net: ignore some errors in windows Accept

Fixes #6987

R=golang-codereviews, dvyukov
CC=golang-codereviews
https://golang.org/cl/49490043

src/pkg/net/fd_windows.go

@@ -513,12 +513,7 @@ func (fd *netFD) WriteTo(buf []byte, sa syscall.Sockaddr) (int, error) {
 	})
 }
 
-func (fd *netFD) accept(toAddr func(syscall.Sockaddr) Addr) (*netFD, error) {
-	if err := fd.readLock(); err != nil {
-		return nil, err
-	}
-	defer fd.readUnlock()
-
+func (fd *netFD) acceptOne(toAddr func(syscall.Sockaddr) Addr, rawsa []syscall.RawSockaddrAny, o *operation) (*netFD, error) {
 	// Get new socket.
 	s, err := sysSocket(fd.family, fd.sotype, 0)
 	if err != nil {
@@ -537,9 +532,7 @@ func (fd *netFD) accept(toAddr func(syscall.Sockaddr) Addr) (*netFD, error) {
 	}
 
 	// Submit accept request.
-	o := &fd.rop
 	o.handle = s
-	var rawsa [2]syscall.RawSockaddrAny
 	o.rsan = int32(unsafe.Sizeof(rawsa[0]))
 	_, err = rsrv.ExecIO(o, "AcceptEx", func(o *operation) error {
 		return syscall.AcceptEx(o.fd.sysfd, o.handle, (*byte)(unsafe.Pointer(&rawsa[0])), 0, uint32(o.rsan), uint32(o.rsan), &o.qty, &o.o)
@@ -556,6 +549,45 @@ func (fd *netFD) accept(toAddr func(syscall.Sockaddr) Addr) (*netFD, error) {
 		return nil, &OpError{"Setsockopt", fd.net, fd.laddr, err}
 	}
 
+	return netfd, nil
+}
+
+func (fd *netFD) accept(toAddr func(syscall.Sockaddr) Addr) (*netFD, error) {
+	if err := fd.readLock(); err != nil {
+		return nil, err
+	}
+	defer fd.readUnlock()
+
+	o := &fd.rop
+	var netfd *netFD
+	var err error
+	var rawsa [2]syscall.RawSockaddrAny
+	for {
+		netfd, err = fd.acceptOne(toAddr, rawsa[:], o)
+		if err == nil {
+			break
+		}
+		// Sometimes we see WSAECONNRESET and ERROR_NETNAME_DELETED is
+		// returned here. These happen if connection reset is received
+		// before AcceptEx could complete. These errors relate to new
+		// connection, not to AcceptEx, so ignore broken connection and
+		// try AcceptEx again for more connections.
+		operr, ok := err.(*OpError)
+		if !ok {
+			return nil, err
+		}
+		errno, ok := operr.Err.(syscall.Errno)
+		if !ok {
+			return nil, err
+		}
+		switch errno {
+		case syscall.ERROR_NETNAME_DELETED, syscall.WSAECONNRESET:
+			// ignore these and try again
+		default:
+			return nil, err
+		}
+	}
+
 	// Get local and peer addr out of AcceptEx buffer.
 	var lrsa, rrsa *syscall.RawSockaddrAny
 	var llen, rlen int32

