Add basic miscellaneous components to TF-PSA-Crypto

tests/scripts/components-basic-checks.sh

@@ -18,3 +18,102 @@ components_tf_psa_crypto_check_python_files () {
     msg "Lint: Python scripts"
     $FRAMEWORK/scripts/check-python-files.sh
 }
+
+component_tf_psa_crypto_check_recursion () {
+    msg "Check: recursion.pl" # < 1s
+    $FRAMEWORK/scripts/recursion.pl ${PSA_CORE_PATH}/*.c
+    $FRAMEWORK/scripts/recursion.pl ${BUILTIN_SRC_PATH}/*.c
+}
+
+component_tf_psa_crypto_check_changelog () {
+    msg "Check: changelog entries" # < 1s
+    rm -f ChangeLog.new
+    $FRAMEWORK/scripts/assemble_changelog.py -o ChangeLog.new
+    if [ -e ChangeLog.new ]; then
+        # Show the diff for information. It isn't an error if the diff is
+        # non-empty.
+        diff -u ChangeLog ChangeLog.new || true
+        rm ChangeLog.new
+    fi
+}
+
+component_tf_psa_crypto_check_test_cases () {
+    msg "Check: test case descriptions" # < 1s
+    if [ $QUIET -eq 1 ]; then
+        opt='--quiet'
+    else
+        opt=''
+    fi
+    $FRAMEWORK/scripts/check_test_cases.py $opt
+    unset opt
+}
+
+component_tf_psa_crypto_check_test_dependencies () {
+    msg "Check: test case dependencies: legacy vs PSA" # < 1s
+    # The purpose of this component is to catch unjustified dependencies on
+    # legacy feature macros (MBEDTLS_xxx) in PSA tests. Generally speaking,
+    # PSA test should use PSA feature macros (PSA_WANT_xxx, more rarely
+    # MBEDTLS_PSA_xxx).
+    #
+    # Most of the time, use of legacy MBEDTLS_xxx macros are mistakes, which
+    # this component is meant to catch. However a few of them are justified,
+    # mostly by the absence of a PSA equivalent, so this component includes a
+    # list of expected exceptions.
+
+    found="check-test-deps-found-$$"
+    expected="check-test-deps-expected-$$"
+
+    # Find legacy dependencies in PSA tests
+    grep 'depends_on' \
+        tests/suites/test_suite_psa*.data \
+        tests/suites/test_suite_psa*.function |
+        grep -Eo '!?MBEDTLS_[^: ]*' |
+        grep -v -e MBEDTLS_PSA_ -e MBEDTLS_TEST_ |
+        sort -u > $found
+
+    # Expected ones with justification - keep in sorted order by ASCII table!
+    rm -f $expected
+    # No PSA equivalent - WANT_KEY_TYPE_AES means all sizes
+    echo "!MBEDTLS_AES_ONLY_128_BIT_KEY_LENGTH" >> $expected
+    # No PSA equivalent - used to skip decryption tests in PSA-ECB, CBC/XTS/NIST_KW/DES
+    echo "!MBEDTLS_BLOCK_CIPHER_NO_DECRYPT" >> $expected
+    # MBEDTLS_ASN1_WRITE_C is used by import_rsa_made_up() in test_suite_psa_crypto
+    # in order to build a fake RSA key of the wanted size based on
+    # PSA_VENDOR_RSA_MAX_KEY_BITS. The legacy module is only used by
+    # the test code and that's probably the most convenient way of achieving
+    # the test's goal.
+    echo "MBEDTLS_ASN1_WRITE_C" >> $expected
+    # No PSA equivalent - used in test_suite_psa_crypto to get some "known" size
+    # for raw key generation.
+    echo "MBEDTLS_CTR_DRBG_MAX_REQUEST" >> $expected
+    # No PSA equivalent - we should probably have one in the future.
+    echo "MBEDTLS_ECP_RESTARTABLE" >> $expected
+    # No PSA equivalent - needed by some init tests
+    echo "MBEDTLS_ENTROPY_NV_SEED" >> $expected
+    # No PSA equivalent - required to run threaded tests.
+    echo "MBEDTLS_THREADING_PTHREAD" >> $expected
+
+    # Compare reality with expectation.
+    # We want an exact match, to ensure the above list remains up-to-date.
+    #
+    # The output should be empty. When it's not:
+    # - Each '+' line is a macro that was found but not expected. You want to
+    # find where that macro occurs, and either replace it with PSA macros, or
+    # add it to the exceptions list above with a justification.
+    # - Each '-' line is a macro that was expected but not found; it means the
+    # exceptions list above should be updated by removing that macro.
+    diff -U0 $expected $found
+
+    rm $found $expected
+}
+
+component_tf_psa_crypto_check_test_helpers () {
+    msg "unit test: generate_test_code.py"
+    # unittest writes out mundane stuff like number or tests run on stderr.
+    # Our convention is to reserve stderr for actual errors, and write
+    # harmless info on stdout so it can be suppress with --quiet.
+    ./framework/scripts/test_generate_test_code.py 2>&1
+
+    msg "unit test: translate_ciphers.py"
+    python3 -m unittest framework/scripts/translate_ciphers.py 2>&1
+}