关于 java 位运算

Question

这段代码中和0x0ff与运算是什么意思？ 获取length的原理是什么？
其中对象in如下

// read and check header 
int type = in.read() & 0x0ff;
int identifier = in.read() & 0x0ff;
int length = (in.read() & 0x0ff) << 8 | (in.read() & 0x0ff);

所有的代码

/**

• Reads a Radius packet from the given input stream and

• creates an appropiate RadiusPacket descendant object.

• Reads in all attributes and returns the object.

• Decodes the encrypted fields and attributes of the packet.

• @param dictionary dictionary to use for attributes

• @param sharedSecret shared secret to be used to decode this packet

• @param request Radius request packet if this is a response packet to be

• decoded, null if this is a request packet to be decoded

• @return new RadiusPacket object

• @exception IOException if an IO error occurred

• @exception RadiusException if the Radius packet is malformed
  */

1. static RadiusPacket decodePacket(Dictionary dictionary, InputStream in, String sharedSecret, RadiusPacket request)

2. IOException, RadiusException {

      // check shared secret
      if (sharedSecret == null || sharedSecret.length() == 0)
          throw new RuntimeException("no shared secret has been set");
   
      // check request authenticator
      if (request != null && request.getAuthenticator() == null)
          throw new RuntimeException("request authenticator not set");
   
      // read and check header
      int type = in.read() & 0x0ff;
      int identifier = in.read() & 0x0ff;
      int length = (in.read() & 0x0ff) << 8 | (in.read() & 0x0ff);
   
      if (request != null && request.getPacketIdentifier() != identifier)
          throw new RadiusException("bad packet: invalid packet identifier (request: " + request.getPacketIdentifier() + ", response: " + identifier);
      if (length < RADIUS_HEADER_LENGTH)
          throw new RadiusException("bad packet: packet too short (" + length + " bytes)");
      if (length > MAX_PACKET_LENGTH)
          throw new RadiusException("bad packet: packet too long (" + length + " bytes)");
   
      // read rest of packet
      byte[] authenticator = new byte[16];
      byte[] attributeData = new byte[length - RADIUS_HEADER_LENGTH];
      in.read(authenticator);
      in.read(attributeData);
   
      // check and count attributes
      int pos = 0;
      int attributeCount = 0;
      while (pos < attributeData.length) {
          if (pos + 1 >= attributeData.length)
              throw new RadiusException("bad packet: attribute length mismatch");
          int attributeLength = attributeData[pos + 1] & 0x0ff;
          if (attributeLength < 2)
              throw new RadiusException("bad packet: invalid attribute length");
          pos += attributeLength;
          attributeCount++;
      }
      if (pos != attributeData.length)
          throw new RadiusException("bad packet: attribute length mismatch");
   
      // create RadiusPacket object; set properties
      RadiusPacket rp = createRadiusPacket(type);
      rp.setPacketType(type);
      rp.setPacketIdentifier(identifier);
      rp.authenticator = authenticator;
      
      // load attributes
      pos = 0;
      while (pos < attributeData.length) {
          int attributeType = attributeData[pos] & 0x0ff;
          int attributeLength = attributeData[pos + 1] & 0x0ff;
          RadiusAttribute a = RadiusAttribute.createRadiusAttribute(dictionary, -1, attributeType);
          a.readAttribute(attributeData, pos, attributeLength);
          rp.addAttribute(a);
          pos += attributeLength;
      }
      
      // request packet?
      if (request == null) {
          // decode attributes
          rp.decodeRequestAttributes(sharedSecret);
          rp.checkRequestAuthenticator(sharedSecret, length, attributeData);
      } else {
          // response packet: check authenticator
          rp.checkResponseAuthenticator(sharedSecret, length, attributeData, request.getAuthenticator());
      }
      
      return rp;

   }

Answer 1

0x0ff 就是

0000 0000 1111 1111

一个数字和 0x0ff 进行与运算，则会把高8位置0，低8位保留。

---

回到你的程序中，通过这个函数猜测应该是解析的 RFC2866 报文：

报文格式：

 0                   1                   2                   3
 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
|     Code      |  Identifier   |            Length             |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
|                                                               |
|                         Authenticator                         |
|                                                               |
|                                                               |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
|  Attributes ...
+-+-+-+-+-+-+-+-+-+-+-+-+-

根据格式， 0-7位是 CODE 信息，在你的函数中用 type 存储。（注意 type 的类型，是 int）

@brayden 提到了，InputStream.read() 一个读一个 byte，正好是 8 位，所以，如果写 byte type = in.read() 就正好了。但是 type 的类型是 int，因此我们需要处理一下。

如果我们读入的是 5，那么当我们把他转换成 int 时，也应该是 5。那么需要在前面添加 0。

            0000 0101
& 0000 0000 1111 1111
-------------------------------
  0000 0000 0000 0101

同理，读报文的 8-15位。

length 的值比较难读取，因为它一共16位，而 in.read() 每次只能读8个字节，那么就需要读取两次，然后拼凑。方法是，先读8位，然后左移，再读8位。

int length = (in.read() & 0x0ff) << 8 | (in.read() & 0x0ff);

Answer 2

java里InputStream.read()一个读一个byte, 但是返回一个int, 四字节; 但实际上有用的只是最后的那个byte. 所以需要每次in.read() & 0x0ff;

具体到这段代码,
读第一个byte到一个int型, 赋值给type;
读第二个byte到一个int型, 赋值给identifier;
读第三个byte, 第四个byte, 第三个字节左移8位, 和第四个字节拼接起来, 赋值给length.