Add a test that does some basic validation of the pkg-config files.
Example run:
./tests/scripts/all.sh test_cmake_as_package
<snip>
******************************************************************
* test_cmake_as_package: build: cmake 'as-package' build
* Wed Sep 11 16:17:41 UTC 2024
******************************************************************
cmake .
make
Built against Mbed TLS 3.6.0
testing package config file: mbedtls ... passed
testing package config file: mbedx509 ... passed
testing package config file: mbedcrypto ... passed
make clean
Signed-off-by: Bill Roberts <bill.roberts@arm.com>
Running test suites from another directory only works when the auto-chdir
code in host_test.function is enabled, which is platform-dependent and
configuration-dependent.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Before this commit, `make test` stopped after running the TLS tests if there
was a failure.
Have `run-test-suites.pl` take care of looking in all the directories, so
that the last line of output from `make test` is an accurate report of all
the test suites, not just the test suites from the last run of
`run-test-suites.pl`.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
In preparation of MBEDTLS_PSA_CRYPTO_CONFIG
being always on, enable MBEDTLS_PSA_CRYPTO_CONFIG
in build_aes_variations.
While enabling MBEDTLS_PSA_CRYPTO_CONFIG, disable
some PSA_WANT_ so that the configuration
remain the same for the modified component.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
In preparation of MBEDTLS_PSA_CRYPTO_CONFIG
being always on, enable MBEDTLS_PSA_CRYPTO_CONFIG
in some TLS 1.3 all.sh test components.
While enabling MBEDTLS_PSA_CRYPTO_CONFIG, disable
some PSA_WANT_ so that the configurations
remain the same for the modified components.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
* restricted/development-restricted: (30 commits)
Tiny fix in ChangeLog pt 2
Tiny fix in ChangeLog
Changelog entry for the RSA memory leak
Edit ChangeLog entry
Update ChangeLog
Add test cases for extKeyUsage
Rationalize extKeyUsage tests
Use P_CLI when O_CLI's status is not reliable
Rationalize keyUsage testing, round 2
Always print detailed cert errors in test programs
Fix 1.3 failure to update flags for (ext)KeyUsage
Rationalize ssl-opt tests for keyUsage
Test cert alert KEY_USAGE -> UNSUPPORTED_CERT
Free allocated memory where methods were returning without freeing
Force MBEDTLS_PSA_HMAC_DRBG_MD_TYPE based on CTR_DRBG
Document that MBEDTLS_PSA_HMAC_DRBG_MD_TYPE does not force HMAC
Clean up constant-flow memsan testing
Improve description of who is affected
More diversified sizes in tests
Fix stack buffer overflow in ECDSA signature format conversions
...
We were only testing the static key store (MBEDTLS_PSA_KEY_STORE_DYNAMIC
disabled) with configs/*.h. Add a component with the static key store and
everything else (including built-in keys), and a component with the static
key store and CTR_DBRG using PSA for AES (which means PSA uses a volatile
key internally).
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Reorganize the structure of the builds of aesce.c and sha256ce.c with
various configurations of AES/SHA256 acceleration. Now everything in a given
configuration is grouped together (building the object file, and checking
the presence or absence of certain instructions in the assembly).
Fix several instances of copypasta in -m options where we were sometimes not
allowing the compiler to use optional instructions even though we use them
in inline assembly, and sometimes allowing the compiler to use optional
instructions even though we then checked their absence
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
In public headers, we want to avoid things that are not standard C++,
including features that GCC and Clang support as extensions, such as
flexible array members. So compile with `-pedantic`.
Non-regression for https://github.com/Mbed-TLS/mbedtls/issues/9020.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Fix local testing leftover when working on
PR 9394. This was causing basic-build-test.sh
to fail on the Open CI (not the internal one
though).
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
Start a server in tf-psa-crypto/tests in addition to
tests to be able to run test suites from
tf-psa-crypto/tests.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
Move the definitions of PSA core and builtin
source paths to pre_initialize_variables
where they are merged with the definition
of the configuration file paths.
Add the 3.6 case by the way as
pre_initialize_variables is planned to
be in the all.sh part that will land in
mbedtls-framework at some point.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
Move the DATA_FILES_PATH_VAR variable into the generated bash rather
than only variablising it in the python that generates the test script.