src/pkg/net/net_windows_test.go

@@ -0,0 +1,146 @@
+// Copyright 2014 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package net
+
+import (
+	"bufio"
+	"fmt"
+	"io"
+	"os"
+	"os/exec"
+	"syscall"
+	"testing"
+	"time"
+)
+
+func TestAcceptIgnoreSomeErrors(t *testing.T) {
+	recv := func(ln Listener) (string, error) {
+		c, err := ln.Accept()
+		if err != nil {
+			// Display windows errno in error message.
+			operr, ok := err.(*OpError)
+			if !ok {
+				return "", err
+			}
+			errno, ok := operr.Err.(syscall.Errno)
+			if !ok {
+				return "", err
+			}
+			return "", fmt.Errorf("%v (windows errno=%d)", err, errno)
+		}
+		defer c.Close()
+
+		b := make([]byte, 100)
+		n, err := c.Read(b)
+		if err != nil && err != io.EOF {
+			return "", err
+		}
+		return string(b[:n]), nil
+	}
+
+	send := func(addr string, data string) error {
+		c, err := Dial("tcp", addr)
+		if err != nil {
+			return err
+		}
+		defer c.Close()
+
+		b := []byte(data)
+		n, err := c.Write(b)
+		if err != nil {
+			return err
+		}
+		if n != len(b) {
+			return fmt.Errorf(`Only %d chars of string "%s" sent`, n, data)
+		}
+		return nil
+	}
+
+	if envaddr := os.Getenv("GOTEST_DIAL_ADDR"); envaddr != "" {
+		// In child process.
+		c, err := Dial("tcp", envaddr)
+		if err != nil {
+			t.Fatalf("Dial failed: %v", err)
+		}
+		fmt.Printf("sleeping\n")
+		time.Sleep(time.Minute) // process will be killed here
+		c.Close()
+	}
+
+	ln, err := Listen("tcp", "127.0.0.1:0")
+	if err != nil {
+		t.Fatalf("Listen failed: %v", err)
+	}
+	defer ln.Close()
+
+	// Start child process that connects to our listener.
+	cmd := exec.Command(os.Args[0], "-test.run=TestAcceptIgnoreSomeErrors")
+	cmd.Env = append(os.Environ(), "GOTEST_DIAL_ADDR="+ln.Addr().String())
+	stdout, err := cmd.StdoutPipe()
+	if err != nil {
+		t.Fatalf("cmd.StdoutPipe failed: %v", err)
+	}
+	err = cmd.Start()
+	if err != nil {
+		t.Fatalf("cmd.Start failed: %v\n%s\n", err)
+	}
+	outReader := bufio.NewReader(stdout)
+	for {
+		s, err := outReader.ReadString('\n')
+		if err != nil {
+			t.Fatalf("reading stdout failed: %v", err)
+		}
+		if s == "sleeping\n" {
+			break
+		}
+	}
+	defer cmd.Wait() // ignore error - we know it is getting killed
+
+	const alittle = 100 * time.Millisecond
+	time.Sleep(alittle)
+	cmd.Process.Kill() // the only way to trigger the errors
+	time.Sleep(alittle)
+
+	// Send second connection data (with delay in a separate goroutine).
+	result := make(chan error)
+	go func() {
+		time.Sleep(alittle)
+		err = send(ln.Addr().String(), "abc")
+		if err != nil {
+			result <- err
+		}
+		result <- nil
+	}()
+	defer func() {
+		err := <-result
+		if err != nil {
+			t.Fatalf("send failed: %v", err)
+		}
+	}()
+
+	// Receive first or second connection.
+	s, err := recv(ln)
+	if err != nil {
+		t.Fatalf("recv failed: %v", err)
+	}
+	switch s {
+	case "":
+		// First connection data is received, lets get second connection data.
+	case "abc":
+		// First connection is lost forever, but that is ok.
+		return
+	default:
+		t.Fatalf(`"%s" received from recv, but "" or "abc" expected`, s)
+	}
+
+	// Get second connection data.
+	s, err = recv(ln)
+	if err != nil {
+		t.Fatalf("recv failed: %v", err)
+	}
+	if s != "abc" {
+		t.Fatalf(`"%s" received from recv, but "abc" expected`, s)
+	}
+}

src/pkg/syscall/ztypes_windows.go

@@ -11,6 +11,7 @@ const (
 	ERROR_ACCESS_DENIED       Errno = 5
 	ERROR_NO_MORE_FILES       Errno = 18
 	ERROR_HANDLE_EOF          Errno = 38
+	ERROR_NETNAME_DELETED     Errno = 64
 	ERROR_FILE_EXISTS         Errno = 80
 	ERROR_BROKEN_PIPE         Errno = 109
 	ERROR_BUFFER_OVERFLOW     Errno = 111
@@ -23,6 +24,7 @@ const (
 	ERROR_IO_PENDING          Errno = 997
 	ERROR_NOT_FOUND           Errno = 1168
 	WSAEACCES                 Errno = 10013
+	WSAECONNRESET             Errno = 10054
 )
 
 const (