Incorrect key length used in KeyBuilder tests with TYPE_HMAC keys

[nc-adnan]

According to the Java Card 3.0.5 API, the keyLength parameter passed to the KeyBuilder.buildKey method is the key size in bits.

In the KeyBuilder tests configs, the key sizes are specified in bytes for TYPE_HMAC keys (64/128) which leads to incorrect results:

see JCAlgTest/AlgTest_JClient/src/algtestjclient/SingleModeTest.java:

KEYBUILDER_TEST_CFGS.put("TYPE_HMAC_TRANSIENT_RESET#2.2.2", new KBTestCfg(JCConsts.KeyBuilder_TYPE_HMAC_TRANSIENT_RESET, (short) 64));
KEYBUILDER_TEST_CFGS.put("TYPE_HMAC_TRANSIENT_DESELECT#2.2.2", new KBTestCfg(JCConsts.KeyBuilder_TYPE_HMAC_TRANSIENT_DESELECT, (short) 64));
KEYBUILDER_TEST_CFGS.put("TYPE_HMAC LENGTH_HMAC_SHA_1_BLOCK_64#2.2.2", new KBTestCfg(JCConsts.KeyBuilder_TYPE_HMAC, (short) 64));
KEYBUILDER_TEST_CFGS.put("TYPE_HMAC LENGTH_HMAC_SHA_256_BLOCK_64#2.2.2", new KBTestCfg(JCConsts.KeyBuilder_TYPE_HMAC, (short) 64));
KEYBUILDER_TEST_CFGS.put("TYPE_HMAC LENGTH_HMAC_SHA_384_BLOCK_64#2.2.2", new KBTestCfg(JCConsts.KeyBuilder_TYPE_HMAC, (short) 128));
KEYBUILDER_TEST_CFGS.put("TYPE_HMAC LENGTH_HMAC_SHA_512_BLOCK_64#2.2.2", new KBTestCfg(JCConsts.KeyBuilder_TYPE_HMAC, (short) 128));

[petrs]

Thank you @nc-adnan for reporting, it will take longer to fix this issue due to need for re-measurement of existing results for these six values. So I'm leaving the issue open for now.

[petrs]

fixed by 6e01f83