests and generalize key derivation test Key derivation test now uses an indirect way to test generator validity as the direct way previously used isn't compatible with the PSA IPC implementation. Additional bad path test for the generator added to check basic bad-path scenarios.
2018-10-25 13:46:09 +02:00
{
uint8_t output_buffer[16];
size_t buffer_size = 16;
size_t capacity = 0;
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
psa_key_derivation_operation_t operation = PSA_KEY_DERIVATION_OPERATION_INIT;
add additional generator tests and generalize key derivation test Key derivation test now uses an indirect way to test generator validity as the direct way previously used isn't compatible with the PSA IPC implementation. Additional bad path test for the generator added to check basic bad-path scenarios.
2018-10-25 13:46:09 +02:00
Wrap and reindent some lines After renaming several identifiers, re-wrap and re-indent some lines to make the code prettier.
2019-05-16 17:53:40 +02:00
TEST_ASSERT( psa_key_derivation_output_bytes( &operation,
output_buffer, buffer_size )
psa: Check generator validity before read Check generator validity (i.e. that alg has been initialized) before allowing reads from the generator or allowing reads of the generator's capacity. This aligns our implementation with the documented error code behavior in our crypto.h and the PSA Crypto API.
2019-02-15 14:05:49 +01:00
== PSA_ERROR_BAD_STATE );
add additional generator tests and generalize key derivation test Key derivation test now uses an indirect way to test generator validity as the direct way previously used isn't compatible with the PSA IPC implementation. Additional bad path test for the generator added to check basic bad-path scenarios.
2018-10-25 13:46:09 +02:00
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
TEST_ASSERT( psa_key_derivation_get_capacity( &operation, &capacity )
psa: Check generator validity before read Check generator validity (i.e. that alg has been initialized) before allowing reads from the generator or allowing reads of the generator's capacity. This aligns our implementation with the documented error code behavior in our crypto.h and the PSA Crypto API.
2019-02-15 14:05:49 +01:00
== PSA_ERROR_BAD_STATE );
add additional generator tests and generalize key derivation test Key derivation test now uses an indirect way to test generator validity as the direct way previously used isn't compatible with the PSA IPC implementation. Additional bad path test for the generator added to check basic bad-path scenarios.
2018-10-25 13:46:09 +02:00
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
PSA_ASSERT( psa_key_derivation_abort( &operation ) );
add additional generator tests and generalize key derivation test Key derivation test now uses an indirect way to test generator validity as the direct way previously used isn't compatible with the PSA IPC implementation. Additional bad path test for the generator added to check basic bad-path scenarios.
2018-10-25 13:46:09 +02:00
Wrap and reindent some lines After renaming several identifiers, re-wrap and re-indent some lines to make the code prettier.
2019-05-16 17:53:40 +02:00
TEST_ASSERT( psa_key_derivation_output_bytes( &operation,
output_buffer, buffer_size )
psa: Check generator validity before read Check generator validity (i.e. that alg has been initialized) before allowing reads from the generator or allowing reads of the generator's capacity. This aligns our implementation with the documented error code behavior in our crypto.h and the PSA Crypto API.
2019-02-15 14:05:49 +01:00
== PSA_ERROR_BAD_STATE );
add additional generator tests and generalize key derivation test Key derivation test now uses an indirect way to test generator validity as the direct way previously used isn't compatible with the PSA IPC implementation. Additional bad path test for the generator added to check basic bad-path scenarios.
2018-10-25 13:46:09 +02:00
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
TEST_ASSERT( psa_key_derivation_get_capacity( &operation, &capacity )
psa: Check generator validity before read Check generator validity (i.e. that alg has been initialized) before allowing reads from the generator or allowing reads of the generator's capacity. This aligns our implementation with the documented error code behavior in our crypto.h and the PSA Crypto API.
2019-02-15 14:05:49 +01:00
== PSA_ERROR_BAD_STATE );
add additional generator tests and generalize key derivation test Key derivation test now uses an indirect way to test generator validity as the direct way previously used isn't compatible with the PSA IPC implementation. Additional bad path test for the generator added to check basic bad-path scenarios.
2018-10-25 13:46:09 +02:00
exit:
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
psa_key_derivation_abort( &operation );
add tests that increase key derivation code coverage slightly added tests that increase code coverage for the key derivation functions slightly by reaching error cases not covered before.
2018-10-22 16:24:55 +02:00
}
/* END_CASE */
HKDF: positive tests
2018-07-12 17:24:54 +02:00
/* BEGIN_CASE */
void derive_output( int alg_arg,
Convert derive_output to the new KDF API
2019-05-29 17:35:49 +02:00
int step1_arg, data_t *input1,
int step2_arg, data_t *input2,
int step3_arg, data_t *input3,
HKDF: positive tests
2018-07-12 17:24:54 +02:00
int requested_capacity_arg,
data_t *expected_output1,
data_t *expected_output2 )
{
psa_algorithm_t alg = alg_arg;
Convert derive_output to the new KDF API
2019-05-29 17:35:49 +02:00
psa_key_derivation_step_t steps[] = {step1_arg, step2_arg, step3_arg};
data_t *inputs[] = {input1, input2, input3};
psa_key_handle_t handles[] = {0, 0, 0};
HKDF: positive tests
2018-07-12 17:24:54 +02:00
size_t requested_capacity = requested_capacity_arg;
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
psa_key_derivation_operation_t operation = PSA_KEY_DERIVATION_OPERATION_INIT;
HKDF: positive tests
2018-07-12 17:24:54 +02:00
uint8_t *expected_outputs[2] =
{expected_output1->x, expected_output2->x};
size_t output_sizes[2] =
{expected_output1->len, expected_output2->len};
size_t output_buffer_size = 0;
uint8_t *output_buffer = NULL;
size_t expected_capacity;
size_t current_capacity;
Implement atomic-creation psa_{generate,generator_import}_key Implement the new, attribute-based psa_generate_key and psa_generator_import_key.
2019-04-18 12:53:30 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
HKDF: positive tests
2018-07-12 17:24:54 +02:00
psa_status_t status;
Convert derive_output to the new KDF API
2019-05-29 17:35:49 +02:00
size_t i;
HKDF: positive tests
2018-07-12 17:24:54 +02:00
for( i = 0; i < ARRAY_LENGTH( expected_outputs ); i++ )
{
if( output_sizes[i] > output_buffer_size )
output_buffer_size = output_sizes[i];
if( output_sizes[i] == 0 )
expected_outputs[i] = NULL;
}
Use ASSERT_ALLOC instead of mbedtls_calloc in PSA tests This commit resolves a bug whereby some test cases failed on systems where mbedtls_calloc returns NULL when the size of 0, because the test case asserted `pointer != NULL` regardless of the size.
2018-09-27 13:54:18 +02:00
ASSERT_ALLOC( output_buffer, output_buffer_size );
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
HKDF: positive tests
2018-07-12 17:24:54 +02:00
Implement atomic-creation psa_{generate,generator_import}_key Implement the new, attribute-based psa_generate_key and psa_generator_import_key.
2019-04-18 12:53:30 +02:00
psa_set_key_usage_flags( &attributes, PSA_KEY_USAGE_DERIVE );
psa_set_key_algorithm( &attributes, alg );
psa_set_key_type( &attributes, PSA_KEY_TYPE_DERIVE );
HKDF: positive tests
2018-07-12 17:24:54 +02:00
/* Extraction phase. */
Convert derive_output to the new KDF API
2019-05-29 17:35:49 +02:00
PSA_ASSERT( psa_key_derivation_setup( &operation, alg ) );
PSA_ASSERT( psa_key_derivation_set_capacity( &operation,
requested_capacity ) );
for( i = 0; i < ARRAY_LENGTH( steps ); i++ )
Key derivation by small input steps: proof-of-concept Document the new API. Keep the old one. Implement for HKDF. Use it in a few test cases. Key agreement is still unchanged.
2019-01-07 22:59:38 +01:00
{
Convert derive_output to the new KDF API
2019-05-29 17:35:49 +02:00
switch( steps[i] )
{
case 0:
break;
case PSA_KEY_DERIVATION_INPUT_SECRET:
PSA_ASSERT( psa_import_key( &attributes,
inputs[i]->x, inputs[i]->len,
&handles[i] ) );
PSA_ASSERT( psa_key_derivation_input_key(
&operation, steps[i],
handles[i] ) );
break;
default:
PSA_ASSERT( psa_key_derivation_input_bytes(
&operation, steps[i],
inputs[i]->x, inputs[i]->len ) );
break;
}
Key derivation by small input steps: proof-of-concept Document the new API. Keep the old one. Implement for HKDF. Use it in a few test cases. Key agreement is still unchanged.
2019-01-07 22:59:38 +01:00
}
Convert derive_output to the new KDF API
2019-05-29 17:35:49 +02:00
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
PSA_ASSERT( psa_key_derivation_get_capacity( &operation,
Wrap and reindent some lines After renaming several identifiers, re-wrap and re-indent some lines to make the code prettier.
2019-05-16 17:53:40 +02:00
&current_capacity ) );
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( current_capacity, requested_capacity );
HKDF: positive tests
2018-07-12 17:24:54 +02:00
expected_capacity = requested_capacity;
/* Expansion phase. */
for( i = 0; i < ARRAY_LENGTH( expected_outputs ); i++ )
{
/* Read some bytes. */
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
status = psa_key_derivation_output_bytes( &operation,
Wrap and reindent some lines After renaming several identifiers, re-wrap and re-indent some lines to make the code prettier.
2019-05-16 17:53:40 +02:00
output_buffer, output_sizes[i] );
HKDF: positive tests
2018-07-12 17:24:54 +02:00
if( expected_capacity == 0 && output_sizes[i] == 0 )
{
/* Reading 0 bytes when 0 bytes are available can go either way. */
TEST_ASSERT( status == PSA_SUCCESS ||
Replace PSA error code definitions with the ones defined in PSA spec
2019-02-14 12:48:10 +01:00
status == PSA_ERROR_INSUFFICIENT_DATA );
HKDF: positive tests
2018-07-12 17:24:54 +02:00
continue;
}
else if( expected_capacity == 0 ||
output_sizes[i] > expected_capacity )
{
/* Capacity exceeded. */
Replace PSA error code definitions with the ones defined in PSA spec
2019-02-14 12:48:10 +01:00
TEST_EQUAL( status, PSA_ERROR_INSUFFICIENT_DATA );
HKDF: positive tests
2018-07-12 17:24:54 +02:00
expected_capacity = 0;
continue;
}
/* Success. Check the read data. */
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( status );
HKDF: positive tests
2018-07-12 17:24:54 +02:00
if( output_sizes[i] != 0 )
Use ASSERT_COMPARE in preference to memcmp in PSA tests
2018-12-18 00:40:50 +01:00
ASSERT_COMPARE( output_buffer, output_sizes[i],
expected_outputs[i], output_sizes[i] );
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
/* Check the operation status. */
HKDF: positive tests
2018-07-12 17:24:54 +02:00
expected_capacity -= output_sizes[i];
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
PSA_ASSERT( psa_key_derivation_get_capacity( &operation,
Wrap and reindent some lines After renaming several identifiers, re-wrap and re-indent some lines to make the code prettier.
2019-05-16 17:53:40 +02:00
&current_capacity ) );
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( expected_capacity, current_capacity );
HKDF: positive tests
2018-07-12 17:24:54 +02:00
}
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
PSA_ASSERT( psa_key_derivation_abort( &operation ) );
HKDF: positive tests
2018-07-12 17:24:54 +02:00
exit:
mbedtls_free( output_buffer );
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
psa_key_derivation_abort( &operation );
Convert derive_output to the new KDF API
2019-05-29 17:35:49 +02:00
for( i = 0; i < ARRAY_LENGTH( handles ); i++ )
psa_destroy_key( handles[i] );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
HKDF: positive tests
2018-07-12 17:24:54 +02:00
}
/* END_CASE */
HKDF: be more robust if we reach the maximum ouptut length In psa_generator_hkdf_read, return BAD_STATE if we're trying to construct more output than the algorithm allows. This can't happen through the API due to the capacity limit, but it could potentially happen in an internal call. Also add a test case that verifies that we can set up HKDF with its maximum capacity and read up to the maximum capacity.
