if uint64(s.counter)+blocksPerBuf > 1<<32 {
		s.buf = [bufSize]byte{}
		numBlocks := (len(src) + blockSize - 1) / blockSize
		buf := s.buf[bufSize-numBlocks*blockSize:]
		copy(buf, src)
		s.xorKeyStreamBlocksGeneric(buf, buf)
		s.len = len(buf) - copy(dst, buf)
		return
	}

	// If we have a partial (multi-)block, pad it for xorKeyStreamBlocks, and
	// keep the leftover keystream for the next XORKeyStream invocation.