: jank : : 2101 : 2020-10-29 08:05 rpc
一 、服务端初始化
go grpc 服务端代码整体采用建造者模式， 服务初始化时，嵌入所有需要加载的接口服务配置，并加载 ，如 是否加密传输 ：
// Create tls based credential.
creds, err := credentials.NewServerTLSFromFile
(testdata.Path("server1.pem"), testdata.Path("server1.key"))
if err != nil {
log.Fatalf("failed to create credentials: %v", err)
}
s := grpc.NewServer(grpc.Creds(creds))
服务初始化 NewServer(...)代码如下：
func NewServer(opt ...ServerOption) *Server {
opts := defaultServerOptions //default server options
for _, o := range opt {
o.apply(&opts) //iterate through opt to implement opts
}
s := &Server{
lis: make(map[net.Listener]bool),
opts: opts,
conns: make(map[transport.ServerTransport]bool), //connect cache
m: make(map[string]*service),
quit: grpcsync.NewEvent(),
done: grpcsync.NewEvent(),
czData: new(channelzData), //rpc_util.go
}
chainUnaryServerInterceptors(s) //only zero(default) / one unary server intercetor, (./intercept.go)
chainStreamServerInterceptors(s) //the same as above
s.cv = sync.NewCond(&s.mu)
if EnableTracing { //global config in ./trace.go use google.org/x/net/trace
_, file, line, _ := runtime.Caller(1)
//golang.org/x/net/trace/event.go
s.events = trace.NewEventLog("grpc.Server", fmt.Sprintf("%s:%d", file, line))
}
if s.opts.numServerWorkers > 0 { //default 0, multi goroutine( <= cpu kernel) to process
s.initServerWorkers()
}
if channelz.IsOn() { //internal/channelz/funcs.go TurnOn()
s.channelzID = channelz.RegisterServer(&channelzServer{s}, "") //id int64, default 1, ... add 1
}
return s
}二、服务启动
1.服务外层启动net 服务并传入grpc Serve中进行监听
2.初始化后获得Sever 结构，通过Sever 结构进行服务的配置启动相应的服务
3.在服务监听过程中如果出现Temporary类型的错误，则进行重试，否则返回结束服务接收
4.在循环处理中每次都启用一个新的goroutine 去运行新的客户端连接
关键代码如下：
// Serve accepts incoming connections on the listener lis, creating a new
// ServerTransport and service goroutine for each. The service goroutines
// read gRPC requests and then call the registered handlers to reply to them.
// Serve returns when lis.Accept fails with fatal errors. lis will be closed when
// this method returns.
// Serve will return a non-nil error unless Stop or GracefulStop is called.
func (s *Server) Serve(lis net.Listener) error {
...
//wait to stop server
s.serveWG.Add(1)
defer func() {
s.serveWG.Done()
if s.quit.HasFired() {
// Stop or GracefulStop called; block until done and return nil.
<-s.done.Done()
}
}()
ls := &listenSocket{Listener: lis}
s.lis[ls] = true //add to net.Listener map
...
var tempDelay time.Duration // how long to sleep on accept failure
//cycle accept client connect
for {
rawConn, err := lis.Accept()
if err != nil {
if ne, ok := err.(interface {
Temporary() bool
}); ok && ne.Temporary() { //failed retry，if err type is Temporary
...
...
}
...
return err
}
tempDelay = 0
s.serveWG.Add(1)
go func() {
s.handleRawConn(rawConn)
s.serveWG.Done()
}()
}
}handleRawConn函数主要做了如下几件事，
1.设置连接超时时间，并校验连接
2.使用当前连接创建一个http2传输结构，并加入当前所有连接缓存中
3.把传输结构传入服务处理函数serveStreams(...) 中
handleRawConn 关键代码如下：
func (s *Server) handleRawConn(rawConn net.Conn) {
if s.quit.HasFired() {
rawConn.Close()
return
}
//set connection timeout(if timeout close connection, ), default 120s
rawConn.SetDeadline(time.Now().Add(s.opts.connectionTimeout))
//if NewServer() add creds, use authenticator transport
//interface in credentials/credentials.go, such as tls creds implement in credentials/tls.go
conn, authInfo, err := s.useTransportAuthenticator(rawConn)
if err != nil {
...
