Add packed-c protocol support for crypto service
To provide a lightweight parameter encoding that is aligned to
conventions used by SCMI, the packed-c parameter serialization has
been added to the crypto service. This builds on generic
components that allow other packed-c service access protocols
to be added easily. Service level tests have been extended to
use both protobuf and packed-c clients.
Signed-off-by: julhal01 <julian.hall@arm.com>
Change-Id: I9279b0814bcc9cf6c4aa4e30629e2f46f2df4c23
diff --git a/components/common/tlv/test/tlv_tests.cpp b/components/common/tlv/test/tlv_tests.cpp
new file mode 100644
index 0000000..661a810
--- /dev/null
+++ b/components/common/tlv/test/tlv_tests.cpp
@@ -0,0 +1,207 @@
+/*
+ * Copyright (c) 2020-2021, Arm Limited and Contributors. All rights reserved.
+ *
+ * SPDX-License-Identifier: BSD-3-Clause
+ */
+
+#include <cstring>
+#include <common/tlv/tlv.h>
+#include <CppUTest/TestHarness.h>
+
+TEST_GROUP(TlvTests)
+{
+
+};
+
+TEST(TlvTests, decodeGoodRecords)
+{
+ struct tlv_const_iterator iter;
+ struct tlv_record decoded_record;
+
+ /* Case 1: Valid single zero length record */
+ const uint8_t case_1[] = {
+ 0x77, 0x31, 0x00, 0x00
+ };
+
+ tlv_const_iterator_begin(&iter, case_1, sizeof(case_1));
+ CHECK_TRUE(tlv_decode(&iter, &decoded_record));
+ UNSIGNED_LONGS_EQUAL(0x7731, decoded_record.tag);
+ UNSIGNED_LONGS_EQUAL(0, decoded_record.length);
+ CHECK_FALSE(tlv_decode(&iter, &decoded_record));
+
+ /* Case 2: Two valid records */
+ const uint8_t case_2[] = {
+ 0x01, 0x10, 0x00, 0x01, 0x50,
+ 0x02, 0x11, 0x00, 0x04, 0x01, 0x02, 0x03, 0x04
+ };
+
+ tlv_const_iterator_begin(&iter, case_2, sizeof(case_2));
+ CHECK_TRUE(tlv_decode(&iter, &decoded_record));
+ UNSIGNED_LONGS_EQUAL(0x0110, decoded_record.tag);
+ UNSIGNED_LONGS_EQUAL(1, decoded_record.length);
+ UNSIGNED_LONGS_EQUAL(0x50, decoded_record.value[0]);
+ CHECK_TRUE(tlv_decode(&iter, &decoded_record));
+ UNSIGNED_LONGS_EQUAL(0x0211, decoded_record.tag);
+ UNSIGNED_LONGS_EQUAL(4, decoded_record.length);
+ UNSIGNED_LONGS_EQUAL(0x01, decoded_record.value[0]);
+ UNSIGNED_LONGS_EQUAL(0x02, decoded_record.value[1]);
+ UNSIGNED_LONGS_EQUAL(0x03, decoded_record.value[2]);
+ UNSIGNED_LONGS_EQUAL(0x04, decoded_record.value[3]);
+ CHECK_FALSE(tlv_decode(&iter, &decoded_record));
+}
+
+TEST(TlvTests, findAndDecode)
+{
+ struct tlv_const_iterator iter;
+ struct tlv_record decoded_record;
+
+ /*
+ * Checks use of the tlv_find_decode method to extract known records
+ * from a sequence of records, encoded in ascending tag order, that includes
+ * ones that the receiver isn't interested in.
