All current usages have this parameter set to 0 (this means the tests are unchanged).
Remove the GENERIC_ERROR return behaviour, in favour of returning the actual status.
Signed-off-by: Ryan Everett <ryan.everett@arm.com>
All current usages have this parameter set to 0 (meaning the behaviour
of these tests hasn't changed). We also now return the actual error code, not GENERIC_ERROR
Signed-off-by: Ryan Everett <ryan.everett@arm.com>
All current usages have this parameter set to 0 (in this case the behaviour of
the test is unchanged)
Signed-off-by: Ryan Everett <ryan.everett@arm.com>
This will allow us to use this smoke test to ensure that key slot content reads are
only performed when we are registered to read a full slot. We will destroy the key
on another thread while the key is being exercised, and fail the test if an unexpected
error code is hit. Future commits will incrementally implement this new parameter.
All current usages of this function have this parameter set to 0, in which case
the new behaviour must be the same as the old behaviour
Signed-off-by: Ryan Everett <ryan.everett@arm.com>
Test that a PK key and a PSA key are consistent, i.e. that they have the
same type (or are a key pair and the corresponding public key) and that
they have the same public key.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
This fixes the ability to exercise keys in configurations where MD5 is
supported for direct use, but not inside some accelerated algorithms. This
is the case in `all.sh test_psa_crypto_config_accel_ecc_ecp_light_only` and
some other accelerated-ECC components of `all.sh`, where the driver is built
without MD5 support but built-in MD5 remains enabled.
This is only a hack, not a theoretically correct fix, but a correct fix is
out of scope of my current work.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Parsing a key and importing it into PSA may result in a policy that
specifies an algorithm that is not included in the build. This happens if
the key type is supported, but not the algorithm, e.g. in a build with
MBEDTLS_ECP_C but not MBEDTLS_ECDSA_C.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
These functions are thread safe, but using them from within multiple
threads at the same time may not have the intended effect, given order
cannot be guaranteed. Also, standardise header comment formatting.
Signed-off-by: Paul Elliott <paul.elliott@arm.com>
The psa_test_wrappers.h inclusion was breaking the examples in programs/
on functions with poisoning added
Signed-off-by: Ryan Everett <ryan.everett@arm.com>
Improve the definition of mbedtls_test_thread_t to assist adding future
threading implementations, when they happen.
Signed-off-by: Paul Elliott <paul.elliott@arm.com>
This option causes nested calls to PSA functions, so is not compatible
with memory poisoning as it currently stands.
Signed-off-by: David Horstmann <david.horstmann@arm.com>
`gcc-14` added a new `-Wcalloc-transposed-args` warning recently. It
detected minor infelicity in `calloc()` API usage in `mbedtls`:
In file included from /build/mbedtls/tests/include/test/ssl_helpers.h:19,
from /build/mbedtls/tests/src/test_helpers/ssl_helpers.c:11:
/build/mbedtls/tests/src/test_helpers/ssl_helpers.c: In function 'mbedtls_test_init_handshake_options':
/build/mbedtls/tests/include/test/macros.h:128:46:
error: 'calloc' sizes specified with 'sizeof' in the earlier argument
and not in the later argument [-Werror=calloc-transposed-args]
128 | (pointer) = mbedtls_calloc(sizeof(*(pointer)), \
| ^
Signed-off-by: Sergei Trofimovich <slyich@gmail.com>
Add ticket write/parse test functions as defined
by mbedtls_ssl_ticket_write/parse_t. They are
intended to be used in negative testing
involving tickets.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
Remove the use of extern and instead use an accessor to get the address
of the test info mutex (defined only if MBEDTLS_TEST_MUTEX_USAGE is
defined, to hopefully stop more general usage)
Signed-off-by: Paul Elliott <paul.elliott@arm.com>
Only mbedtls_test_opaque_unwrap_key() is actually needed by other
test drivers to deal with opaque keys. mbedtls_test_opaque_wrap_key()
can be kept private to test_driver_key_management.c.
Signed-off-by: Valerio Setti <valerio.setti@nordicsemi.no>
Make sure that extra UNPOISON calls do not cause the poisoning counter
to underflow and wrap around.
Memory that is unpoisoned multiple times should remain unpoisoned.
Signed-off-by: David Horstmann <david.horstmann@arm.com>
This allows unusually-nested memory poisoning to work correctly, since
it keeps track of whether any buffers are still poisoned, rather than
just disabling poisoning at the first call to the UNPOISON() macro.
Signed-off-by: David Horstmann <david.horstmann@arm.com>
`psa_collect_statuses.py` runs `make RECORD_PSA_STATUS_COVERAGE_LOG=1`,
which builds with `RECORD_PSA_STATUS_COVERAGE_LOG`. In this mode, the build
includes wrappers for PSA functions, which conflict with the newly
introduced wrappers that are enabled whenever `MBEDTLS_TEST_HOOKS` is
enabled. In the future, the collect-statuses mechanism should use the new
generic wrapper mechanism. For the time being, keep the old wrappers and
avoid the new wrappers when doing the collect-statuses build.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Code in unit tests (`tests/suites/*.function`) and in test support
code (`tests/src/**.c`) will now go through the wrapper functions when they
call a PSA API function and `MBEDTLS_TEST_HOOKS` is enabled.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Commit files generated by `tests/scripts/generate_psa_wrappers.py`. As of
this commit, the new code is neither useful (the wrappers just call the
underlying functions) nor used (the wrapper functions are not called from
anywhere). This will change in subsequent commits.
This is a deviation from our normal practice of generating
configuration-independent platform-independent files as part of the build in
the development branch. The PSA test wrappers will be committed to the
repository for some time for two reasons:
* In the short term, we will review the generated code but not fully review
the generator script.