2018-07-17 21:06:59 +02:00
/* BEGIN_CASE */
void derive_full( int alg_arg,
data_t *key_data,
Convert derive_full test to the new KDF API
2019-06-25 14:24:52 +02:00
data_t *input1,
data_t *input2,
HKDF: be more robust if we reach the maximum ouptut length In psa_generator_hkdf_read, return BAD_STATE if we're trying to construct more output than the algorithm allows. This can't happen through the API due to the capacity limit, but it could potentially happen in an internal call. Also add a test case that verifies that we can set up HKDF with its maximum capacity and read up to the maximum capacity.
2018-07-17 21:06:59 +02:00
int requested_capacity_arg )
{
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t handle = 0;
HKDF: be more robust if we reach the maximum ouptut length In psa_generator_hkdf_read, return BAD_STATE if we're trying to construct more output than the algorithm allows. This can't happen through the API due to the capacity limit, but it could potentially happen in an internal call. Also add a test case that verifies that we can set up HKDF with its maximum capacity and read up to the maximum capacity.
2018-07-17 21:06:59 +02:00
psa_algorithm_t alg = alg_arg;
size_t requested_capacity = requested_capacity_arg;
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
psa_key_derivation_operation_t operation = PSA_KEY_DERIVATION_OPERATION_INIT;
HKDF: be more robust if we reach the maximum ouptut length In psa_generator_hkdf_read, return BAD_STATE if we're trying to construct more output than the algorithm allows. This can't happen through the API due to the capacity limit, but it could potentially happen in an internal call. Also add a test case that verifies that we can set up HKDF with its maximum capacity and read up to the maximum capacity.
2018-07-17 21:06:59 +02:00
unsigned char output_buffer[16];
size_t expected_capacity = requested_capacity;
size_t current_capacity;
Implement atomic-creation psa_{generate,generator_import}_key Implement the new, attribute-based psa_generate_key and psa_generator_import_key.
2019-04-18 12:53:30 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
HKDF: be more robust if we reach the maximum ouptut length In psa_generator_hkdf_read, return BAD_STATE if we're trying to construct more output than the algorithm allows. This can't happen through the API due to the capacity limit, but it could potentially happen in an internal call. Also add a test case that verifies that we can set up HKDF with its maximum capacity and read up to the maximum capacity.
2018-07-17 21:06:59 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
HKDF: be more robust if we reach the maximum ouptut length In psa_generator_hkdf_read, return BAD_STATE if we're trying to construct more output than the algorithm allows. This can't happen through the API due to the capacity limit, but it could potentially happen in an internal call. Also add a test case that verifies that we can set up HKDF with its maximum capacity and read up to the maximum capacity.
2018-07-17 21:06:59 +02:00
Implement atomic-creation psa_{generate,generator_import}_key Implement the new, attribute-based psa_generate_key and psa_generator_import_key.
2019-04-18 12:53:30 +02:00
psa_set_key_usage_flags( &attributes, PSA_KEY_USAGE_DERIVE );
psa_set_key_algorithm( &attributes, alg );
psa_set_key_type( &attributes, PSA_KEY_TYPE_DERIVE );
HKDF: be more robust if we reach the maximum ouptut length In psa_generator_hkdf_read, return BAD_STATE if we're trying to construct more output than the algorithm allows. This can't happen through the API due to the capacity limit, but it could potentially happen in an internal call. Also add a test case that verifies that we can set up HKDF with its maximum capacity and read up to the maximum capacity.
2018-07-17 21:06:59 +02:00
Split long lines after psa_import_key refactoring
2019-05-15 20:22:09 +02:00
PSA_ASSERT( psa_import_key( &attributes, key_data->x, key_data->len,
&handle ) );
HKDF: be more robust if we reach the maximum ouptut length In psa_generator_hkdf_read, return BAD_STATE if we're trying to construct more output than the algorithm allows. This can't happen through the API due to the capacity limit, but it could potentially happen in an internal call. Also add a test case that verifies that we can set up HKDF with its maximum capacity and read up to the maximum capacity.
2018-07-17 21:06:59 +02:00
Refactor key derivation setup in tests
2019-07-03 13:41:36 +02:00
if( !setup_key_derivation_wrap( &operation, handle, alg,
input1->x, input1->len,
input2->x, input2->len,
requested_capacity ) )
goto exit;
Convert derive_full test to the new KDF API
2019-06-25 14:24:52 +02:00
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
PSA_ASSERT( psa_key_derivation_get_capacity( &operation,
Wrap and reindent some lines After renaming several identifiers, re-wrap and re-indent some lines to make the code prettier.
2019-05-16 17:53:40 +02:00
&current_capacity ) );
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( current_capacity, expected_capacity );
HKDF: be more robust if we reach the maximum ouptut length In psa_generator_hkdf_read, return BAD_STATE if we're trying to construct more output than the algorithm allows. This can't happen through the API due to the capacity limit, but it could potentially happen in an internal call. Also add a test case that verifies that we can set up HKDF with its maximum capacity and read up to the maximum capacity.
2018-07-17 21:06:59 +02:00
/* Expansion phase. */
while( current_capacity > 0 )
{
size_t read_size = sizeof( output_buffer );
if( read_size > current_capacity )
read_size = current_capacity;
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
PSA_ASSERT( psa_key_derivation_output_bytes( &operation,
Wrap and reindent some lines After renaming several identifiers, re-wrap and re-indent some lines to make the code prettier.
2019-05-16 17:53:40 +02:00
output_buffer,
read_size ) );
HKDF: be more robust if we reach the maximum ouptut length In psa_generator_hkdf_read, return BAD_STATE if we're trying to construct more output than the algorithm allows. This can't happen through the API due to the capacity limit, but it could potentially happen in an internal call. Also add a test case that verifies that we can set up HKDF with its maximum capacity and read up to the maximum capacity.
2018-07-17 21:06:59 +02:00
expected_capacity -= read_size;
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
PSA_ASSERT( psa_key_derivation_get_capacity( &operation,
Wrap and reindent some lines After renaming several identifiers, re-wrap and re-indent some lines to make the code prettier.
2019-05-16 17:53:40 +02:00
&current_capacity ) );
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( current_capacity, expected_capacity );
HKDF: be more robust if we reach the maximum ouptut length In psa_generator_hkdf_read, return BAD_STATE if we're trying to construct more output than the algorithm allows. This can't happen through the API due to the capacity limit, but it could potentially happen in an internal call. Also add a test case that verifies that we can set up HKDF with its maximum capacity and read up to the maximum capacity.
2018-07-17 21:06:59 +02:00
}
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
/* Check that the operation refuses to go over capacity. */
TEST_EQUAL( psa_key_derivation_output_bytes( &operation, output_buffer, 1 ),
Replace PSA error code definitions with the ones defined in PSA spec
2019-02-14 12:48:10 +01:00
PSA_ERROR_INSUFFICIENT_DATA );
HKDF: be more robust if we reach the maximum ouptut length In psa_generator_hkdf_read, return BAD_STATE if we're trying to construct more output than the algorithm allows. This can't happen through the API due to the capacity limit, but it could potentially happen in an internal call. Also add a test case that verifies that we can set up HKDF with its maximum capacity and read up to the maximum capacity.
2018-07-17 21:06:59 +02:00
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
PSA_ASSERT( psa_key_derivation_abort( &operation ) );
HKDF: be more robust if we reach the maximum ouptut length In psa_generator_hkdf_read, return BAD_STATE if we're trying to construct more output than the algorithm allows. This can't happen through the API due to the capacity limit, but it could potentially happen in an internal call. Also add a test case that verifies that we can set up HKDF with its maximum capacity and read up to the maximum capacity.
2018-07-17 21:06:59 +02:00
exit:
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
psa_key_derivation_abort( &operation );
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_destroy_key( handle );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
HKDF: be more robust if we reach the maximum ouptut length In psa_generator_hkdf_read, return BAD_STATE if we're trying to construct more output than the algorithm allows. This can't happen through the API due to the capacity limit, but it could potentially happen in an internal call. Also add a test case that verifies that we can set up HKDF with its maximum capacity and read up to the maximum capacity.
2018-07-17 21:06:59 +02:00
}
/* END_CASE */
Convert derive_key_exercise to the new KDF API
2019-07-03 14:51:30 +02:00
/* BEGIN_CASE */
Key derivation: test deriving a key from the KDF output
2018-07-12 17:29:22 +02:00
void derive_key_exercise( int alg_arg,
data_t *key_data,
Convert derive_key_exercise to the new KDF API
2019-07-03 14:51:30 +02:00
data_t *input1,
data_t *input2,
Key derivation: test deriving a key from the KDF output
2018-07-12 17:29:22 +02:00
int derived_type_arg,
int derived_bits_arg,
int derived_usage_arg,
int derived_alg_arg )
{
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t base_handle = 0;
psa_key_handle_t derived_handle = 0;
Key derivation: test deriving a key from the KDF output
2018-07-12 17:29:22 +02:00
psa_algorithm_t alg = alg_arg;
psa_key_type_t derived_type = derived_type_arg;
size_t derived_bits = derived_bits_arg;
psa_key_usage_t derived_usage = derived_usage_arg;
psa_algorithm_t derived_alg = derived_alg_arg;
size_t capacity = PSA_BITS_TO_BYTES( derived_bits );
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
psa_key_derivation_operation_t operation = PSA_KEY_DERIVATION_OPERATION_INIT;
Implement atomic-creation psa_{generate,generator_import}_key Implement the new, attribute-based psa_generate_key and psa_generator_import_key.
2019-04-18 12:53:30 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
psa_key_attributes_t got_attributes = PSA_KEY_ATTRIBUTES_INIT;
Key derivation: test deriving a key from the KDF output
2018-07-12 17:29:22 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
Key derivation: test deriving a key from the KDF output
2018-07-12 17:29:22 +02:00
Implement atomic-creation psa_{generate,generator_import}_key Implement the new, attribute-based psa_generate_key and psa_generator_import_key.
2019-04-18 12:53:30 +02:00
psa_set_key_usage_flags( &attributes, PSA_KEY_USAGE_DERIVE );
psa_set_key_algorithm( &attributes, alg );
psa_set_key_type( &attributes, PSA_KEY_TYPE_DERIVE );
Split long lines after psa_import_key refactoring
2019-05-15 20:22:09 +02:00
PSA_ASSERT( psa_import_key( &attributes, key_data->x, key_data->len,
&base_handle ) );
Key derivation: test deriving a key from the KDF output
2018-07-12 17:29:22 +02:00
/* Derive a key. */
Convert derive_key_exercise to the new KDF API
2019-07-03 14:51:30 +02:00
if ( setup_key_derivation_wrap( &operation, base_handle, alg,
input1->x, input1->len,
input2->x, input2->len, capacity ) )
goto exit;
Implement atomic-creation psa_{generate,generator_import}_key Implement the new, attribute-based psa_generate_key and psa_generator_import_key.
2019-04-18 12:53:30 +02:00
psa_set_key_usage_flags( &attributes, derived_usage );
psa_set_key_algorithm( &attributes, derived_alg );
psa_set_key_type( &attributes, derived_type );
Set the key size as an attribute Instead of passing a separate parameter for the key size to psa_generate_key and psa_generator_import_key, set it through the attributes, like the key type and other metadata.
2019-04-26 13:49:28 +02:00
psa_set_key_bits( &attributes, derived_bits );
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
PSA_ASSERT( psa_key_derivation_output_key( &attributes, &operation,
Wrap and reindent some lines After renaming several identifiers, re-wrap and re-indent some lines to make the code prettier.
2019-05-16 17:53:40 +02:00
&derived_handle ) );
Key derivation: test deriving a key from the KDF output
2018-07-12 17:29:22 +02:00
/* Test the key information */
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
PSA_ASSERT( psa_get_key_attributes( derived_handle, &got_attributes ) );
TEST_EQUAL( psa_get_key_type( &got_attributes ), derived_type );
TEST_EQUAL( psa_get_key_bits( &got_attributes ), derived_bits );
Key derivation: test deriving a key from the KDF output
2018-07-12 17:29:22 +02:00
/* Exercise the derived key. */
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
if( ! exercise_key( derived_handle, derived_usage, derived_alg ) )
Key derivation: test deriving a key from the KDF output
2018-07-12 17:29:22 +02:00
goto exit;
exit:
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
psa_key_derivation_abort( &operation );
Call psa_reset_key_attributes after psa_get_key_attributes After calling psa_get_key_attributes(), call psa_reset_key_attributes() if the key may have domain parameters, because that's the way to free the domain parameter substructure in the attribute structure. Keep not calling reset() in some places where the key can only be a symmetric key which doesn't have domain parameters.