Signed-off-by: David Horstmann <david.horstmann@arm.com>
Now that we have PSASIM we can really test CRYPTO_CLIENT
functionality and those functions are not needed anymore.
Moreover new test suites that are going to rely on
CRYPTO_CLIENT && !CRYPTO_C would be tested from
test_default_psa_crypto_client_without_crypto_provider()
leading to failures due to stub functions being empty.
Signed-off-by: Valerio Setti <valerio.setti@nordicsemi.no>
Disable asm in memsan constant-flow testing and adjust
test_suit_bignum_core.function accordingly
Signed-off-by: Elena Uziunaite <elena.uziunaite@arm.com>
The header line said “driver vs reference” whereas the error line said
“reference -> driver”. Updated the header line to be consistent with the
error line.
Signed-off-by: Sam Berry <sam.berry@arm.com>
The driver-vs-reference checks compare test results in different
configurations. Ignore the test results that report differences in
configurations that were the point of the comparison. Do compare other
configuration reports: this will let us know if the configurations diverge
in an unexpected way.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Generate option-on and option-off cases for test_suite_config, for all
boolean options (MBEDTLS_xxx and PSA_WANT_xxx, collected from the mbedtls
and PSA config files).
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Remove test_full_no_bignum as duplicate
of the test of the reference config
confgi-symmetric-only.h and
component_test_psa_crypto_config_accel_ecc_no_bignum().
Specifically, component_test_full_no_bignum was
added as part of preparation work for that component.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
Adapt test_full_no_cipher_ components with
MBEDTLS_PSA_CRYPTO_CONFIG enabled.
Remove the component with no PSA crypto and
the one with MBEDTLS_PSA_CRYPTO_CONFIG disabled.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
Adapt test_crypto_full_md_light_only with
MBEDTLS_PSA_CRYPTO_CONFIG enabled.
No need to disable PSA_WANT_ALG_HKDF as
the PSA implementation of HKDF is independent
of hkdf.c and thus of MAC through md.c.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
With replacing the `MD_CAN` macros with `PSA_WANT` counterparts
the pure legacy test cases are needing the config options
from `crypto_config.h`.
Signed-off-by: Gabor Mezei <gabor.mezei@arm.com>
This allows to keep the same CONFIG_H while building the Mbed TLS
library and the application (no matter if it's client or server).
Signed-off-by: Valerio Setti <valerio.setti@nordicsemi.no>
- smoke test client.c becomes a trivial call to psa_crypto_init()
- server.c now uses psa_sim_crypto_server.c's psa_crypto_call()
- Makefile is updated to build all the modules, and allow a different MAIN
- all.sh's test_psasim now tests the simulation of psa_hash_compute() too
Signed-off-by: Tom Cosgrove <tom.cosgrove@arm.com>
Change the way we decide if for a given
configuration we need to run tests with
and without MBEDTLS_USE_PSA_CRYPTO enabled.
That makes the script suitable for 3.6 and
development branch.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
This allows to re-enable MBEDTLS_ENTROPY_NV_SEED since the
seedfile is correctly found in the "test" folder at runtime.
Signed-off-by: Valerio Setti <valerio.setti@nordicsemi.no>
- add quotes to the $@ parameter in helper_crypto_client_build()
- instead of copying mbedtls_config.h to build static libraries,
we rely on the already existing backup/cleanup mechanism which
is available in all.sh.
Signed-off-by: Valerio Setti <valerio.setti@nordicsemi.no>
Since this python module resides in the framework submodule we must
add an extra explicit path to it for mypy.
Signed-off-by: David Horstmann <david.horstmann@arm.com>
Add DEBUG=1 in test_psasim() to helpers and final make to build
the libraries and the final binaries with debug symbols
enabled.
Signed-off-by: Valerio Setti <valerio.setti@nordicsemi.no>
Instead of copying the entire library & include folders twice
to build libraries for client and server:
- change the main config file (mbedtls_config.h)
- build in the root library folder
- move the generated library in the psasim folder
- use those library for linking the client/server binaries
Signed-off-by: Valerio Setti <valerio.setti@nordicsemi.no>
This is necessary because otherwise the library is not able to
find the seedfile at runtime and it fails the initialization.
However since this test runs on a standard PC we can rely on
platform entropy as source of entropy.