rawConn.SetDeadline(time.Time{}) //reset deadline，close timer
return
}
// Finish handshaking (HTTP2), new http2 transport (implement in internal/transport/http2_server.go)
st := s.newHTTP2Transport(conn, authInfo)
if st == nil {
return
}
rawConn.SetDeadline(time.Time{}) //reset deadline
if !s.addConn(st) { //add to conns map ( cache connect)
return
}
go func() {
s.serveStreams(st)
s.removeConn(st) //remove connection
}()
}s.serverStreams(st), 通过调用st.HandleStreams嵌入handleStream() 流处理方法，及上下文函数，进行循环处理，嵌入外层函数运行的好处是服务底层不需要关心服务的具体实现，只需要在运行中调用即可。其中handleStream 的并行个数可在服务初始化时进行配置，默认只运行一个。
func (s *Server) serveStreams(st transport.ServerTransport) {
...
//cycle handle stream and trace, it is embedded to process
st.HandleStreams(func(stream *transport.Stream) {
wg.Add(1)
if s.opts.numServerWorkers > 0 { //default 0
//if has multi server workers, and get a worker by modulus
...
} else {
//default server worker handle
go func() {
defer wg.Done()
s.handleStream(st, stream, s.traceInfo(st, stream))
}()
}
}, func(ctx context.Context, method string) context.Context {
...
})
...
}HandleStreams() 是transport 接口的中的一个方法，其中http2的实现，通过循环读取framer中的frame结构，并对其进行分类处理，其中包括头帧、数据帧处理等
关键代码如下：
func (t *http2Server) HandleStreams(handle func(*Stream), traceCtx func(context.Context, string) context.Context) {
defer close(t.readerDone)
for {
t.controlBuf.throttle()
frame, err := t.framer.fr.ReadFrame()
...
switch frame := frame.(type) {
//handle header frame
case *http2.MetaHeadersFrame:
//new stream register
if t.operateHeaders(frame, handle, traceCtx) {
t.Close()
break
}
//handle data frame
case *http2.DataFrame:
t.handleData(frame)
...
}
}
}t.operateHeaders() , 处理头信息。
首先进行decode，并获取全部的头信息， 创建一个新的流处理结构体， 并丰富其字段.
如果客户端设置了超时，则在stream中设置带有相应超时时间的context
注册流信息，并调用handleStream() 方法处理
func (t *http2Server) operateHeaders(frame *http2.MetaHeadersFrame, handle func(*Stream), traceCtx func(context.Context, string) context.Context) (fatal bool) {
//get stream id
streamID := frame.Header().StreamID
state := &decodeState{
serverSide: true, //is server side
}
//decode header
//// get state.data.encoding、method、contentSubtype and so on , ./http_util.go
//if had grpc-timeout header, set state.data.timeoutSet = true
if err := state.decodeHeader(frame); err != nil {
if se, ok := status.FromError(err); ok {
t.controlBuf.put(&cleanupStream{
streamID: streamID,
rst: true, //close stream
rstCode: statusCodeConvTab[se.Code()],
onWrite: func() {},
})
}
return false
}
//create a new Stream struct
buf := newRecvBuffer()
s := &Stream{
id: streamID,
st: t,
buf: buf,
recvCompress: state.data.encoding,
method: state.data.method,
contentSubtype: state.data.contentSubtype,
}
...
//if client set timeout, set stream context with timeout
if state.data.timeoutSet {
s.ctx, s.cancel = context.WithTimeout(t.ctx, state.data.timeout)
} else {
s.ctx, s.cancel = context.WithCancel(t.ctx)
}
//peer/peer.go, set get remote addr and auth info
pr := &peer.Peer{
Addr: t.remoteAddr,
}
// Attach Auth info if there is any.
if t.authInfo != nil {
pr.AuthInfo = t.authInfo
}
//set get context by peer
s.ctx = peer.NewContext(s.ctx, pr)
// Attach the received metadata to the context.
...