2019-04-26 16:03:33 +02:00
psa_reset_key_attributes( &got_attributes );
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_destroy_key( base_handle );
psa_destroy_key( derived_handle );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
Key derivation: test deriving a key from the KDF output
2018-07-12 17:29:22 +02:00
}
/* END_CASE */
Convert derive_key_export to the new KDF API
2019-07-03 15:17:09 +02:00
/* BEGIN_CASE */
Key derivation: test deriving a key from the KDF output
2018-07-12 17:29:22 +02:00
void derive_key_export( int alg_arg,
data_t *key_data,
Convert derive_key_export to the new KDF API
2019-07-03 15:17:09 +02:00
data_t *input1,
data_t *input2,
Key derivation: test deriving a key from the KDF output
2018-07-12 17:29:22 +02:00
int bytes1_arg,
int bytes2_arg )
{
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t base_handle = 0;
psa_key_handle_t derived_handle = 0;
Key derivation: test deriving a key from the KDF output
2018-07-12 17:29:22 +02:00
psa_algorithm_t alg = alg_arg;
size_t bytes1 = bytes1_arg;
size_t bytes2 = bytes2_arg;
size_t capacity = bytes1 + bytes2;
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
psa_key_derivation_operation_t operation = PSA_KEY_DERIVATION_OPERATION_INIT;
Use ASSERT_ALLOC instead of mbedtls_calloc in PSA tests This commit resolves a bug whereby some test cases failed on systems where mbedtls_calloc returns NULL when the size of 0, because the test case asserted `pointer != NULL` regardless of the size.
2018-09-27 13:54:18 +02:00
uint8_t *output_buffer = NULL;
uint8_t *export_buffer = NULL;
Implement atomic-creation psa_{generate,generator_import}_key Implement the new, attribute-based psa_generate_key and psa_generator_import_key.
2019-04-18 12:53:30 +02:00
psa_key_attributes_t base_attributes = PSA_KEY_ATTRIBUTES_INIT;
psa_key_attributes_t derived_attributes = PSA_KEY_ATTRIBUTES_INIT;
Key derivation: test deriving a key from the KDF output
2018-07-12 17:29:22 +02:00
size_t length;
Use ASSERT_ALLOC instead of mbedtls_calloc in PSA tests This commit resolves a bug whereby some test cases failed on systems where mbedtls_calloc returns NULL when the size of 0, because the test case asserted `pointer != NULL` regardless of the size.
2018-09-27 13:54:18 +02:00
ASSERT_ALLOC( output_buffer, capacity );
ASSERT_ALLOC( export_buffer, capacity );
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
Key derivation: test deriving a key from the KDF output
2018-07-12 17:29:22 +02:00
Implement atomic-creation psa_{generate,generator_import}_key Implement the new, attribute-based psa_generate_key and psa_generator_import_key.
2019-04-18 12:53:30 +02:00
psa_set_key_usage_flags( &base_attributes, PSA_KEY_USAGE_DERIVE );
psa_set_key_algorithm( &base_attributes, alg );
psa_set_key_type( &base_attributes, PSA_KEY_TYPE_DERIVE );
Split long lines after psa_import_key refactoring
2019-05-15 20:22:09 +02:00
PSA_ASSERT( psa_import_key( &base_attributes, key_data->x, key_data->len,
&base_handle ) );
Key derivation: test deriving a key from the KDF output
2018-07-12 17:29:22 +02:00
/* Derive some material and output it. */
Convert derive_key_export to the new KDF API
2019-07-03 15:17:09 +02:00
if( !setup_key_derivation_wrap( &operation, base_handle, alg,
input1->x, input1->len,
input2->x, input2->len, capacity ) )
goto exit;
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
PSA_ASSERT( psa_key_derivation_output_bytes( &operation,
Wrap and reindent some lines After renaming several identifiers, re-wrap and re-indent some lines to make the code prettier.
2019-05-16 17:53:40 +02:00
output_buffer,
capacity ) );
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
PSA_ASSERT( psa_key_derivation_abort( &operation ) );
Key derivation: test deriving a key from the KDF output
2018-07-12 17:29:22 +02:00
/* Derive the same output again, but this time store it in key objects. */
Convert derive_key_export to the new KDF API
2019-07-03 15:17:09 +02:00
if( !setup_key_derivation_wrap( &operation, base_handle, alg,
input1->x, input1->len,
input2->x, input2->len, capacity ) )
goto exit;
Implement atomic-creation psa_{generate,generator_import}_key Implement the new, attribute-based psa_generate_key and psa_generator_import_key.
2019-04-18 12:53:30 +02:00
psa_set_key_usage_flags( &derived_attributes, PSA_KEY_USAGE_EXPORT );
psa_set_key_algorithm( &derived_attributes, 0 );
psa_set_key_type( &derived_attributes, PSA_KEY_TYPE_RAW_DATA );
Set the key size as an attribute Instead of passing a separate parameter for the key size to psa_generate_key and psa_generator_import_key, set it through the attributes, like the key type and other metadata.
2019-04-26 13:49:28 +02:00
psa_set_key_bits( &derived_attributes, PSA_BYTES_TO_BITS( bytes1 ) );
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
PSA_ASSERT( psa_key_derivation_output_key( &derived_attributes, &operation,
Wrap and reindent some lines After renaming several identifiers, re-wrap and re-indent some lines to make the code prettier.
2019-05-16 17:53:40 +02:00
&derived_handle ) );
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_export_key( derived_handle,
export_buffer, bytes1,
&length ) );
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( length, bytes1 );
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_destroy_key( derived_handle ) );
Set the key size as an attribute Instead of passing a separate parameter for the key size to psa_generate_key and psa_generator_import_key, set it through the attributes, like the key type and other metadata.
2019-04-26 13:49:28 +02:00
psa_set_key_bits( &derived_attributes, PSA_BYTES_TO_BITS( bytes2 ) );
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
PSA_ASSERT( psa_key_derivation_output_key( &derived_attributes, &operation,
Wrap and reindent some lines After renaming several identifiers, re-wrap and re-indent some lines to make the code prettier.
2019-05-16 17:53:40 +02:00
&derived_handle ) );
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_export_key( derived_handle,
export_buffer + bytes1, bytes2,
&length ) );
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( length, bytes2 );
Key derivation: test deriving a key from the KDF output
2018-07-12 17:29:22 +02:00
/* Compare the outputs from the two runs. */
Use ASSERT_COMPARE in preference to memcmp in PSA tests
2018-12-18 00:40:50 +01:00
ASSERT_COMPARE( output_buffer, bytes1 + bytes2,
export_buffer, capacity );
Key derivation: test deriving a key from the KDF output
2018-07-12 17:29:22 +02:00
exit:
mbedtls_free( output_buffer );
mbedtls_free( export_buffer );
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
psa_key_derivation_abort( &operation );
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_destroy_key( base_handle );
Limit keys to 65528 bits 65528 bits is more than any reasonable key until we start supporting post-quantum cryptography. This limit is chosen to allow bit-sizes to be stored in 16 bits, with 65535 left to indicate an invalid value. It's a whole number of bytes, which facilitates some calculations, in particular allowing a key of exactly PSA_CRYPTO_MAX_STORAGE_SIZE to be created but not one bit more. As a resource usage limit, this is arguably too large, but that's out of scope of the current commit. Test that key import, generation and derivation reject overly large sizes.
2019-07-30 17:26:54 +02:00
psa_destroy_key( derived_handle );
PSA_DONE( );
}
/* END_CASE */
/* BEGIN_CASE */
Test key creation with an invalid type (0 and nonzero)
2019-07-31 15:14:44 +02:00
void derive_key( int alg_arg,
data_t *key_data, data_t *input1, data_t *input2,
int type_arg, int bits_arg,
int expected_status_arg )
Limit keys to 65528 bits 65528 bits is more than any reasonable key until we start supporting post-quantum cryptography. This limit is chosen to allow bit-sizes to be stored in 16 bits, with 65535 left to indicate an invalid value. It's a whole number of bytes, which facilitates some calculations, in particular allowing a key of exactly PSA_CRYPTO_MAX_STORAGE_SIZE to be created but not one bit more. As a resource usage limit, this is arguably too large, but that's out of scope of the current commit. Test that key import, generation and derivation reject overly large sizes.
2019-07-30 17:26:54 +02:00
{
psa_key_handle_t base_handle = 0;
psa_key_handle_t derived_handle = 0;
psa_algorithm_t alg = alg_arg;
Test key creation with an invalid type (0 and nonzero)
2019-07-31 15:14:44 +02:00
psa_key_type_t type = type_arg;
Limit keys to 65528 bits 65528 bits is more than any reasonable key until we start supporting post-quantum cryptography. This limit is chosen to allow bit-sizes to be stored in 16 bits, with 65535 left to indicate an invalid value. It's a whole number of bytes, which facilitates some calculations, in particular allowing a key of exactly PSA_CRYPTO_MAX_STORAGE_SIZE to be created but not one bit more. As a resource usage limit, this is arguably too large, but that's out of scope of the current commit. Test that key import, generation and derivation reject overly large sizes.
2019-07-30 17:26:54 +02:00
size_t bits = bits_arg;
psa_status_t expected_status = expected_status_arg;
psa_key_derivation_operation_t operation = PSA_KEY_DERIVATION_OPERATION_INIT;
psa_key_attributes_t base_attributes = PSA_KEY_ATTRIBUTES_INIT;
psa_key_attributes_t derived_attributes = PSA_KEY_ATTRIBUTES_INIT;
PSA_ASSERT( psa_crypto_init( ) );
psa_set_key_usage_flags( &base_attributes, PSA_KEY_USAGE_DERIVE );
psa_set_key_algorithm( &base_attributes, alg );
psa_set_key_type( &base_attributes, PSA_KEY_TYPE_DERIVE );
PSA_ASSERT( psa_import_key( &base_attributes, key_data->x, key_data->len,
&base_handle ) );
if( !setup_key_derivation_wrap( &operation, base_handle, alg,
input1->x, input1->len,
input2->x, input2->len, SIZE_MAX ) )
goto exit;
psa_set_key_usage_flags( &derived_attributes, PSA_KEY_USAGE_EXPORT );
psa_set_key_algorithm( &derived_attributes, 0 );
Test key creation with an invalid type (0 and nonzero)
2019-07-31 15:14:44 +02:00
psa_set_key_type( &derived_attributes, type );
Limit keys to 65528 bits 65528 bits is more than any reasonable key until we start supporting post-quantum cryptography. This limit is chosen to allow bit-sizes to be stored in 16 bits, with 65535 left to indicate an invalid value. It's a whole number of bytes, which facilitates some calculations, in particular allowing a key of exactly PSA_CRYPTO_MAX_STORAGE_SIZE to be created but not one bit more. As a resource usage limit, this is arguably too large, but that's out of scope of the current commit. Test that key import, generation and derivation reject overly large sizes.
2019-07-30 17:26:54 +02:00
psa_set_key_bits( &derived_attributes, bits );
TEST_EQUAL( psa_key_derivation_output_key( &derived_attributes, &operation,
&derived_handle ),
expected_status );
exit:
psa_key_derivation_abort( &operation );
psa_destroy_key( base_handle );
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_destroy_key( derived_handle );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
Key derivation: test deriving a key from the KDF output
2018-07-12 17:29:22 +02:00
}
/* END_CASE */
New API function: psa_key_agreement Set up a generator from a key agreement.
2018-09-18 12:01:02 +02:00
/* BEGIN_CASE */
void key_agreement_setup( int alg_arg,
int our_key_type_arg, data_t *our_key_data,
data_t *peer_key_data,
int expected_status_arg )
{
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t our_key = 0;
New API function: psa_key_agreement Set up a generator from a key agreement.
