vlc源码分析(三) 调用live555接收RTSP数据

http://www.cnblogs.com/jiayayao/p/6819722.html

 

 首先了解RTSP/RTP/RTCP相关概念,尤其是了解RTP协议:RTP与RTCP协议介绍(转载)

  vlc使用模块加载机制调用live555,调用live555的文件是live555.cpp。

一、几个重要的类  

  以下向左箭头(“<-”)为继承关系。

1. RTPInterface

  RTPInterface是RTPSource的成员变量,其成员函数handleRead会读取网络数据存入BufferedPacket内,该类最终会调到UDP的发送接收函数。

Boolean RTPInterface::handleRead(unsigned char* buffer, unsigned bufferMaxSize,
                 unsigned& bytesRead, struct sockaddr_in& fromAddress, Boolean& packetReadWasIncomplete)

2. BufferedPacket

  BufferedPacket:用于存储媒体的RTP数据包

  BufferedPacket<-H264BufferedPacket:用于存储H264媒体RTP数据包

  该类有一个重要函数fillInData,是由RTPInterface读取数据存入包中。

Boolean BufferedPacket::fillInData(RTPInterface& rtpInterface, Boolean& packetReadWasIncomplete);

  相对于BufferedPacket,有对应的工厂类:

  BufferedPacketFactory:工厂模式生成BufferedPacket包
  BufferedPacketFactory<-H264BufferedPacketFactory:专门生产H264BufferedPacket的工厂

  在SessionsSetup的时候(也是模块加载的时候),会根据Source类型,选定生产BufferedPacket的工厂类型,即如果Source是H264格式的话,就会new H264BufferedPacketFactory,之后在接收数据的时候就会生产H264BufferedPacket用于存储H264媒体数据。

  ReorderingPacketBuffer:MultiFramedRTPSource的成员变量,用于管理多个BufferedPacket。

3. Source相关类  

  Source相关类的继承关系:Medium<-MediaSource<-FramedSource<-RTPSource<-MultiFramedRTPSource<-H264VideoRTPSource。
  在SessionsSetup的时候,会根据数据源的类型,选定Source的类型,即如果数据源是H264格式的话,就会调用

static H264VideoRTPSource* createNew(UsageEnvironment& env, Groupsock* RTPgs,
  unsigned char rtpPayloadFormat,
  unsigned rtpTimestampFrequency = 90000);

二、播放流程的建立

  播放流程的建立可以参考vlc源码分析之播放流程

三、接收RTSP数据

  vlc在播放IPC时,会开启一个线程接收网络数据,该线程接收网络数据后会调用Demux()进行分离(因为可能是音频,也可能是视频)。Demux()首先将必要的接口,如StreamRead、StreamClose注册下去,然后就进入事件循环:

p_sys->scheduler->doEventLoop( &p_sys->event_data );

  如果有网络数据到来了,Demux()会做两件事,第一件事是分析RTP包,放入ReorderingPacketBuffer管理的BufferedPacket中,堆栈如下图所示:

  第二件事是读取的BufferedPacket,进行一系列拆包操作后,将数据放入数据fifo中,堆栈如下图所示:

  doEventLoop会进入死循环,直到p_sys->event_data的值被中断或者超时改变,从而退出循环。当有网络数据到来的时候,doEventLoop会执行SingleStep->...->doGetNextFrame1(),在doGetNextFrame1()函数中读取RTP数据。这个过程的代码及注释如下:

复制代码
// 做了两件事,一件是分析RTP包,放入ReorderingPacketBuffer管理的BufferedPacket中;
// 另一件是读取的BufferedPacket,进行一系列拆包操作后,将数据放入数据fifo中
void MultiFramedRTPSource::networkReadHandler1() {
  BufferedPacket* bPacket = fPacketReadInProgress;
  if (bPacket == NULL) {
    // Normal case: Get a free BufferedPacket descriptor to hold the new network packet:
    bPacket = fReorderingBuffer->getFreePacket(this);
  }

  // Read the network packet, and perform sanity checks on the RTP header:
  Boolean readSuccess = False;
  // do-while(0)结构,出现错误直接break
  do {
    Boolean packetReadWasIncomplete = fPacketReadInProgress != NULL;
    if (!bPacket->fillInData(fRTPInterface, packetReadWasIncomplete)) break;
    if (packetReadWasIncomplete) {
      // We need additional read(s) before we can process the incoming packet:
      fPacketReadInProgress = bPacket;
      return;
    } else {
      fPacketReadInProgress = NULL;
    }
#ifdef TEST_LOSS
    setPacketReorderingThresholdTime(0);
       // don't wait for 'lost' packets to arrive out-of-order later
    if ((our_random()%10) == 0) break; // simulate 10% packet loss
#endif