+ */
+ const uint8_t encoded_records[] = {
+ 0x00, 0x01, 0x00, 0x00,
+ 0x00, 0x02, 0x00, 0x03, 0x88, 0x88, 0x88,
+ 0x00, 0x07, 0x00, 0x02, 0x77, 0x77,
+ 0x00, 0x09, 0x00, 0x01, 0x77
+ };
+
+ tlv_const_iterator_begin(&iter, encoded_records, sizeof(encoded_records));
+ CHECK_TRUE(tlv_find_decode(&iter, 0x0001, &decoded_record));
+ UNSIGNED_LONGS_EQUAL(0, decoded_record.length);
+ CHECK_TRUE(tlv_find_decode(&iter, 0x0007, &decoded_record));
+ UNSIGNED_LONGS_EQUAL(2, decoded_record.length);
+ CHECK_TRUE(tlv_find_decode(&iter, 0x0009, &decoded_record));
+ UNSIGNED_LONGS_EQUAL(1, decoded_record.length);
+}
+
+TEST(TlvTests, decodeBadRecords)
+{
+ struct tlv_const_iterator iter;
+ struct tlv_record decoded_record;
+
+ /* Case 1: Too short to accommodate a valid header */
+ const uint8_t case_1[] = {
+ 0x77, 0x31, 0x00
+ };
+
+ tlv_const_iterator_begin(&iter, case_1, sizeof(case_1));
+ CHECK_FALSE(tlv_decode(&iter, &decoded_record));
+
+ /* Case 1: A complete record followed by a truncated one */
+ const uint8_t case_2[] = {
+ 0x77, 0x31, 0x00, 0x00,
+ 0x03, 0x21, 0x00, 0x03, 0xaa
+ };
+
+ tlv_const_iterator_begin(&iter, case_2, sizeof(case_2));
+ CHECK_TRUE(tlv_decode(&iter, &decoded_record));
+ CHECK_FALSE(tlv_decode(&iter, &decoded_record));
+}
+
+TEST(TlvTests, encodeRecords)
+{
+ struct tlv_iterator iter;
+ struct tlv_record record_to_encode;
+ size_t required_space;
+
+ /* Case 1: Encode zero length record */
+ const uint8_t case_1_expected[] = {
+ 0x66, 0x77, 0x00, 0x00
+ };
+ record_to_encode.tag = 0x6677;
+ record_to_encode.length = 0;
+ record_to_encode.value = NULL;
+ required_space = tlv_required_space(record_to_encode.length);
+ uint8_t case_1_actual[required_space];
+ tlv_iterator_begin(&iter, case_1_actual, required_space);
+ CHECK_TRUE(tlv_encode(&iter, &record_to_encode));
+ MEMCMP_EQUAL(case_1_expected, case_1_actual, required_space);
+
+ /* Case 2: Encode two records */
+ const uint8_t case_2_expected[] = {
+ 0xa8, 0xa9, 0x00, 0x01, 0x88,
+ 0xa8, 0xaa, 0x00, 0x02, 0x01, 0x02
+ };
+
+ required_space = tlv_required_space(1) + tlv_required_space(2);
+ uint8_t case_2_actual[required_space];
+ tlv_iterator_begin(&iter, case_2_actual, required_space);
+
+ record_to_encode.tag = 0xa8a9;
+ record_to_encode.length = 1;
+ record_to_encode.value = &case_2_expected[4];
+ CHECK_TRUE(tlv_encode(&iter, &record_to_encode));
+
+ record_to_encode.tag = 0xa8aa;
+ record_to_encode.length = 2;
+ record_to_encode.value = &case_2_expected[9];
+ CHECK_TRUE(tlv_encode(&iter, &record_to_encode));
+
+ MEMCMP_EQUAL(case_2_expected, case_2_actual, required_space);
+
+ /* Check that you can't encode beyond the limit of the buffer */
+ CHECK_FALSE(tlv_encode(&iter, &record_to_encode));
+}
+
+TEST(TlvTests, encodeInsufficientSpace)
+{
+ struct tlv_iterator iter;
+ struct tlv_record record_to_encode;
+ size_t required_space;
+
+ /* Case 1: Encode record into buffer that isn't big enough */
+ const uint8_t case_1_expected[] = {
+ 0x66, 0x77, 0x00, 0x03, 0x01, 0x02, 0x03
+ };
+ record_to_encode.tag = 0x6677;
+ record_to_encode.length = 3;
+ record_to_encode.value = &case_1_expected[4];
+ required_space = tlv_required_space(record_to_encode.length) - 1;
+ uint8_t case_1_actual[required_space];
+ tlv_iterator_begin(&iter, case_1_actual, required_space);
+ CHECK_FALSE(tlv_encode(&iter, &record_to_encode));
+}
+
+TEST(TlvTests, encodeWrongOrder)
+{
+ struct tlv_iterator iter;
+ struct tlv_record record_to_encode;
+ size_t required_space;
+
+ /* Check defence against encoding successive records
+ * out of tag value order. Encoding rules require
+ * records to be in ascending tag value order.
+ */
+ const uint8_t record_value[] = {
+ 0x11, 0x22, 0x33, 0x44
+ };
+
+ /* Attemps to encode 4 records, the first 3 obey order
+ * rule, the last one doesn't
+ */
+ required_space = tlv_required_space(sizeof(record_value)) * 4;
+ uint8_t encode_buffer[required_space];
+ tlv_iterator_begin(&iter, encode_buffer, required_space);
+
+ record_to_encode.tag = 1;
+ record_to_encode.length = sizeof(record_value);
+ record_to_encode.value = record_value;
+ CHECK_TRUE(tlv_encode(&iter, &record_to_encode));
+
+ record_to_encode.tag = 2;
+ record_to_encode.length = sizeof(record_value);
+ record_to_encode.value = record_value;
+ CHECK_TRUE(tlv_encode(&iter, &record_to_encode));
+
+ record_to_encode.tag = 2;
+ record_to_encode.length = sizeof(record_value);
+ record_to_encode.value = record_value;
+ CHECK_TRUE(tlv_encode(&iter, &record_to_encode));
+
+ record_to_encode.tag = 1;
+ record_to_encode.length = sizeof(record_value);
+ record_to_encode.value = record_value;
+ CHECK_FALSE(tlv_encode(&iter, &record_to_encode));
+}