Signed-off-by: Valerio Setti <valerio.setti@nordicsemi.no>
This commit implements the first useful IPC communication between
the client and the server. The implemented command is simple,
psa_crypto_init(), and its return value is sent back to the client.
Note: the newly added file psa_functions_codes.h is temporary
and it's probably the one that needs to be automatically
generated by a python script to support all crypto functions.
Signed-off-by: Valerio Setti <valerio.setti@nordicsemi.no>
Ensure that both server and client can call mbedtls_version_get_string_full()
to verify that they are linked against proper libraries.
Note: each side (client/server) performs the call against its own
MbedTLS library. There is no IPC communication involved in this
test. Client/server communication will come later.
Signed-off-by: Valerio Setti <valerio.setti@nordicsemi.no>
It includes changes to:
- tests/Makefile: build the library for client and server in different
folders. It mimica the libtestdriver1 behavior (without functions
renaming though).
- tests/scripts/all.sh: helper function to build for client and
server with some default configuration for each of them.
- crypto_spe.h: this is dummy file taken from the already existing
tests. It's just meant to pacify the compiler, not to provide
something useful. It will likely be changed in the future.
Signed-off-by: Valerio Setti <valerio.setti@nordicsemi.no>
- remove new line at beginning of test_keys.h
- add footer at the end of both generated files
Signed-off-by: Valerio Setti <valerio.setti@nordicsemi.no>
This helps removing the previous pylint exception.
Also use "with" statement for opening the file in order to
ensure that all the content is flushed to the file before
exiting.
Signed-off-by: Valerio Setti <valerio.setti@nordicsemi.no>
Without this fix keys could be listed differently on Ubuntu 16
between different runs therefore causing
check_generated_files() to fail.
Signed-off-by: Valerio Setti <valerio.setti@nordicsemi.no>
On the CI, the git version when running on
Ubuntu 16.04 is 2.7 and it does not support
the "--recurse-submodules" option of
"git ls-files" thus do not use it.
Another argument to not use it is that
when TF-PSA-Crypto will be a submodule of
mbedtls we will not want check_files.py to
check the TF-PSA-Crypto files as well.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
- rewrite output file (do not append)
- remove useless "os" import
- move pylint for main() function
Signed-off-by: Valerio Setti <valerio.setti@nordicsemi.no>
- Embed input arguments inside the script so as to simplify the
calls in Makefiles/CMakeLists.
- add a new "--list-dependencies" command line option to print
out the list of dependencies.
- Modify tests/Makefile accordinlgy.
Signed-off-by: Valerio Setti <valerio.setti@nordicsemi.no>
The previous commit had:
- one obvious mistake (-f NULL with default -e runs nothing)
- one unforeseen issue: OPENSSL_NEXT skips static ECDH
- arguably scope creep: the stated goal was to simplify the full
invocation (in particular, make it obvious that everything is run
without having to remember the default value of EXCLUDE), but it also
made an unrelated change: running most tests with OPENSSL_NEXT (hence
the previous point).
This commit should fix all this, in particular it switches back to
running most tests with OPENSSL and using OPENSSL_NEXT only when needed.
Hopefully in the future we'll do the opposite: most tests will run with
a recent OpenSSL, and only those that need an older one will use
something older. But that will be another PR.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
Prepare this component for PSA_CRYPTO_CONFIG to be on by default.
Rename it so that the name is still accurate when we remove legacy symbols
Signed-off-by: Ryan Everett <ryan.everett@arm.com>
We actually only need two invocations.
This also moves all the default tests to OPENSSL_NEXT, which is good
because OPENSSL is ancient.
I have no idea why NULL doesn't work with OPENSSL_NEXT (1.1.1a) server,
because according to the manpage [1], "ALL,COMPLEMENTOFALL" (which is
what we are using) should do it, and indeed
$OPENSSL_NEXT ciphers "ALL,COMPLEMENTOFALL" | tr ':' '\n'
lists NULL ciphersuites, and also they work client-side with
OPENSSL_NEXT...
[1] https://www.openssl.org/docs/man1.1.1/man1/ciphers.html
Also, while at it, remove partial invocation (only non-default) from one
component, as we already have a full invocation in the same config (plus
ASan) in another component.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
Remove static declaration of look-up table from test_suite_pk
and generate it automatically with Python.
