Use CodedInputStream.newInstance(ByteBuffer) for non-segmented messages

This is to improve the decode performance of small messages (<16 KB) to avoid a message-sized allocation and copy.

Protobuf already has an optimized code path for [decoding a direct `ByteBuffer`](https://github.com/google/protobuf/blob/v3.2.0/java/core/src/main/java/com/google/protobuf/CodedInputStream.java#L141). We currently [copy to a `byte[]`](https://github.com/grpc/grpc-java/blob/v1.2.0/protobuf-lite/src/main/java/io/grpc/protobuf/lite/ProtoLiteUtils.java#L139), but that could be avoided because small messages will commonly be in a single `ByteBuffer`.

This will require adding a new interface for retrieving the `ByteBuffer` from the `InputStream`. I'm uncertain whether the interface should support returning multiple `ByteBuffer`s.

Read protobuf-lite/src/main/java/io/grpc/protobuf/lite/ProtoLiteUtils.java around the current byte[] copy, then compare it with CodedInputStream.java's direct ByteBuffer path. Determine how the InputStream interface should expose ByteBuffer data, including the stated question about multiple buffers; done means non-segmented small messages can use the optimized path without the message-sized allocation and copy.

索引モデルが issue の本文から書いたものです。