C2: allow vectorization of HeapByteBuffer.putInt loops

Loops like this:
    static final ByteBuffer buf = ByteBuffer.allocate(SIZE * 4);

    void floss(ByteBuffer b, int n) {
        for (int i = 0; i < SIZE; i++) {
            buf.putInt(i<<2, n);
        }
    }
are currently not vectorized.