Signed-off-by: Valerio Setti <valerio.setti@nordicsemi.no>
This is to manage RSA and EC keys in the same way in order to
prepare for the following commits.
Signed-off-by: Valerio Setti <valerio.setti@nordicsemi.no>
Most of them (2 exceptions, see below) are of the "driver + built-in"
type, so they're all a subset of test_psa_crypto_driver which tests
everything with driver + built-in at once.
Furthermore, all those components were build-only, while
test_psa_crypto_driver runs the test suites.
Special cases: two of the components looked like they were trying to go for
driver-only (ecdh disabling ECDH_C and hkdf disabling HKDF_C).
For ECDH, built-in would actually be re-enabled because not enough was
accelerated: you also need ECC key types and curves - see
component_test_psa_crypto_config_accel_ecdh which does this correctly.
For HKDF, we don't have test driver support for key derivation yet. I
guess that shows how little testing value these build-only components
really had.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
As the comment says, this component's only goal was to make sure the
legacy+driver test cases in test_suite_md.psa were executed.
But actually these are already executed in
component_test_psa_crypto_drivers which tests with everything having
both a driver and the built-in, as can be seen in the outcomes file.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
asymmetric_key_data.py already provides EC/RSA key pair values that
are suitable for generate_test_keys.py. So instead of re-generating
the keys using gen_key program, we use those keys.
This commit also:
- extends asymmetric_key_data.py to introduce
RSA bit sizes that are used in test_suite_pk but were missing from
asymmetric_key_data.py.
- updates test_keys.h with new keys.
Signed-off-by: Valerio Setti <valerio.setti@nordicsemi.no>
- remove BEGIN_FILE/END_FILE lines from output header file.
- add single disclaimer at the beginning of the file instead
of having it repeated for every array.
- improved exception message for missing key generation program.
This commits also regenerates "test_keys.h" in order to fully
comply with the new format.
Signed-off-by: Valerio Setti <valerio.setti@nordicsemi.no>
This commit adds "generate_test_keys.py" script to generate
predefined keys used in test_suite_pk. Keys are generated with
"programs/pkey/gen_key" tool and converted to C array using
the python script.
tests/src/test_keys.h is automatically generated using the
above mentioned script.
test_suite_pk is updated in order to use the new format.
Signed-off-by: Valerio Setti <valerio.setti@nordicsemi.no>
Update the guards generated by the wrapper generation script to use
!MBEDTLS_PSA_ASSUME_EXCLUSIVE_BUFFERS and regenerate the PSA test
wrappers.
Signed-off-by: David Horstmann <david.horstmann@arm.com>
Replace MBEDTLS_PSA_COPY_CALLER_BUFFERS with inverse:
!MBEDTLS_PSA_ASSUME_EXCLUSIVE_BUFFERS. This ensures that buffer
protection is enabled by default without any change to the Mbed TLS
config file.
Signed-off-by: David Horstmann <david.horstmann@arm.com>
The already existing component_test_psa_crypto_client() is renamed
as component_test_default_psa_crypto_client_without_crypto_provider()
while component_build_full_psa_crypto_client_without_crypto_provider()
was added.
- Both of them check that the missing symbols at link time (if any)
belong to the psa_xxx() family.
- The former builds with default config + CRYPTO_CLIENT - CRYPTO_C and
then runs test suites.
- The latter only perform the builds using the full config and then
it checks that PK-PSA bridge functions are present.
Signed-off-by: Valerio Setti <valerio.setti@nordicsemi.no>
This is possible because after #8740 RSA_C no longer depends on
PK to parse and write private/public keys.
This commit also solves related issues that arose after this change
in "pk.c" and "test_suite_pk". In particular now we can use
rsa's module functions for parsing and writing keys without need
to rely on pk_parse and pk_write functions.
Signed-off-by: Valerio Setti <valerio.setti@nordicsemi.no>
Enabling this causes TSan warnings, as some self-tests use unprotected globals
(see X_count variables in ecp.c). This isn't an issue, as these globals are only
read in self tests, which do not use threads.
Signed-off-by: Ryan Everett <ryan.everett@arm.com>
Custom-e RSA key generation is not yet supported in the test driver, and we
don't support fallback from the test driver to the built-in
implementation (even though we're testing with MBEDTLS_RSA_C). So for the
time being, only run psa_generate_key_ext test cases for RSA with a custom
public exponent when using the built-in implementation.