2018-09-18 12:01:02 +02:00
psa_algorithm_t alg = alg_arg;
psa_key_type_t our_key_type = our_key_type_arg;
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
psa_key_derivation_operation_t operation = PSA_KEY_DERIVATION_OPERATION_INIT;
Implement atomic-creation psa_{generate,generator_import}_key Implement the new, attribute-based psa_generate_key and psa_generator_import_key.
2019-04-18 12:53:30 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
Quick fix of key agreement setup tests for the new derivation API Allow either the key derivation step or the key agreement step to fail. These tests should be split into three groups: key derivation setup tests with an algorithm that includes a key agreement step, and multipart key agreement failure tests, and raw key agreement failure tests.
2019-04-11 21:27:06 +02:00
psa_status_t expected_status = expected_status_arg;
psa_status_t status;
New API function: psa_key_agreement Set up a generator from a key agreement.
2018-09-18 12:01:02 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
New API function: psa_key_agreement Set up a generator from a key agreement.
2018-09-18 12:01:02 +02:00
Implement atomic-creation psa_{generate,generator_import}_key Implement the new, attribute-based psa_generate_key and psa_generator_import_key.
2019-04-18 12:53:30 +02:00
psa_set_key_usage_flags( &attributes, PSA_KEY_USAGE_DERIVE );
psa_set_key_algorithm( &attributes, alg );
psa_set_key_type( &attributes, our_key_type );
Split long lines after psa_import_key refactoring
2019-05-15 20:22:09 +02:00
PSA_ASSERT( psa_import_key( &attributes,
our_key_data->x, our_key_data->len,
&our_key ) );
New API function: psa_key_agreement Set up a generator from a key agreement.
2018-09-18 12:01:02 +02:00
Quick fix of key agreement setup tests for the new derivation API Allow either the key derivation step or the key agreement step to fail. These tests should be split into three groups: key derivation setup tests with an algorithm that includes a key agreement step, and multipart key agreement failure tests, and raw key agreement failure tests.
2019-04-11 21:27:06 +02:00
/* The tests currently include inputs that should fail at either step.
* Test cases that fail at the setup step should be changed to call
* key_derivation_setup instead, and this function should be renamed
* to key_agreement_fail. */
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
status = psa_key_derivation_setup( &operation, alg );
Quick fix of key agreement setup tests for the new derivation API Allow either the key derivation step or the key agreement step to fail. These tests should be split into three groups: key derivation setup tests with an algorithm that includes a key agreement step, and multipart key agreement failure tests, and raw key agreement failure tests.
2019-04-11 21:27:06 +02:00
if( status == PSA_SUCCESS )
{
Wrap and reindent some lines After renaming several identifiers, re-wrap and re-indent some lines to make the code prettier.
2019-05-16 17:53:40 +02:00
TEST_EQUAL( psa_key_derivation_key_agreement(
&operation, PSA_KEY_DERIVATION_INPUT_SECRET,
our_key,
peer_key_data->x, peer_key_data->len ),
Quick fix of key agreement setup tests for the new derivation API Allow either the key derivation step or the key agreement step to fail. These tests should be split into three groups: key derivation setup tests with an algorithm that includes a key agreement step, and multipart key agreement failure tests, and raw key agreement failure tests.
2019-04-11 21:27:06 +02:00
expected_status );
}
else
{
TEST_ASSERT( status == expected_status );
}
New API function: psa_key_agreement Set up a generator from a key agreement.
2018-09-18 12:01:02 +02:00
exit:
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
psa_key_derivation_abort( &operation );
New API function: psa_key_agreement Set up a generator from a key agreement.
2018-09-18 12:01:02 +02:00
psa_destroy_key( our_key );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
New API function: psa_key_agreement Set up a generator from a key agreement.
2018-09-18 12:01:02 +02:00
}
/* END_CASE */
New test function for raw agreement Change test cases with test data for raw agreement to this new test function.
2019-04-11 22:12:38 +02:00
/* BEGIN_CASE */
void raw_key_agreement( int alg_arg,
int our_key_type_arg, data_t *our_key_data,
data_t *peer_key_data,
data_t *expected_output )
{
psa_key_handle_t our_key = 0;
psa_algorithm_t alg = alg_arg;
psa_key_type_t our_key_type = our_key_type_arg;
Implement atomic-creation psa_{generate,generator_import}_key Implement the new, attribute-based psa_generate_key and psa_generator_import_key.
2019-04-18 12:53:30 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
New test function for raw agreement Change test cases with test data for raw agreement to this new test function.
2019-04-11 22:12:38 +02:00
unsigned char *output = NULL;
size_t output_length = ~0;
ASSERT_ALLOC( output, expected_output->len );
PSA_ASSERT( psa_crypto_init( ) );
Implement atomic-creation psa_{generate,generator_import}_key Implement the new, attribute-based psa_generate_key and psa_generator_import_key.
2019-04-18 12:53:30 +02:00
psa_set_key_usage_flags( &attributes, PSA_KEY_USAGE_DERIVE );
psa_set_key_algorithm( &attributes, alg );
psa_set_key_type( &attributes, our_key_type );
Split long lines after psa_import_key refactoring
2019-05-15 20:22:09 +02:00
PSA_ASSERT( psa_import_key( &attributes,
our_key_data->x, our_key_data->len,
&our_key ) );
New test function for raw agreement Change test cases with test data for raw agreement to this new test function.
2019-04-11 22:12:38 +02:00
Shorten the name of psa_key_agreement_raw_shared_secret There is less of a risk of confusion with the KA+KDF function now.
2019-05-16 18:00:41 +02:00
PSA_ASSERT( psa_raw_key_agreement( alg, our_key,
peer_key_data->x, peer_key_data->len,
output, expected_output->len,
&output_length ) );
New test function for raw agreement Change test cases with test data for raw agreement to this new test function.
2019-04-11 22:12:38 +02:00
ASSERT_COMPARE( output, output_length,
expected_output->x, expected_output->len );
exit:
mbedtls_free( output );
psa_destroy_key( our_key );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
New test function for raw agreement Change test cases with test data for raw agreement to this new test function.
2019-04-11 22:12:38 +02:00
}
/* END_CASE */
Key agreement test functions
2018-09-18 12:11:34 +02:00
/* BEGIN_CASE */
void key_agreement_capacity( int alg_arg,
int our_key_type_arg, data_t *our_key_data,
data_t *peer_key_data,
int expected_capacity_arg )
{
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t our_key = 0;
Key agreement test functions
2018-09-18 12:11:34 +02:00
psa_algorithm_t alg = alg_arg;
psa_key_type_t our_key_type = our_key_type_arg;
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
psa_key_derivation_operation_t operation = PSA_KEY_DERIVATION_OPERATION_INIT;
Implement atomic-creation psa_{generate,generator_import}_key Implement the new, attribute-based psa_generate_key and psa_generator_import_key.
2019-04-18 12:53:30 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
Key agreement test functions
2018-09-18 12:11:34 +02:00
size_t actual_capacity;
key_agreement_capacity: test the actual capacity as well After testing that the advertized capacity is what the test data says, read that many bytes to test that this is also actual capacity.
2018-10-25 22:36:12 +02:00
unsigned char output[16];
Key agreement test functions
2018-09-18 12:11:34 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
Key agreement test functions
2018-09-18 12:11:34 +02:00
Implement atomic-creation psa_{generate,generator_import}_key Implement the new, attribute-based psa_generate_key and psa_generator_import_key.
2019-04-18 12:53:30 +02:00
psa_set_key_usage_flags( &attributes, PSA_KEY_USAGE_DERIVE );
psa_set_key_algorithm( &attributes, alg );
psa_set_key_type( &attributes, our_key_type );
Split long lines after psa_import_key refactoring
2019-05-15 20:22:09 +02:00
PSA_ASSERT( psa_import_key( &attributes,
our_key_data->x, our_key_data->len,
&our_key ) );
Key agreement test functions
2018-09-18 12:11:34 +02:00
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
PSA_ASSERT( psa_key_derivation_setup( &operation, alg ) );
Wrap and reindent some lines After renaming several identifiers, re-wrap and re-indent some lines to make the code prettier.
2019-05-16 17:53:40 +02:00
PSA_ASSERT( psa_key_derivation_key_agreement(
&operation,
PSA_KEY_DERIVATION_INPUT_SECRET, our_key,
peer_key_data->x, peer_key_data->len ) );
Test multipart key agreement with ECDH+HKDF Basic coverage with one algorithm only and a restricted choice of output lengths.
2019-04-11 22:13:20 +02:00
if( PSA_ALG_IS_HKDF( PSA_ALG_KEY_AGREEMENT_GET_KDF( alg ) ) )
{
/* The test data is for info="" */
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
PSA_ASSERT( psa_key_derivation_input_bytes( &operation,
Rename PSA_KDF_STEP_xxx -> PSA_KEY_DERIVATION_INPUT_xxx More consistent with the new function names.
2019-05-16 16:05:19 +02:00
PSA_KEY_DERIVATION_INPUT_INFO,
Test multipart key agreement with ECDH+HKDF Basic coverage with one algorithm only and a restricted choice of output lengths.
2019-04-11 22:13:20 +02:00
NULL, 0 ) );
}
Key agreement test functions
2018-09-18 12:11:34 +02:00
key_agreement_capacity: test the actual capacity as well After testing that the advertized capacity is what the test data says, read that many bytes to test that this is also actual capacity.
2018-10-25 22:36:12 +02:00
/* Test the advertized capacity. */
Rename generator functions to psa_key_derivation_xxx Generators are mostly about key derivation (currently: only about key derivation). "Generator" is not a commonly used term in cryptography. So favor "derivation" as terminology. Call a generator a key derivation operation structure, since it behaves like other multipart operation structures. Furthermore, the function names are not fully consistent. In this commit, I rename the functions to consistently have the prefix "psa_key_derivation_". I used the following command: perl -i -pe '%t = ( psa_crypto_generator_t => "psa_key_derivation_operation_t", psa_crypto_generator_init => "psa_key_derivation_init", psa_key_derivation_setup => "psa_key_derivation_setup", psa_key_derivation_input_key => "psa_key_derivation_input_key", psa_key_derivation_input_bytes => "psa_key_derivation_input_bytes", psa_key_agreement => "psa_key_derivation_key_agreement", psa_set_generator_capacity => "psa_key_derivation_set_capacity", psa_get_generator_capacity => "psa_key_derivation_get_capacity", psa_generator_read => "psa_key_derivation_output_bytes", psa_generate_derived_key => "psa_key_derivation_output_key", psa_generator_abort => "psa_key_derivation_abort", PSA_CRYPTO_GENERATOR_INIT => "PSA_KEY_DERIVATION_OPERATION_INIT", PSA_GENERATOR_UNBRIDLED_CAPACITY => "PSA_KEY_DERIVATION_UNLIMITED_CAPACITY", ); s/\b(@{[join("|", keys %t)]})\b/$t{$1}/ge' $(git ls-files)
2019-05-16 15:28:51 +02:00
PSA_ASSERT( psa_key_derivation_get_capacity(
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
&operation, &actual_capacity ) );
Use TEST_EQUAL(a,b) in preference to TEST_ASSERT(a==b) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==([^;=]*)\);$/${1}TEST_EQUAL($2,$3);/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:24:04 +01:00
TEST_EQUAL( actual_capacity, (size_t) expected_capacity_arg );
Key agreement test functions
2018-09-18 12:11:34 +02:00
key_agreement_capacity: test the actual capacity as well After testing that the advertized capacity is what the test data says, read that many bytes to test that this is also actual capacity.
2018-10-25 22:36:12 +02:00
/* Test the actual capacity by reading the output. */
while( actual_capacity > sizeof( output ) )
{
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
PSA_ASSERT( psa_key_derivation_output_bytes( &operation,
Wrap and reindent some lines After renaming several identifiers, re-wrap and re-indent some lines to make the code prettier.
2019-05-16 17:53:40 +02:00
output, sizeof( output ) ) );
key_agreement_capacity: test the actual capacity as well After testing that the advertized capacity is what the test data says, read that many bytes to test that this is also actual capacity.
