Fix bugs impacting fuzzing coverage

- Update ConvertDataToValidEnum to not use FuzzedDataProvider since it
  causes unexpected parsing results.
- Add OEMCryptoLicenseAPIFuzz::LoadLicenseWithGenericCryptoKeys so that
  generic crypto fuzz tests can load appropriate keys.
- Remove custom mutator from oemcrypto_generic_verify_fuzz because it
  provides minimal additional coverage.
- Refresh affected corpus files.

Merged from https://widevine-internal-review.googlesource.com/168557
Merged from https://widevine-internal-review.googlesource.com/171191
Merged from https://widevine-internal-review.googlesource.com/172170
Merged from https://widevine-internal-review.googlesource.com/172250

Change-Id: Ie676a36cbf4c12bdda9566fad3590a7b69168d9c

libwvdrmengine/oemcrypto/test/fuzz_tests/oemcrypto_decrypt_cenc_fuzz.cc

@@ -45,7 +45,7 @@ extern "C" int LLVMFuzzerTestOneInput(const uint8_t* data, size_t size) {
   FuzzedDataProvider fuzzed_data(inputs[0].data, inputs[0].size);
   fuzzed_data.ConsumeData(&fuzzed_structure, sizeof(fuzzed_structure));
   wvoec::ConvertDataToValidEnum(OEMCrypto_CipherMode_MaxValue,
-                                &fuzzed_structure.cipher_mode);
+                                fuzzed_structure.cipher_mode);
 
   // Allocate sample descriptions.
   std::vector<OEMCrypto_SampleDescription> sample_descriptions(
@@ -100,7 +100,7 @@ extern "C" int LLVMFuzzerTestOneInput(const uint8_t* data, size_t size) {
         MAX_FUZZ_SAMPLE_SIZE + 1;
     wvoec::ConvertDataToValidEnum(
         OEMCrypto_BufferType_MaxValue,
-        &fuzzed_sample_description.buffers.output_descriptor.type);
+        fuzzed_sample_description.buffers.output_descriptor.type);
     fuzzed_sample_description.buffers.output_descriptor.buffer_config %=
         MAX_FUZZ_SAMPLE_SIZE + 1;