* The build scripts cannot yet accommodate a generated header file.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Allow memory poisoning to be enabled and disabled at runtime using a
thread-local flag. This allows poisoning to be disabled whenever a PSA
function is called but not through the test wrappers, removing false
positive use-after-poisons.
Signed-off-by: David Horstmann <david.horstmann@arm.com>
Make sure that we don't enable memory poisoning when
MBEDTLS_PSA_COPY_CALLER_BUFFERS is disabled.
Signed-off-by: David Horstmann <david.horstmann@arm.com>
Since we are applying hooks transparently to all tests, we cannot setup
and teardown test hooks in the tests. Instead we must do this in the
test wrappers which are used to pre-poison and unpoison memory.
Signed-off-by: David Horstmann <david.horstmann@arm.com>
Use the preprocessor to wrap psa_cipher_encrypt in a wrapper that
poisons the input and output buffers.
Signed-off-by: David Horstmann <david.horstmann@arm.com>
This commit
- Reverts changes previously done to psa_crypto_helpers.[c,h]
- Implements a new check for open key slots in
mbedtls_test_helper_is_psa_leaking():
- when CTR_DRBG does not use AES_C or PSA does not have an external
RNG, then we allow 1 key slot (it's the one holding the AES key)
- when the above conditions are not met, then we fallback to the
usual check for "no open key slots remaining"
Signed-off-by: Valerio Setti <valerio.setti@nordicsemi.no>
When AES_C is not defined CTR_DRBG relies on PSA to get AES-ECB. This
means that, when AES-ECB is accelerated, each random operation goes through
driver access as well. This might result in unexpectedly increased
counters for driver's access.
We add extra counters in test_driver_[cipher/key_management].c to be
more specific on which driver functions are accessed and ignore
extra accesses due to CTR_DRBG.
Signed-off-by: Valerio Setti <valerio.setti@nordicsemi.no>
This commit also adds AES_PSA_[INIT/DONE] in "psa_crypto_helpers.h". Its
scope is to call PSA_[INIT/DONE] only when AES_C is not defined (which is
when PSA is effectively required for CTR_DRBG).
Signed-off-by: Valerio Setti <valerio.setti@nordicsemi.no>
After adding a check in ssl_conf_check(), we have configured RNG via
mbedtls_ssl_conf_rng() for TLS tests in both test_suite_ssl.function
and test_suite_debug.function. As a result, rng_get() is not only
available when MBEDTLS_SSL_HANDSHAKE_WITH_CERT_ENABLED enabled.
Therefore, we remove the guard for rng_get() to make it accessible for
TLS tests which have call for mbedtls_ssl_setup().
Signed-off-by: Yanray Wang <yanray.wang@arm.com>
Asan poisons memory with an 8-byte granularity. We want to make sure that
the whole specified region is poisoned (our typical use case is a
heap-allocated object, and we want to poison the whole object, and we don't
care about the bytes after the end of the object and up to the beginning of
the next object). So align the start and end of the region to (un)poison to
an 8-byte boundary.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
While an area of memory is poisoned, reading or writing from it triggers a
sanitizer violation.
Implemented for ASan.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Conflicts:
* `include/mbedtls/build_info.h`: a new fragment to auto-enable
`MBEDTLS_CIPHER_PADDING_PKCS7` was added in
c9f4040f7f3356293e90c58d11f6567def641e08 in `development-restricted`.
In `development`, this section of the file has moved to
`include/mbedtls/config_adjust_legacy_crypto.h`.
* `library/bignum.c`: function name change in `development-restricted` vs
comment change in development. The comment change in `development` is not
really relevant, so just take the line from `development-restricted`.
Since 3.0.0, mbedtls_config.h (formerly config.h) no longer needs to include
config_psa.h or check_config.h: build_info.h takes care of that.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Besides being unnecessary, it was causing problem when build SSL test
programs, which include this header, then in turn trying to include the
internal header from library, which didn't work.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
Most functions in psa_util.h are going to end up there (except those
that can be static in one file), but I wanted to have separate commits
for file creation and moving code around, so for now the new file's
pretty empty but that will change in the next few commits.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
- deprecate legacy PSA_WANT_KEY_TYPE_xxx_KEY_PAIR
- introduce new PSA_WANT_KEY_TYPE_xxx_KEY_PAIR_yyy where
- xxx is either RSA, DH or ECC
- yyy can be USE, IMPORT, EXPORT, GENERATE, DERIVE
Signed-off-by: Valerio Setti <valerio.setti@nordicsemi.no>
The seed file must exist before running tests. Because the location is
somewhat platform- and configuration-dependent, and to be friendly to
developers who run test suites individually and aren't familiar with this
feature, rely on the test framework code rather than on test scripts to
create the seed file.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
The build option MBEDTLS_PSA_INJECT_ENTROPY requires some extra platform
functions, for historical reasons. To enable us to test this option, provide
a version of these functions for testing.
(These versions would actually work in production, but providing them in the
library in a way that doesn't break existing users might be slightly tricky,
so it's out of scope of this commit.)
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Change the type of signed integer arguments from int32_t to intmax_t.
This allows the C code to work with test function arguments with a range
larger than int32_t. A subsequent commit will change the .datax generator
to support larger types.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
The test framework stores size_t and int32_t values in the parameter store
by converting them all to int. This is ok in practice, since we assume int
covers int32_t and we don't have test data larger than 2GB. But it's
confusing and error-prone. So make the parameter store a union, which allows
size_t values not to be potentially truncated and makes the code a little
clearer.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
