Address ANAPIC review comments -- hal.

- Document that IPreparedModel::execute*WithConfig are for single-time
  usages, and IPreparedModel::createReusableExecution is for reusable
  usages.
- Make PrepareModelConfig::cacheToken a fixed sized array.

Fixes: 215566186
Test: VtsHalNeuralnetworksTargetTest
Test: NNT_static
Change-Id: I9c5a49c46a3eac2828b9eb666f3d742038493121
diff --git a/neuralnetworks/aidl/vts/functional/Utils.h b/neuralnetworks/aidl/vts/functional/Utils.h
index 4e0a4aa..ccb0778 100644
--- a/neuralnetworks/aidl/vts/functional/Utils.h
+++ b/neuralnetworks/aidl/vts/functional/Utils.h
@@ -21,6 +21,7 @@
 #include <gtest/gtest.h>
 
 #include <algorithm>
+#include <array>
 #include <iosfwd>
 #include <string>
 #include <utility>
@@ -47,6 +48,7 @@
 inline constexpr int64_t kOmittedTimeoutDuration = -1;
 inline constexpr int64_t kNoDuration = -1;
 inline const std::vector<uint8_t> kEmptyCacheToken(IDevice::BYTE_SIZE_OF_CACHE_TOKEN);
+inline const std::array<uint8_t, IDevice::BYTE_SIZE_OF_CACHE_TOKEN> kEmptyCacheTokenArray{};
 
 // Returns the amount of space needed to store a value of the specified type.
 //