    // Check for the 12-byte RTP header:
    if (bPacket->dataSize() < 12) break;
    // 读取RTP头,向前移4个字节
    unsigned rtpHdr = ntohl(*(u_int32_t*)(bPacket->data())); ADVANCE(4);
    // 读取RTP头中的标记位
    Boolean rtpMarkerBit = (rtpHdr&0x00800000) != 0;
    // 读取时间戳,向前移4个字节
    unsigned rtpTimestamp = ntohl(*(u_int32_t*)(bPacket->data()));ADVANCE(4);
    // 读取SSRC,向前移4个字节
    unsigned rtpSSRC = ntohl(*(u_int32_t*)(bPacket->data())); ADVANCE(4);

    // Check the RTP version number (it should be 2):
    // 检查RTP头版本,不是2的话,break
    if ((rtpHdr&0xC0000000) != 0x80000000) break;

    // Skip over any CSRC identifiers in the header:
    // 跳过CSRC计数字节
    unsigned cc = (rtpHdr>>24)&0xF;
    if (bPacket->dataSize() < cc) break;
    ADVANCE(cc*4);

    // Check for (& ignore) any RTP header extension
    // 如果扩展头标志被置位
    if (rtpHdr&0x10000000) {
      if (bPacket->dataSize() < 4) break;
      // 获取扩展头
      unsigned extHdr = ntohl(*(u_int32_t*)(bPacket->data())); ADVANCE(4);
      // 获取扩展字节数
      unsigned remExtSize = 4*(extHdr&0xFFFF);
      if (bPacket->dataSize() < remExtSize) break;
      // 直接跳过扩展字节???
      ADVANCE(remExtSize);
    }

    // Discard any padding bytes:
    // 如果填充标志被置位,直接丢弃不处理
    if (rtpHdr&0x20000000) {
      if (bPacket->dataSize() == 0) break;
      unsigned numPaddingBytes
    = (unsigned)(bPacket->data())[bPacket->dataSize()-1];
      if (bPacket->dataSize() < numPaddingBytes) break;
      bPacket->removePadding(numPaddingBytes);
    }
    // Check the Payload Type.
    // 检查载荷类型,如果源数据H264类型,则其值为96
    // 如果与我们生成的source类型不同,则break
    if ((unsigned char)((rtpHdr&0x007F0000)>>16)
    != rtpPayloadFormat()) {
      break;
    }

    // The rest of the packet is the usable data.  Record and save it:
    if (rtpSSRC != fLastReceivedSSRC) {
      // The SSRC of incoming packets has changed.  Unfortunately we don't yet handle streams that contain multiple SSRCs,
      // but we can handle a single-SSRC stream where the SSRC changes occasionally:
      fLastReceivedSSRC = rtpSSRC;
      fReorderingBuffer->resetHaveSeenFirstPacket();
    }
    // RTP包序号,随RTP数据包而自增,由接收者用来探测包损失
    unsigned short rtpSeqNo = (unsigned short)(rtpHdr&0xFFFF);
    Boolean usableInJitterCalculation
      = packetIsUsableInJitterCalculation((bPacket->data()),
                          bPacket->dataSize());
    struct timeval presentationTime; // computed by:
    Boolean hasBeenSyncedUsingRTCP; // computed by:
    // 根据数据包的一些信息,进行一些计算和记录
    receptionStatsDB()
      .noteIncomingPacket(rtpSSRC, rtpSeqNo, rtpTimestamp,
              timestampFrequency(),
              usableInJitterCalculation, presentationTime,
              hasBeenSyncedUsingRTCP, bPacket->dataSize());

    // Fill in the rest of the packet descriptor, and store it:
    struct timeval timeNow;
    gettimeofday(&timeNow, NULL);
    // 将计算所得的一些参数再赋值到包中
    bPacket->assignMiscParams(rtpSeqNo, rtpTimestamp, presentationTime,
                  hasBeenSyncedUsingRTCP, rtpMarkerBit,
                  timeNow);
    // 经过以上判断和检查,没有发现问题,则由管理类fReorderingBuffer存储包
    if (!fReorderingBuffer->storePacket(bPacket)) break;

    readSuccess = True;// 读取成功
  } while (0);
  if (!readSuccess) fReorderingBuffer->freePacket(bPacket);// 如果读取不成功,则释放内存

  // 将读取到的数据包送至数据fifo中,等待解码线程解码
  doGetNextFrame1();
  // If we didn't get proper data this time, we'll get another chance
}
复制代码

  将读取到的数据包送至数据fifo中,之后就是等待解码线程从数据fifo中拿数据,解码和渲染了,具体可参考vlc源码分析之播放流程

  附:

  配置好的Windows版vlc工程下载:https://github.com/jiayayao/vlc_2.1.0-vs_2010,下载后使用vs2010可以直接编译运行,调试学习非常方便。

 

posted @ 2017-07-18 18:38  STN_LCD  阅读(829)  评论(0编辑  收藏  举报