//flow control, implement by userself at NewServer(...)
if t.inTapHandle != nil {
...
s.ctx, err = t.inTapHandle(s.ctx, info)
...
}
...
//control limit max active stream
if uint32(len(t.activeStreams)) >= t.maxStreams {
...
return false
}
//set max streamID now
t.maxStreamID = streamID
//register streamID
t.activeStreams[streamID] = s
...
//call traceCtx(),new trace and attach context value
s.ctx = traceCtx(s.ctx, s.method)
// if config， for statistical information，and attach context value
if t.stats != nil {
s.ctx = t.stats.TagRPC(s.ctx, &stats.RPCTagInfo{FullMethodName: s.method})
inHeader := &stats.InHeader{
FullMethod: s.method,
RemoteAddr: t.remoteAddr,
LocalAddr: t.localAddr,
Compression: s.recvCompress,
WireLength: int(frame.Header().Length),
Header: metadata.MD(state.data.mdata).Copy(),
}
t.stats.HandleRPC(s.ctx, inHeader)
}
s.ctxDone = s.ctx.Done()
s.wq = newWriteQuota(defaultWriteQuota, s.ctxDone)
//set transport reader
s.trReader = &transportReader{
//implement io.Reader interface to read from s.buf(*recvBuffer)
reader: &recvBufferReader{
ctx: s.ctx,
ctxDone: s.ctxDone,
recv: s.buf,
freeBuffer: t.bufferPool.put,
},
windowHandler: func(n int) {
t.updateWindow(s, uint32(n))
},
}
// Register the stream with loopy (implement by list link)
// async set l.estdStreams *outSteam map
t.controlBuf.put(®isterStream{
streamID: s.id,
wq: s.wq,
})
//call s.hanldeStream()
handle(s)
return false
}
func (t *http2Server) handleData(f *http2.DataFrame) {
size := f.Header().Length
...
// Select the right stream to dispatch.
s, ok := t.getStream(f)
if !ok {
return
}
if size > 0 {
...
if len(f.Data()) > 0 {
buffer := t.bufferPool.get()
buffer.Reset()
buffer.Write(f.Data())
s.write(recvMsg{buffer: buffer}) //received the stream data and append the s.buf.backlog
}
}
if f.Header().Flags.Has(http2.FlagDataEndStream) {
// Received the end of stream from the client.
//set s.state = streamReadDone
s.compareAndSwapState(streamActive, streamReadDone)
s.write(recvMsg{err: io.EOF}) //received over
}
}handleStream(), grpc服务业务层处理方法，在这个方法中会根据stream中的信息获取对应的服务和方法，
1.先从当前已注册的单元处理方法中查询对应的方法， 并通过调用processUnaryRPC 函数来处理并返回。
2.从当前已注册的流处理方法中查询对应的方法，并通过调用processStreamingRPC 函数俩处理并返回。
3.没有该方法，则进行相应的错误处理，并返回结束
//st is http2 transport, stream is from http2
func (s *Server) handleStream(t transport.ServerTransport, stream *transport.Stream, trInfo *traceInfo) {
sm := stream.Method()
...
pos := strings.LastIndex(sm, "/")
...
//get service name
service := sm[:pos]
//get method name
method := sm[pos+1:]
//process by method, process unary rpc or streaming rpc
srv, knownService := s.m[service]
if knownService { //if exist this service
if md, ok := srv.md[method]; ok { //if it is simple rpc
s.processUnaryRPC(t, stream, srv, md, trInfo)
return
}
if sd, ok := srv.sd[method]; ok { //if it is stream rpc
s.processStreamingRPC(t, stream, srv, sd, trInfo)
return
}
}
// Unknown service, or known server unknown method. handle
...
} processUnaryRPC(...)