2018-10-25 22:36:12 +02:00
actual_capacity -= sizeof( output );
}
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
PSA_ASSERT( psa_key_derivation_output_bytes( &operation,
Wrap and reindent some lines After renaming several identifiers, re-wrap and re-indent some lines to make the code prettier.
2019-05-16 17:53:40 +02:00
output, actual_capacity ) );
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
TEST_EQUAL( psa_key_derivation_output_bytes( &operation, output, 1 ),
Replace PSA error code definitions with the ones defined in PSA spec
2019-02-14 12:48:10 +01:00
PSA_ERROR_INSUFFICIENT_DATA );
key_agreement_capacity: test the actual capacity as well After testing that the advertized capacity is what the test data says, read that many bytes to test that this is also actual capacity.
2018-10-25 22:36:12 +02:00
Key agreement test functions
2018-09-18 12:11:34 +02:00
exit:
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
psa_key_derivation_abort( &operation );
Key agreement test functions
2018-09-18 12:11:34 +02:00
psa_destroy_key( our_key );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
Key agreement test functions
2018-09-18 12:11:34 +02:00
}
/* END_CASE */
/* BEGIN_CASE */
void key_agreement_output( int alg_arg,
int our_key_type_arg, data_t *our_key_data,
data_t *peer_key_data,
Add multipart key agreement tests Add test cases that do key agreement with raw selection in pieces, to validate that selection works even when the application doesn't read everything in one chunk.
2018-10-25 22:37:15 +02:00
data_t *expected_output1, data_t *expected_output2 )
Key agreement test functions
2018-09-18 12:11:34 +02:00
{
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t our_key = 0;
Key agreement test functions
2018-09-18 12:11:34 +02:00
psa_algorithm_t alg = alg_arg;
psa_key_type_t our_key_type = our_key_type_arg;
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
psa_key_derivation_operation_t operation = PSA_KEY_DERIVATION_OPERATION_INIT;
Implement atomic-creation psa_{generate,generator_import}_key Implement the new, attribute-based psa_generate_key and psa_generator_import_key.
2019-04-18 12:53:30 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
Add multipart key agreement tests Add test cases that do key agreement with raw selection in pieces, to validate that selection works even when the application doesn't read everything in one chunk.
2018-10-25 22:37:15 +02:00
uint8_t *actual_output = NULL;
Key agreement test functions
2018-09-18 12:11:34 +02:00
Add multipart key agreement tests Add test cases that do key agreement with raw selection in pieces, to validate that selection works even when the application doesn't read everything in one chunk.
2018-10-25 22:37:15 +02:00
ASSERT_ALLOC( actual_output, MAX( expected_output1->len,
expected_output2->len ) );
Key agreement test functions
2018-09-18 12:11:34 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
Key agreement test functions
2018-09-18 12:11:34 +02:00
Implement atomic-creation psa_{generate,generator_import}_key Implement the new, attribute-based psa_generate_key and psa_generator_import_key.
2019-04-18 12:53:30 +02:00
psa_set_key_usage_flags( &attributes, PSA_KEY_USAGE_DERIVE );
psa_set_key_algorithm( &attributes, alg );
psa_set_key_type( &attributes, our_key_type );
Split long lines after psa_import_key refactoring
2019-05-15 20:22:09 +02:00
PSA_ASSERT( psa_import_key( &attributes,
our_key_data->x, our_key_data->len,
&our_key ) );
Key agreement test functions
2018-09-18 12:11:34 +02:00
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
PSA_ASSERT( psa_key_derivation_setup( &operation, alg ) );
Wrap and reindent some lines After renaming several identifiers, re-wrap and re-indent some lines to make the code prettier.
2019-05-16 17:53:40 +02:00
PSA_ASSERT( psa_key_derivation_key_agreement(
&operation,
PSA_KEY_DERIVATION_INPUT_SECRET, our_key,
peer_key_data->x, peer_key_data->len ) );
Test multipart key agreement with ECDH+HKDF Basic coverage with one algorithm only and a restricted choice of output lengths.
2019-04-11 22:13:20 +02:00
if( PSA_ALG_IS_HKDF( PSA_ALG_KEY_AGREEMENT_GET_KDF( alg ) ) )
{
/* The test data is for info="" */
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
PSA_ASSERT( psa_key_derivation_input_bytes( &operation,
Rename PSA_KDF_STEP_xxx -> PSA_KEY_DERIVATION_INPUT_xxx More consistent with the new function names.
2019-05-16 16:05:19 +02:00
PSA_KEY_DERIVATION_INPUT_INFO,
Test multipart key agreement with ECDH+HKDF Basic coverage with one algorithm only and a restricted choice of output lengths.
2019-04-11 22:13:20 +02:00
NULL, 0 ) );
}
Key agreement test functions
2018-09-18 12:11:34 +02:00
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
PSA_ASSERT( psa_key_derivation_output_bytes( &operation,
Wrap and reindent some lines After renaming several identifiers, re-wrap and re-indent some lines to make the code prettier.
2019-05-16 17:53:40 +02:00
actual_output,
expected_output1->len ) );
Use ASSERT_COMPARE in preference to memcmp in PSA tests
2018-12-18 00:40:50 +01:00
ASSERT_COMPARE( actual_output, expected_output1->len,
expected_output1->x, expected_output1->len );
Add multipart key agreement tests Add test cases that do key agreement with raw selection in pieces, to validate that selection works even when the application doesn't read everything in one chunk.
2018-10-25 22:37:15 +02:00
if( expected_output2->len != 0 )
{
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
PSA_ASSERT( psa_key_derivation_output_bytes( &operation,
Wrap and reindent some lines After renaming several identifiers, re-wrap and re-indent some lines to make the code prettier.
2019-05-16 17:53:40 +02:00
actual_output,
expected_output2->len ) );
Use ASSERT_COMPARE in preference to memcmp in PSA tests
2018-12-18 00:40:50 +01:00
ASSERT_COMPARE( actual_output, expected_output2->len,
expected_output2->x, expected_output2->len );
Add multipart key agreement tests Add test cases that do key agreement with raw selection in pieces, to validate that selection works even when the application doesn't read everything in one chunk.
2018-10-25 22:37:15 +02:00
}
Key agreement test functions
2018-09-18 12:11:34 +02:00
exit:
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
psa_key_derivation_abort( &operation );
Key agreement test functions
2018-09-18 12:11:34 +02:00
psa_destroy_key( our_key );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
Key agreement test functions
2018-09-18 12:11:34 +02:00
mbedtls_free( actual_output );
}
/* END_CASE */
Implement psa_generate_random
2018-06-19 22:00:52 +02:00
/* BEGIN_CASE */
test of generate_random: focus on testing the output buffer size In the test generate_random, focus on testing that psa_generate_random is writing all the bytes of the output buffer and no more. Add a check that it is writing to each byte of the output buffer. Do not try to look for repeating output as the structure of a unit test isn't likely to catch that sort of problem anyway.
2018-06-21 10:22:13 +02:00
void generate_random( int bytes_arg )
Implement psa_generate_random
2018-06-19 22:00:52 +02:00
{
test of generate_random: focus on testing the output buffer size In the test generate_random, focus on testing that psa_generate_random is writing all the bytes of the output buffer and no more. Add a check that it is writing to each byte of the output buffer. Do not try to look for repeating output as the structure of a unit test isn't likely to catch that sort of problem anyway.
2018-06-21 10:22:13 +02:00
size_t bytes = bytes_arg;
const unsigned char trail[] = "don't overwrite me";
Use ASSERT_ALLOC instead of mbedtls_calloc in PSA tests This commit resolves a bug whereby some test cases failed on systems where mbedtls_calloc returns NULL when the size of 0, because the test case asserted `pointer != NULL` regardless of the size.
2018-09-27 13:54:18 +02:00
unsigned char *output = NULL;
unsigned char *changed = NULL;
test of generate_random: focus on testing the output buffer size In the test generate_random, focus on testing that psa_generate_random is writing all the bytes of the output buffer and no more. Add a check that it is writing to each byte of the output buffer. Do not try to look for repeating output as the structure of a unit test isn't likely to catch that sort of problem anyway.
2018-06-21 10:22:13 +02:00
size_t i;
unsigned run;
Implement psa_generate_random
2018-06-19 22:00:52 +02:00
Use ASSERT_ALLOC instead of mbedtls_calloc in PSA tests This commit resolves a bug whereby some test cases failed on systems where mbedtls_calloc returns NULL when the size of 0, because the test case asserted `pointer != NULL` regardless of the size.
2018-09-27 13:54:18 +02:00
ASSERT_ALLOC( output, bytes + sizeof( trail ) );
ASSERT_ALLOC( changed, bytes );
test of generate_random: focus on testing the output buffer size In the test generate_random, focus on testing that psa_generate_random is writing all the bytes of the output buffer and no more. Add a check that it is writing to each byte of the output buffer. Do not try to look for repeating output as the structure of a unit test isn't likely to catch that sort of problem anyway.
2018-06-21 10:22:13 +02:00
memcpy( output + bytes, trail, sizeof( trail ) );
Implement psa_generate_random
2018-06-19 22:00:52 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
Implement psa_generate_random
2018-06-19 22:00:52 +02:00
test of generate_random: focus on testing the output buffer size In the test generate_random, focus on testing that psa_generate_random is writing all the bytes of the output buffer and no more. Add a check that it is writing to each byte of the output buffer. Do not try to look for repeating output as the structure of a unit test isn't likely to catch that sort of problem anyway.
2018-06-21 10:22:13 +02:00
/* Run several times, to ensure that every output byte will be
* nonzero at least once with overwhelming probability
* (2^(-8*number_of_runs)). */
for( run = 0; run < 10; run++ )
{
Skip calling memset when the size is 0 memset(NULL, c, 0) has undefined behavior, so don't do it. clang-asan complains.
2018-09-26 18:19:24 +02:00
if( bytes != 0 )
memset( output, 0, bytes );
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_generate_random( output, bytes ) );
Implement psa_generate_random
2018-06-19 22:00:52 +02:00
test of generate_random: focus on testing the output buffer size In the test generate_random, focus on testing that psa_generate_random is writing all the bytes of the output buffer and no more. Add a check that it is writing to each byte of the output buffer. Do not try to look for repeating output as the structure of a unit test isn't likely to catch that sort of problem anyway.
2018-06-21 10:22:13 +02:00
/* Check that no more than bytes have been overwritten */
Use ASSERT_COMPARE in preference to memcmp in PSA tests
2018-12-18 00:40:50 +01:00
ASSERT_COMPARE( output + bytes, sizeof( trail ),
trail, sizeof( trail ) );
Implement psa_generate_random
2018-06-19 22:00:52 +02:00
test of generate_random: focus on testing the output buffer size In the test generate_random, focus on testing that psa_generate_random is writing all the bytes of the output buffer and no more. Add a check that it is writing to each byte of the output buffer. Do not try to look for repeating output as the structure of a unit test isn't likely to catch that sort of problem anyway.
2018-06-21 10:22:13 +02:00
for( i = 0; i < bytes; i++ )
{
if( output[i] != 0 )
++changed[i];
}
}
Implement psa_generate_random
2018-06-19 22:00:52 +02:00
test of generate_random: focus on testing the output buffer size In the test generate_random, focus on testing that psa_generate_random is writing all the bytes of the output buffer and no more. Add a check that it is writing to each byte of the output buffer. Do not try to look for repeating output as the structure of a unit test isn't likely to catch that sort of problem anyway.
2018-06-21 10:22:13 +02:00
/* Check that every byte was changed to nonzero at least once. This
* validates that psa_generate_random is overwriting every byte of
* the output buffer. */
for( i = 0; i < bytes; i++ )
Implement psa_generate_random
2018-06-19 22:00:52 +02:00
{
test of generate_random: focus on testing the output buffer size In the test generate_random, focus on testing that psa_generate_random is writing all the bytes of the output buffer and no more. Add a check that it is writing to each byte of the output buffer. Do not try to look for repeating output as the structure of a unit test isn't likely to catch that sort of problem anyway.