Add a test case to validate that psa_generate_key_ext test cases for RSA
with a custom public exponent returns NOT_SUPPORTED (rather than silently
doing the wrong thing) when not using built-in RSA (which is subtly
different from when having accelerated RSA: if both are enabled, which we
currently don't do in all.sh, then this should be supported and this is
validated by the test cases above).
This wart will be resolved when we add support for drivers with a
generate_key_ext entry point.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Set unique configuration names in the outcome file. This was lost in the
rewrite from depends-*.pl to depends.py.
Fix#7290
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Release: v23.06_API1.5_ADAC_EAC
This fixes all the issues that were previously added as exceptions.
Signed-off-by: Valerio Setti <valerio.setti@nordicsemi.no>
Since we officially disabled support for importing of PEM formatted keys
into PSA we removed dedicated tests from test_suite_psa_crypto. As a
consequence MBEDTLS_PEM_PARSE_C is no more an exception for
component_check_test_dependencies().
Signed-off-by: Valerio Setti <valerio.setti@nordicsemi.no>
Enable memory poisoning for all functions whose names start with
'psa_pake'. Regenerate the wrappers and commit the result.
Signed-off-by: David Horstmann <david.horstmann@arm.com>
Do not disable RSA_C and related modules because now it does not
automatically re-enable PK module.
Signed-off-by: Valerio Setti <valerio.setti@nordicsemi.no>
When building without PSA crypto functions, disable
MBEDTLS_PSA_CRYPTO_CLIENT as well as MBEDTLS_PSA_CRYPTO_C. With
just MBEDTLS_PSA_CRYPTO_CLIENT, PSA crypto API functions are supposed to
exist at link time but be provided by a third party.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Some platforms may support ASan but be C99-only (no C11 support).
These platforms will support ASan metatests but not memory poisoning,
which requires C11 features.
To allow for this, create a separate platform requirement, "poison",
in metatest.c to distinguish generic ASan metatests from ones that
require suppport for memory poisoning.
In practice our platforms support both, so run "poison" tests in
the same all.sh components where we run "asan" ones.
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>
For now, only instrument the one function for which buffer copying has been
implemented, namely `psa_cipher_encrypt`.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
The new script `tests/scripts/generate_psa_wrappers.py` generates the
implementation of wrapper functions for PSA API functions, as well as a
header that defines macros that redirect calls to the wrapper functions. By
default, the wrapper functions just call the underlying library function.
With `--log`, the wrapper functions log the arguments and return values.
This commit only introduces the new script. Subsequent commits will
integrate the wrappers in the build.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
All no_cipher components have crypto (as in libmbedcrypto.a), but the
difference is one doesn't have PSA crypto while the other two do.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
These were probably leftovers from the development phase of the
associated PR that were not removed in the end.
Signed-off-by: Valerio Setti <valerio.setti@nordicsemi.no>
This reverts commit be978a8c4fc52965b486125f2993251025b1a399.
The feature is no longer needed, and the script is broken if you don't pass
--make-vars.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
This reverts commit 6587959a32f978aeb02766c27cf30b04d8a245e1.
The feature is no longer needed, and the script is broken if you don't pass
--make-vars.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
When building with Make on a Unix-like platform (shell and compiler),
auto-detect configurations that may require linking with pthread.
This removes the need for MAKE_THREADING_FLAGS in all.sh.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
This commit also:
- rename the reference component as component_test_full_block_cipher_legacy_dispatch()
- add a common configuration function, named common_block_cipher_dispatch() that
is used from both accelerated and reference components
Signed-off-by: Valerio Setti <valerio.setti@nordicsemi.no>
This would break platforms that do not have pthread. Put the linking
instead behind a define and add this define where required to all.sh.
Signed-off-by: Paul Elliott <paul.elliott@arm.com>
component_test_tsan now builds and tests the library with
clang ThreadSanitizer enabled.
There are no multi-threaded unit tests so far, the goal is that they are
automatically tested with TSan when they are added.
Signed-off-by: Janos Follath <janos.follath@arm.com>
Support for each test was checked before the command line had been
parsed, causing the support_ functions to ignore arguments that set a
tool's location.
Signed-off-by: Bence Szépkúti <bence.szepkuti@arm.com>