单元处理方法，grpc 服务普通调用处理的核心函数。
1）在该函数中会进行统计信息、日志记录
2）根据头信息获取对应的压缩解压方法并对当前stream中的包内容进行解压
3）调用该rpc请求对应的用户自定义的服务方法，并获取函数返回结果
4）调用rpc 服务返回函数，把调用结果压缩返回客户端。
5）最后在写完body内容后，需要发送处理成功的状态码
//st is http2 transport, stream is from http2, srv is user defined service, md is user defined method
func (s *Server) processUnaryRPC(t transport.ServerTransport, stream *transport.Stream, srv *service, md *MethodDesc, trInfo *traceInfo) (err error) { //stream is from http2
sh := s.opts.statsHandler
//if add stats handler, implement by userself,
//./stats/handler.go
if sh != nil || trInfo != nil || channelz.IsOn() {
...
}
//binlog, default init at internal/binarylog/binarylog.go
binlog := binarylog.GetMethodLogger(stream.Method()) //internal/binarylog/binarylog.go, default get nil, unless user set method at first
//record and write method log
if binlog != nil {
...
}
// comp and cp are used for compression. decomp and dc are used for
// decompression. If comp and decomp are both set, they are the same;
// however they are kept separate to ensure that at most one of the
// compressor/decompressor variable pairs are set for use later.
var comp, decomp encoding.Compressor
var cp Compressor
var dc Decompressor
...
//rpc_util.go, get decompress data d
d, err := recvAndDecompress(&parser{r: stream}, stream, dc, s.opts.maxReceiveMessageSize, payInfo, decomp)
...
df := func(v interface{}) error {
if err := s.getCodec(stream.ContentSubtype()).Unmarshal(d, v); err != nil {
return status.Errorf(codes.Internal, "grpc: error unmarshalling request: %v", err)
}
...
//record stats and log
return nil
}
ctx := NewContextWithServerTransportStream(stream.Context(), stream)
reply, appErr := md.Handler(srv.server, ctx, df, s.opts.unaryInt) //run user defined source method func, if s.opts.unaryInt(user defined unary interceptor) not nil, run it at first
...
opts := &transport.Options{Last: true}
//response to client
if err := s.sendResponse(t, stream, reply, cp, opts, comp); err != nil {
if err == io.EOF {
// The entire stream is done (for unary RPC only).
return err
}
//handle err
...
return err
}
...
err = t.WriteStatus(stream, statusOK) //writer header, after writer data
...
return err
} recvAndDecompress(...) 函数主要是读取包内容，并根据包头信息获取当前的包内容是否加密，如果加密则解密返回，否则直接返回包内容。下面主要看看读取包内容的过程，通过实现io.Reader 接口读取包内容，采用tlv读取方式，先读取头5个字节，第一个字节是type类型加密/不加密， 后4个字节是包的长度，后面的全部字节即包的内容，核心代码如下：
func (p *parser) recvMsg(maxReceiveMessageSize int) (pf payloadFormat, msg []byte, err error) {
if _, err := p.r.Read(p.header[:]); err != nil { //read header and is 5 byte
return 0, nil, err
}
pf = payloadFormat(p.header[0]) //0 none compression, 1 compression
length := binary.BigEndian.Uint32(p.header[1:]) // get data length
...
msg = make([]byte, int(length))
if _, err := p.r.Read(msg); err != nil { //read all data/msg
if err == io.EOF {
err = io.ErrUnexpectedEOF
}
return 0, nil, err
}
return pf, msg, nil
}s.sendResponse(...) encode and compress response body, 调用底层写入方法 t.Write(...) , 至此grpc服务端接收到发送简单过程结束
func (t *http2Server) Write(s *Stream, hdr []byte, data []byte, opts *Options) error {
if !s.isHeaderSent() { // Headers haven't been written yet., transport.go
//write header
if err := t.WriteHeader(s, nil); err != nil {
...
return status.Errorf(codes.Internal, "transport: %v", err)
}
} else {
// Writing headers checks for this condition.
if s.getState() == streamDone { //when fininsh stream - fininshStrem() , or close stream - closeStream()
...
return ...
}
}
...
//append header
hdr = append(hdr, data[:emptyLen]...)
data = data[emptyLen:]
//create data frame
df := &dataFrame{
streamID: s.id,
h: hdr,
d: data,
onEachWrite: t.setResetPingStrikes,
}
...
//put data to list tail, and send a chan to wakeup consumer list head
return t.controlBuf.put(df)
}