2018-06-21 10:22:13 +02:00
TEST_ASSERT( changed[i] != 0 );
Implement psa_generate_random
2018-06-19 22:00:52 +02:00
}
exit:
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
test of generate_random: focus on testing the output buffer size In the test generate_random, focus on testing that psa_generate_random is writing all the bytes of the output buffer and no more. Add a check that it is writing to each byte of the output buffer. Do not try to look for repeating output as the structure of a unit test isn't likely to catch that sort of problem anyway.
2018-06-21 10:22:13 +02:00
mbedtls_free( output );
mbedtls_free( changed );
Implement psa_generate_random
2018-06-19 22:00:52 +02:00
}
/* END_CASE */
Implement psa_generate_key: AES, DES, RSA, ECP In the test cases, try exporting the generated key and perform sanity checks on it.
2018-06-20 00:20:32 +02:00
/* BEGIN_CASE */
void generate_key( int type_arg,
int bits_arg,
int usage_arg,
int alg_arg,
int expected_status_arg )
{
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t handle = 0;
Implement psa_generate_key: AES, DES, RSA, ECP In the test cases, try exporting the generated key and perform sanity checks on it.
2018-06-20 00:20:32 +02:00
psa_key_type_t type = type_arg;
psa_key_usage_t usage = usage_arg;
size_t bits = bits_arg;
psa_algorithm_t alg = alg_arg;
psa_status_t expected_status = expected_status_arg;
Implement atomic-creation psa_{generate,generator_import}_key Implement the new, attribute-based psa_generate_key and psa_generator_import_key.
2019-04-18 12:53:30 +02:00
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
psa_key_attributes_t got_attributes = PSA_KEY_ATTRIBUTES_INIT;
Implement psa_generate_key: AES, DES, RSA, ECP In the test cases, try exporting the generated key and perform sanity checks on it.
2018-06-20 00:20:32 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init( ) );
Implement psa_generate_key: AES, DES, RSA, ECP In the test cases, try exporting the generated key and perform sanity checks on it.
2018-06-20 00:20:32 +02:00
Implement atomic-creation psa_{generate,generator_import}_key Implement the new, attribute-based psa_generate_key and psa_generator_import_key.
2019-04-18 12:53:30 +02:00
psa_set_key_usage_flags( &attributes, usage );
psa_set_key_algorithm( &attributes, alg );
psa_set_key_type( &attributes, type );
Set the key size as an attribute Instead of passing a separate parameter for the key size to psa_generate_key and psa_generator_import_key, set it through the attributes, like the key type and other metadata.
2019-04-26 13:49:28 +02:00
psa_set_key_bits( &attributes, bits );
Implement psa_generate_key: AES, DES, RSA, ECP In the test cases, try exporting the generated key and perform sanity checks on it.
2018-06-20 00:20:32 +02:00
/* Generate a key */
Rename psa_generate_random_key back to psa_generate_key generate_key is a more classical name. The longer name was only introduced to avoid confusion with getting a key from a generator, which is key derivation, but we no longer use the generator terminology so this reason no longer applies. perl -i -pe 's/psa_generate_random_key/psa_generate_key/g' $(git ls-files)
2019-05-16 19:42:05 +02:00
TEST_EQUAL( psa_generate_key( &attributes, &handle ), expected_status );
Remove `extra` parameter from psa_generate_key Read extra data from the domain parameters in the attribute structure instead of taking an argument on the function call. Implement this for RSA key generation, where the public exponent can be set as a domain parameter. Add tests that generate RSA keys with various public exponents.
2019-04-26 17:34:02 +02:00
if( expected_status != PSA_SUCCESS )
Implement atomic-creation psa_{generate,generator_import}_key Implement the new, attribute-based psa_generate_key and psa_generator_import_key.
2019-04-18 12:53:30 +02:00
goto exit;
Implement psa_generate_key: AES, DES, RSA, ECP In the test cases, try exporting the generated key and perform sanity checks on it.
2018-06-20 00:20:32 +02:00
/* Test the key information */
Implement psa_get_key_attributes Implement attribute querying. Test attribute getters and setters. Use psa_get_key_attributes instead of the deprecated functions psa_get_key_policy or psa_get_key_information in most tests.
2019-04-18 21:44:46 +02:00
PSA_ASSERT( psa_get_key_attributes( handle, &got_attributes ) );
TEST_EQUAL( psa_get_key_type( &got_attributes ), type );
TEST_EQUAL( psa_get_key_bits( &got_attributes ), bits );
Implement psa_generate_key: AES, DES, RSA, ECP In the test cases, try exporting the generated key and perform sanity checks on it.
2018-06-20 00:20:32 +02:00
generate_key tests: exercise the key After generating a key, perform a smoke test: run one operation with it and check that the operation has the expected status.
2018-06-20 18:16:48 +02:00
/* Do something with the key according to its type and permitted usage. */
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
if( ! exercise_key( handle, usage, alg ) )
Factor duplicated code into exercise_key Also fail the test if the test code lacks a way to exercise the key.
2018-07-01 22:31:34 +02:00
goto exit;
Implement psa_generate_key: AES, DES, RSA, ECP In the test cases, try exporting the generated key and perform sanity checks on it.
2018-06-20 00:20:32 +02:00
exit:
Call psa_reset_key_attributes after psa_get_key_attributes After calling psa_get_key_attributes(), call psa_reset_key_attributes() if the key may have domain parameters, because that's the way to free the domain parameter substructure in the attribute structure. Keep not calling reset() in some places where the key can only be a symmetric key which doesn't have domain parameters.
2019-04-26 16:03:33 +02:00
psa_reset_key_attributes( &got_attributes );
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_destroy_key( handle );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
Implement psa_generate_key: AES, DES, RSA, ECP In the test cases, try exporting the generated key and perform sanity checks on it.
2018-06-20 00:20:32 +02:00
}
/* END_CASE */
Ensure the module is initialized in psa_generate_random
2018-09-06 15:24:41 +02:00
Remove `extra` parameter from psa_generate_key Read extra data from the domain parameters in the attribute structure instead of taking an argument on the function call. Implement this for RSA key generation, where the public exponent can be set as a domain parameter. Add tests that generate RSA keys with various public exponents.
2019-04-26 17:34:02 +02:00
/* BEGIN_CASE depends_on:MBEDTLS_RSA_C:MBEDTLS_GENPRIME:MBEDTLS_PKCS1_V15 */
void generate_key_rsa( int bits_arg,
data_t *e_arg,
int expected_status_arg )
{
psa_key_handle_t handle = 0;
Rename *KEYPAIR* to *KEY_PAIR* Be consistent with PUBLIC_KEY. perl -i -pe 's/KEYPAIR/KEY_PAIR/g' $(git ls-files)
2019-05-16 19:39:54 +02:00
psa_key_type_t type = PSA_KEY_TYPE_RSA_KEY_PAIR;
Remove `extra` parameter from psa_generate_key Read extra data from the domain parameters in the attribute structure instead of taking an argument on the function call. Implement this for RSA key generation, where the public exponent can be set as a domain parameter. Add tests that generate RSA keys with various public exponents.
2019-04-26 17:34:02 +02:00
size_t bits = bits_arg;
psa_key_usage_t usage = PSA_KEY_USAGE_ENCRYPT | PSA_KEY_USAGE_DECRYPT;
psa_algorithm_t alg = PSA_ALG_RSA_PKCS1V15_SIGN_RAW;
psa_status_t expected_status = expected_status_arg;
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
uint8_t *exported = NULL;
size_t exported_size =
PSA_KEY_EXPORT_MAX_SIZE( PSA_KEY_TYPE_RSA_PUBLIC_KEY, bits );
size_t exported_length = SIZE_MAX;
uint8_t *e_read_buffer = NULL;
int is_default_public_exponent = 0;
Add buffer size macro for psa_get_key_domain_parameters
2019-04-28 11:44:17 +02:00
size_t e_read_size = PSA_KEY_DOMAIN_PARAMETERS_SIZE( type, bits );
Remove `extra` parameter from psa_generate_key Read extra data from the domain parameters in the attribute structure instead of taking an argument on the function call. Implement this for RSA key generation, where the public exponent can be set as a domain parameter. Add tests that generate RSA keys with various public exponents.
2019-04-26 17:34:02 +02:00
size_t e_read_length = SIZE_MAX;
if( e_arg->len == 0 ||
( e_arg->len == 3 &&
e_arg->x[0] == 1 && e_arg->x[1] == 0 && e_arg->x[2] == 1 ) )
{
is_default_public_exponent = 1;
e_read_size = 0;
}
ASSERT_ALLOC( e_read_buffer, e_read_size );
ASSERT_ALLOC( exported, exported_size );
PSA_ASSERT( psa_crypto_init( ) );
psa_set_key_usage_flags( &attributes, usage );
psa_set_key_algorithm( &attributes, alg );
PSA_ASSERT( psa_set_key_domain_parameters( &attributes, type,
e_arg->x, e_arg->len ) );
psa_set_key_bits( &attributes, bits );
/* Generate a key */
Rename psa_generate_random_key back to psa_generate_key generate_key is a more classical name. The longer name was only introduced to avoid confusion with getting a key from a generator, which is key derivation, but we no longer use the generator terminology so this reason no longer applies. perl -i -pe 's/psa_generate_random_key/psa_generate_key/g' $(git ls-files)
2019-05-16 19:42:05 +02:00
TEST_EQUAL( psa_generate_key( &attributes, &handle ), expected_status );
Remove `extra` parameter from psa_generate_key Read extra data from the domain parameters in the attribute structure instead of taking an argument on the function call. Implement this for RSA key generation, where the public exponent can be set as a domain parameter. Add tests that generate RSA keys with various public exponents.
2019-04-26 17:34:02 +02:00
if( expected_status != PSA_SUCCESS )
goto exit;
/* Test the key information */
PSA_ASSERT( psa_get_key_attributes( handle, &attributes ) );
TEST_EQUAL( psa_get_key_type( &attributes ), type );
TEST_EQUAL( psa_get_key_bits( &attributes ), bits );
PSA_ASSERT( psa_get_key_domain_parameters( &attributes,
e_read_buffer, e_read_size,
&e_read_length ) );
if( is_default_public_exponent )
TEST_EQUAL( e_read_length, 0 );
else
ASSERT_COMPARE( e_read_buffer, e_read_length, e_arg->x, e_arg->len );
/* Do something with the key according to its type and permitted usage. */
if( ! exercise_key( handle, usage, alg ) )
goto exit;
/* Export the key and check the public exponent. */
PSA_ASSERT( psa_export_public_key( handle,
exported, exported_size,
&exported_length ) );
{
uint8_t *p = exported;
uint8_t *end = exported + exported_length;
size_t len;
/* RSAPublicKey ::= SEQUENCE {
* modulus INTEGER, -- n
* publicExponent INTEGER } -- e
*/
TEST_EQUAL( 0, mbedtls_asn1_get_tag( &p, end, &len,
Wrap and reindent some lines After renaming several identifiers, re-wrap and re-indent some lines to make the code prettier.
2019-05-16 17:53:40 +02:00
MBEDTLS_ASN1_SEQUENCE |
MBEDTLS_ASN1_CONSTRUCTED ) );
Remove `extra` parameter from psa_generate_key Read extra data from the domain parameters in the attribute structure instead of taking an argument on the function call. Implement this for RSA key generation, where the public exponent can be set as a domain parameter. Add tests that generate RSA keys with various public exponents.
2019-04-26 17:34:02 +02:00
TEST_ASSERT( asn1_skip_integer( &p, end, bits, bits, 1 ) );
TEST_EQUAL( 0, mbedtls_asn1_get_tag( &p, end, &len,
MBEDTLS_ASN1_INTEGER ) );
if( len >= 1 && p[0] == 0 )
{
++p;
--len;
}
if( e_arg->len == 0 )
{
TEST_EQUAL( len, 3 );
TEST_EQUAL( p[0], 1 );
TEST_EQUAL( p[1], 0 );
TEST_EQUAL( p[2], 1 );
}
else
ASSERT_COMPARE( p, len, e_arg->x, e_arg->len );
}
exit:
psa_reset_key_attributes( &attributes );
psa_destroy_key( handle );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE( );
Remove `extra` parameter from psa_generate_key Read extra data from the domain parameters in the attribute structure instead of taking an argument on the function call. Implement this for RSA key generation, where the public exponent can be set as a domain parameter. Add tests that generate RSA keys with various public exponents.
2019-04-26 17:34:02 +02:00
mbedtls_free( e_read_buffer );
mbedtls_free( exported );
}
/* END_CASE */
psa: Implement persistent keys Allow use of persistent keys, including configuring them, importing and exporting them, and destroying them. When getting a slot using psa_get_key_slot, there are 3 scenarios that can occur if the keys lifetime is persistent: 1. Key type is PSA_KEY_TYPE_NONE, no persistent storage entry: - The key slot is treated as a standard empty key slot 2. Key type is PSA_KEY_TYPE_NONE, persistent storage entry exists: - Attempt to load the key from persistent storage 3. Key type is not PSA_KEY_TYPE_NONE: - As checking persistent storage on every use of the key could be expensive, the persistent key is assumed to be saved in persistent storage, the in-memory key is continued to be used.
2018-06-18 18:27:26 +02:00
/* BEGIN_CASE depends_on:MBEDTLS_PSA_CRYPTO_STORAGE_C */
Update persistent_key_load_key_from_storage to use attributes Update persistent_key_load_key_from_storage to the new attribute-based key creation interface. I tweaked the code a little to make it simpler and more robust without changing the core logic.
2019-04-19 14:06:53 +02:00
void persistent_key_load_key_from_storage( data_t *data,
int type_arg, int bits_arg,
int usage_flags_arg, int alg_arg,
int generation_method )
psa: Implement persistent keys Allow use of persistent keys, including configuring them, importing and exporting them, and destroying them. When getting a slot using psa_get_key_slot, there are 3 scenarios that can occur if the keys lifetime is persistent: 1. Key type is PSA_KEY_TYPE_NONE, no persistent storage entry: - The key slot is treated as a standard empty key slot 2. Key type is PSA_KEY_TYPE_NONE, persistent storage entry exists: - Attempt to load the key from persistent storage 3. Key type is not PSA_KEY_TYPE_NONE: - As checking persistent storage on every use of the key could be expensive, the persistent key is assumed to be saved in persistent storage, the in-memory key is continued to be used.
2018-06-18 18:27:26 +02:00
{
Update persistent_key_load_key_from_storage to use attributes Update persistent_key_load_key_from_storage to the new attribute-based key creation interface. I tweaked the code a little to make it simpler and more robust without changing the core logic.
2019-04-19 14:06:53 +02:00
psa_key_id_t key_id = 1;
psa_key_attributes_t attributes = PSA_KEY_ATTRIBUTES_INIT;
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_key_handle_t handle = 0;
Update persistent_key_load_key_from_storage to use attributes Update persistent_key_load_key_from_storage to the new attribute-based key creation interface. I tweaked the code a little to make it simpler and more robust without changing the core logic.
2019-04-19 14:06:53 +02:00
psa_key_handle_t base_key = 0;
psa_key_type_t type = type_arg;
size_t bits = bits_arg;
psa_key_usage_t usage_flags = usage_flags_arg;
psa_algorithm_t alg = alg_arg;
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
psa_key_derivation_operation_t operation = PSA_KEY_DERIVATION_OPERATION_INIT;
psa: Implement persistent keys Allow use of persistent keys, including configuring them, importing and exporting them, and destroying them. When getting a slot using psa_get_key_slot, there are 3 scenarios that can occur if the keys lifetime is persistent: 1. Key type is PSA_KEY_TYPE_NONE, no persistent storage entry: - The key slot is treated as a standard empty key slot 2. Key type is PSA_KEY_TYPE_NONE, persistent storage entry exists: - Attempt to load the key from persistent storage 3. Key type is not PSA_KEY_TYPE_NONE: - As checking persistent storage on every use of the key could be expensive, the persistent key is assumed to be saved in persistent storage, the in-memory key is continued to be used.
2018-06-18 18:27:26 +02:00
unsigned char *first_export = NULL;
unsigned char *second_export = NULL;
size_t export_size = PSA_KEY_EXPORT_MAX_SIZE( type, bits );
size_t first_exported_length;
size_t second_exported_length;
Update persistent_key_load_key_from_storage to use attributes Update persistent_key_load_key_from_storage to the new attribute-based key creation interface. I tweaked the code a little to make it simpler and more robust without changing the core logic.
2019-04-19 14:06:53 +02:00
if( usage_flags & PSA_KEY_USAGE_EXPORT )
{
ASSERT_ALLOC( first_export, export_size );
ASSERT_ALLOC( second_export, export_size );
}
psa: Implement persistent keys Allow use of persistent keys, including configuring them, importing and exporting them, and destroying them. When getting a slot using psa_get_key_slot, there are 3 scenarios that can occur if the keys lifetime is persistent: 1. Key type is PSA_KEY_TYPE_NONE, no persistent storage entry: - The key slot is treated as a standard empty key slot 2. Key type is PSA_KEY_TYPE_NONE, persistent storage entry exists: - Attempt to load the key from persistent storage 3. Key type is not PSA_KEY_TYPE_NONE: - As checking persistent storage on every use of the key could be expensive, the persistent key is assumed to be saved in persistent storage, the in-memory key is continued to be used.
2018-06-18 18:27:26 +02:00
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init() );
psa: Implement persistent keys Allow use of persistent keys, including configuring them, importing and exporting them, and destroying them. When getting a slot using psa_get_key_slot, there are 3 scenarios that can occur if the keys lifetime is persistent: 1. Key type is PSA_KEY_TYPE_NONE, no persistent storage entry: - The key slot is treated as a standard empty key slot 2. Key type is PSA_KEY_TYPE_NONE, persistent storage entry exists: - Attempt to load the key from persistent storage 3. Key type is not PSA_KEY_TYPE_NONE: - As checking persistent storage on every use of the key could be expensive, the persistent key is assumed to be saved in persistent storage, the in-memory key is continued to be used.
2018-06-18 18:27:26 +02:00
Replace psa_make_key_persistent by id/lifetime setters in tests Remove all internal uses of psa_make_key_persistent.
2019-05-15 16:12:22 +02:00
psa_set_key_id( &attributes, key_id );
Update persistent_key_load_key_from_storage to use attributes Update persistent_key_load_key_from_storage to the new attribute-based key creation interface. I tweaked the code a little to make it simpler and more robust without changing the core logic.
2019-04-19 14:06:53 +02:00
psa_set_key_usage_flags( &attributes, usage_flags );
psa_set_key_algorithm( &attributes, alg );
psa_set_key_type( &attributes, type );
Set the key size as an attribute Instead of passing a separate parameter for the key size to psa_generate_key and psa_generator_import_key, set it through the attributes, like the key type and other metadata.
2019-04-26 13:49:28 +02:00
psa_set_key_bits( &attributes, bits );
psa: Implement persistent keys Allow use of persistent keys, including configuring them, importing and exporting them, and destroying them. When getting a slot using psa_get_key_slot, there are 3 scenarios that can occur if the keys lifetime is persistent: 1. Key type is PSA_KEY_TYPE_NONE, no persistent storage entry: - The key slot is treated as a standard empty key slot 2. Key type is PSA_KEY_TYPE_NONE, persistent storage entry exists: - Attempt to load the key from persistent storage 3. Key type is not PSA_KEY_TYPE_NONE: - As checking persistent storage on every use of the key could be expensive, the persistent key is assumed to be saved in persistent storage, the in-memory key is continued to be used.
2018-06-18 18:27:26 +02:00
psa: Extend psa_generate_key to support persistent lifetimes
2018-11-07 17:05:30 +01:00
switch( generation_method )
{
case IMPORT_KEY:
/* Import the key */
Split long lines after psa_import_key refactoring
2019-05-15 20:22:09 +02:00
PSA_ASSERT( psa_import_key( &attributes, data->x, data->len,
&handle ) );
psa: Extend psa_generate_key to support persistent lifetimes
2018-11-07 17:05:30 +01:00
break;
psa: Implement persistent keys Allow use of persistent keys, including configuring them, importing and exporting them, and destroying them. When getting a slot using psa_get_key_slot, there are 3 scenarios that can occur if the keys lifetime is persistent: 1. Key type is PSA_KEY_TYPE_NONE, no persistent storage entry: - The key slot is treated as a standard empty key slot 2. Key type is PSA_KEY_TYPE_NONE, persistent storage entry exists: - Attempt to load the key from persistent storage 3. Key type is not PSA_KEY_TYPE_NONE: - As checking persistent storage on every use of the key could be expensive, the persistent key is assumed to be saved in persistent storage, the in-memory key is continued to be used.
2018-06-18 18:27:26 +02:00
psa: Extend psa_generate_key to support persistent lifetimes
2018-11-07 17:05:30 +01:00
case GENERATE_KEY:
/* Generate a key */
Rename psa_generate_random_key back to psa_generate_key generate_key is a more classical name. The longer name was only introduced to avoid confusion with getting a key from a generator, which is key derivation, but we no longer use the generator terminology so this reason no longer applies. perl -i -pe 's/psa_generate_random_key/psa_generate_key/g' $(git ls-files)
2019-05-16 19:42:05 +02:00
PSA_ASSERT( psa_generate_key( &attributes, &handle ) );
psa: Extend psa_generate_key to support persistent lifetimes
2018-11-07 17:05:30 +01:00
break;
case DERIVE_KEY:
Update persistent_key_load_key_from_storage to use attributes Update persistent_key_load_key_from_storage to the new attribute-based key creation interface. I tweaked the code a little to make it simpler and more robust without changing the core logic.
2019-04-19 14:06:53 +02:00
{
/* Create base key */
psa_algorithm_t derive_alg = PSA_ALG_HKDF( PSA_ALG_SHA_256 );
psa_key_attributes_t base_attributes = PSA_KEY_ATTRIBUTES_INIT;
psa_set_key_usage_flags( &base_attributes,
PSA_KEY_USAGE_DERIVE );
psa_set_key_algorithm( &base_attributes, derive_alg );
psa_set_key_type( &base_attributes, PSA_KEY_TYPE_DERIVE );
Split long lines after psa_import_key refactoring
2019-05-15 20:22:09 +02:00
PSA_ASSERT( psa_import_key( &base_attributes,
data->x, data->len,
&base_key ) );
Update persistent_key_load_key_from_storage to use attributes Update persistent_key_load_key_from_storage to the new attribute-based key creation interface. I tweaked the code a little to make it simpler and more robust without changing the core logic.
2019-04-19 14:06:53 +02:00
/* Derive a key. */
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
PSA_ASSERT( psa_key_derivation_setup( &operation, derive_alg ) );
Wrap and reindent some lines After renaming several identifiers, re-wrap and re-indent some lines to make the code prettier.
2019-05-16 17:53:40 +02:00
PSA_ASSERT( psa_key_derivation_input_key(
&operation,
PSA_KEY_DERIVATION_INPUT_SECRET, base_key ) );
Update persistent_key_load_key_from_storage to use attributes Update persistent_key_load_key_from_storage to the new attribute-based key creation interface. I tweaked the code a little to make it simpler and more robust without changing the core logic.
2019-04-19 14:06:53 +02:00
PSA_ASSERT( psa_key_derivation_input_bytes(
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
&operation, PSA_KEY_DERIVATION_INPUT_INFO,
Update persistent_key_load_key_from_storage to use attributes Update persistent_key_load_key_from_storage to the new attribute-based key creation interface. I tweaked the code a little to make it simpler and more robust without changing the core logic.
2019-04-19 14:06:53 +02:00
NULL, 0 ) );
Wrap and reindent some lines After renaming several identifiers, re-wrap and re-indent some lines to make the code prettier.
2019-05-16 17:53:40 +02:00
PSA_ASSERT( psa_key_derivation_output_key( &attributes,
&operation,
&handle ) );
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
PSA_ASSERT( psa_key_derivation_abort( &operation ) );
Update persistent_key_load_key_from_storage to use attributes Update persistent_key_load_key_from_storage to the new attribute-based key creation interface. I tweaked the code a little to make it simpler and more robust without changing the core logic.
2019-04-19 14:06:53 +02:00
PSA_ASSERT( psa_destroy_key( base_key ) );
base_key = 0;
}
Wrap and reindent some lines After renaming several identifiers, re-wrap and re-indent some lines to make the code prettier.
2019-05-16 17:53:40 +02:00
break;
psa: Extend psa_generate_key to support persistent lifetimes
2018-11-07 17:05:30 +01:00
}
Update persistent_key_load_key_from_storage to use attributes Update persistent_key_load_key_from_storage to the new attribute-based key creation interface. I tweaked the code a little to make it simpler and more robust without changing the core logic.
2019-04-19 14:06:53 +02:00
psa_reset_key_attributes( &attributes );
psa: Implement persistent keys Allow use of persistent keys, including configuring them, importing and exporting them, and destroying them. When getting a slot using psa_get_key_slot, there are 3 scenarios that can occur if the keys lifetime is persistent: 1. Key type is PSA_KEY_TYPE_NONE, no persistent storage entry: - The key slot is treated as a standard empty key slot 2. Key type is PSA_KEY_TYPE_NONE, persistent storage entry exists: - Attempt to load the key from persistent storage 3. Key type is not PSA_KEY_TYPE_NONE: - As checking persistent storage on every use of the key could be expensive, the persistent key is assumed to be saved in persistent storage, the in-memory key is continued to be used.
2018-06-18 18:27:26 +02:00
Update persistent_key_load_key_from_storage to use attributes Update persistent_key_load_key_from_storage to the new attribute-based key creation interface. I tweaked the code a little to make it simpler and more robust without changing the core logic.
2019-04-19 14:06:53 +02:00
/* Export the key if permitted by the key policy. */
if( usage_flags & PSA_KEY_USAGE_EXPORT )
{
PSA_ASSERT( psa_export_key( handle,
first_export, export_size,
&first_exported_length ) );
if( generation_method == IMPORT_KEY )
ASSERT_COMPARE( data->x, data->len,
first_export, first_exported_length );
}
psa: Implement persistent keys Allow use of persistent keys, including configuring them, importing and exporting them, and destroying them. When getting a slot using psa_get_key_slot, there are 3 scenarios that can occur if the keys lifetime is persistent: 1. Key type is PSA_KEY_TYPE_NONE, no persistent storage entry: - The key slot is treated as a standard empty key slot 2. Key type is PSA_KEY_TYPE_NONE, persistent storage entry exists: - Attempt to load the key from persistent storage 3. Key type is not PSA_KEY_TYPE_NONE: - As checking persistent storage on every use of the key could be expensive, the persistent key is assumed to be saved in persistent storage, the in-memory key is continued to be used.
2018-06-18 18:27:26 +02:00
/* Shutdown and restart */
Close or destroy keys explicitly in tests
2019-05-28 14:08:50 +02:00
PSA_ASSERT( psa_close_key( handle ) );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE();
Use PSA_ASSERT(a) in preference to TEST_ASSERT(a==PSA_SUCCESS) This commit is the result of the following command, followed by reindenting (but not wrapping lines): perl -00 -i -pe 's/^( *)TEST_ASSERT\(([^;=]*)(?: |\n *)==\s*PSA_SUCCESS\s*\);$/${1}PSA_ASSERT($2 );/gm' tests/suites/test_suite_psa_*.function
2018-12-18 00:18:46 +01:00
PSA_ASSERT( psa_crypto_init() );
psa: Implement persistent keys Allow use of persistent keys, including configuring them, importing and exporting them, and destroying them. When getting a slot using psa_get_key_slot, there are 3 scenarios that can occur if the keys lifetime is persistent: 1. Key type is PSA_KEY_TYPE_NONE, no persistent storage entry: - The key slot is treated as a standard empty key slot 2. Key type is PSA_KEY_TYPE_NONE, persistent storage entry exists: - Attempt to load the key from persistent storage 3. Key type is not PSA_KEY_TYPE_NONE: - As checking persistent storage on every use of the key could be expensive, the persistent key is assumed to be saved in persistent storage, the in-memory key is continued to be used.
2018-06-18 18:27:26 +02:00
/* Check key slot still contains key data */
Remove lifetime parameter from psa_open_key Change the scope of key identifiers to be global, rather than per lifetime. As a result, you now need to specify the lifetime of a key only when creating it.
2019-05-06 18:44:55 +02:00
PSA_ASSERT( psa_open_key( key_id, &handle ) );
Update persistent_key_load_key_from_storage to use attributes Update persistent_key_load_key_from_storage to the new attribute-based key creation interface. I tweaked the code a little to make it simpler and more robust without changing the core logic.
2019-04-19 14:06:53 +02:00
PSA_ASSERT( psa_get_key_attributes( handle, &attributes ) );
TEST_EQUAL( psa_get_key_id( &attributes ), key_id );
TEST_EQUAL( psa_get_key_lifetime( &attributes ),
PSA_KEY_LIFETIME_PERSISTENT );
TEST_EQUAL( psa_get_key_type( &attributes ), type );
TEST_EQUAL( psa_get_key_bits( &attributes ), bits );
TEST_EQUAL( psa_get_key_usage_flags( &attributes ), usage_flags );
TEST_EQUAL( psa_get_key_algorithm( &attributes ), alg );
psa: Implement persistent keys Allow use of persistent keys, including configuring them, importing and exporting them, and destroying them. When getting a slot using psa_get_key_slot, there are 3 scenarios that can occur if the keys lifetime is persistent: 1. Key type is PSA_KEY_TYPE_NONE, no persistent storage entry: - The key slot is treated as a standard empty key slot 2. Key type is PSA_KEY_TYPE_NONE, persistent storage entry exists: - Attempt to load the key from persistent storage 3. Key type is not PSA_KEY_TYPE_NONE: - As checking persistent storage on every use of the key could be expensive, the persistent key is assumed to be saved in persistent storage, the in-memory key is continued to be used.
2018-06-18 18:27:26 +02:00
Update persistent_key_load_key_from_storage to use attributes Update persistent_key_load_key_from_storage to the new attribute-based key creation interface. I tweaked the code a little to make it simpler and more robust without changing the core logic.
2019-04-19 14:06:53 +02:00
/* Export the key again if permitted by the key policy. */
if( usage_flags & PSA_KEY_USAGE_EXPORT )
psa: Extend psa_generate_key to support persistent lifetimes
2018-11-07 17:05:30 +01:00
{
Update persistent_key_load_key_from_storage to use attributes Update persistent_key_load_key_from_storage to the new attribute-based key creation interface. I tweaked the code a little to make it simpler and more robust without changing the core logic.
2019-04-19 14:06:53 +02:00
PSA_ASSERT( psa_export_key( handle,
second_export, export_size,
&second_exported_length ) );
psa: Extend psa_generate_key to support persistent lifetimes
2018-11-07 17:05:30 +01:00
ASSERT_COMPARE( first_export, first_exported_length,
second_export, second_exported_length );
}
/* Do something with the key according to its type and permitted usage. */
Update persistent_key_load_key_from_storage to use attributes Update persistent_key_load_key_from_storage to the new attribute-based key creation interface. I tweaked the code a little to make it simpler and more robust without changing the core logic.
2019-04-19 14:06:53 +02:00
if( ! exercise_key( handle, usage_flags, alg ) )
psa: Extend psa_generate_key to support persistent lifetimes
2018-11-07 17:05:30 +01:00
goto exit;
psa: Implement persistent keys Allow use of persistent keys, including configuring them, importing and exporting them, and destroying them. When getting a slot using psa_get_key_slot, there are 3 scenarios that can occur if the keys lifetime is persistent: 1. Key type is PSA_KEY_TYPE_NONE, no persistent storage entry: - The key slot is treated as a standard empty key slot 2. Key type is PSA_KEY_TYPE_NONE, persistent storage entry exists: - Attempt to load the key from persistent storage 3. Key type is not PSA_KEY_TYPE_NONE: - As checking persistent storage on every use of the key could be expensive, the persistent key is assumed to be saved in persistent storage, the in-memory key is continued to be used.
2018-06-18 18:27:26 +02:00
exit:
Call psa_reset_key_attributes after psa_get_key_attributes After calling psa_get_key_attributes(), call psa_reset_key_attributes() if the key may have domain parameters, because that's the way to free the domain parameter substructure in the attribute structure. Keep not calling reset() in some places where the key can only be a symmetric key which doesn't have domain parameters.
2019-04-26 16:03:33 +02:00
psa_reset_key_attributes( &attributes );
psa: Implement persistent keys Allow use of persistent keys, including configuring them, importing and exporting them, and destroying them. When getting a slot using psa_get_key_slot, there are 3 scenarios that can occur if the keys lifetime is persistent: 1. Key type is PSA_KEY_TYPE_NONE, no persistent storage entry: - The key slot is treated as a standard empty key slot 2. Key type is PSA_KEY_TYPE_NONE, persistent storage entry exists: - Attempt to load the key from persistent storage 3. Key type is not PSA_KEY_TYPE_NONE: - As checking persistent storage on every use of the key could be expensive, the persistent key is assumed to be saved in persistent storage, the in-memory key is continued to be used.
2018-06-18 18:27:26 +02:00
mbedtls_free( first_export );
mbedtls_free( second_export );
Rename "generator" to "operation" Generators are now key derivation operations. Keep "random generator" intact.
2019-05-16 17:31:03 +02:00
psa_key_derivation_abort( &operation );
Update persistent_key_load_key_from_storage to use attributes Update persistent_key_load_key_from_storage to the new attribute-based key creation interface. I tweaked the code a little to make it simpler and more robust without changing the core logic.
2019-04-19 14:06:53 +02:00
psa_destroy_key( base_key );
if( handle == 0 )
{
/* In case there was a test failure after creating the persistent key
* but while it was not open, try to re-open the persistent key
* to delete it. */
Remove lifetime parameter from psa_open_key Change the scope of key identifiers to be global, rather than per lifetime. As a result, you now need to specify the lifetime of a key only when creating it.
2019-05-06 18:44:55 +02:00
psa_open_key( key_id, &handle );
Update persistent_key_load_key_from_storage to use attributes Update persistent_key_load_key_from_storage to the new attribute-based key creation interface. I tweaked the code a little to make it simpler and more robust without changing the core logic.
2019-04-19 14:06:53 +02:00
}
Convert the PSA crypto cryptography tests to the new handle API Switch from the direct use of slot numbers to handles allocated by psa_allocate_key. This commit does not affect persistent key tests except for the one test function in test_suite_psa_crypto that uses persistent keys (persistent_key_load_key_from_storage). The general principle for each function is: * Change `psa_key_slot_t slot` to `psa_key_handle_t handle`. * Call psa_allocate_key() before setting the policy of the slot, or before creating key material in functions that don't set a policy. * Some PSA_ERROR_EMPTY_SLOT errors become PSA_ERROR_INVALID_HANDLE because there is now a distinction between not having a valid handle, and having a valid handle to a slot that doesn't contain key material. * In tests that use symmetric keys, calculate the max_bits parameters of psa_allocate_key() from the key data size. In tests where the key may be asymmetric, call an auxiliary macro KEY_BITS_FROM_DATA which returns an overapproximation. There's no good way to find a good value for max_bits with the API, I think the API should be tweaked.
2018-12-03 15:36:32 +01:00
psa_destroy_key( handle );
Replace all calls to mbedtls_psa_crypto_free by PSA_DONE Replace all calls to mbedtls_psa_crypto_free in tests by PSA_DONE. This is correct for most tests, because most tests close open keys. A few tests now fail; these tests need to be reviewed and switched back to mbedtls_psa_crypto_free if they genuinely expected to end with some slots still in use.
2019-05-28 15:10:21 +02:00
PSA_DONE();
psa: Implement persistent keys Allow use of persistent keys, including configuring them, importing and exporting them, and destroying them. When getting a slot using psa_get_key_slot, there are 3 scenarios that can occur if the keys lifetime is persistent: 1. Key type is PSA_KEY_TYPE_NONE, no persistent storage entry: - The key slot is treated as a standard empty key slot 2. Key type is PSA_KEY_TYPE_NONE, persistent storage entry exists: - Attempt to load the key from persistent storage 3. Key type is not PSA_KEY_TYPE_NONE: - As checking persistent storage on every use of the key could be expensive, the persistent key is assumed to be saved in persistent storage, the in-memory key is continued to be used.
2018-06-18 18:27:26 +02:00
}
/* END_CASE */
Reference in New Issue Copy Permalink