Block for link address resolution

Previously, if address resolution for UDP or Ping sockets required sending
packets using Write in Transport layer, Resolve would return ErrWouldBlock
and Write would return ErrNoLinkAddress. Meanwhile startAddressResolution
would run in background. Further calls to Write using same address would also
return ErrNoLinkAddress until resolution has been completed successfully.

Since Write is not allowed to block and System Calls need to be
interruptible in System Call layer, the caller to Write is responsible for
blocking upon return of ErrWouldBlock.

Now, when startAddressResolution is called a notification channel for
the completion of the address resolution is returned.
The channel will traverse up to the calling function of Write as well as
ErrNoLinkAddress. Once address resolution is complete (success or not) the
channel is closed. The caller would call Write again to send packets and
check if address resolution was compeleted successfully or not.

Fixes google/gvisor#5

Change-Id: Idafaf31982bee1915ca084da39ae7bd468cebd93
PiperOrigin-RevId: 214962200

1 files changed, 8 insertions, 8 deletions

diff --git a/pkg/tcpip/transport/tcp/endpoint.go b/pkg/tcpip/transport/tcp/endpoint.go
index e82e25233..707d6be96 100644
--- a/pkg/tcpip/transport/tcp/endpoint.go
+++ b/pkg/tcpip/transport/tcp/endpoint.go
@@ -492,7 +492,7 @@ func (e *endpoint) readLocked() (buffer.View, *tcpip.Error) {
 }
 
 // Write writes data to the endpoint's peer.
-func (e *endpoint) Write(p tcpip.Payload, opts tcpip.WriteOptions) (uintptr, *tcpip.Error) {
+func (e *endpoint) Write(p tcpip.Payload, opts tcpip.WriteOptions) (uintptr, <-chan struct{}, *tcpip.Error) {
 	// Linux completely ignores any address passed to sendto(2) for TCP sockets
 	// (without the MSG_FASTOPEN flag). Corking is unimplemented, so opts.More
 	// and opts.EndOfRecord are also ignored.
@@ -504,15 +504,15 @@ func (e *endpoint) Write(p tcpip.Payload, opts tcpip.WriteOptions) (uintptr, *tc
 	if e.state != stateConnected {
 		switch e.state {
 		case stateError:
-			return 0, e.hardError
+			return 0, nil, e.hardError
 		default:
-			return 0, tcpip.ErrClosedForSend
+			return 0, nil, tcpip.ErrClosedForSend
 		}
 	}
 
 	// Nothing to do if the buffer is empty.
 	if p.Size() == 0 {
-		return 0, nil
+		return 0, nil, nil
 	}
 
 	e.sndBufMu.Lock()
@@ -520,20 +520,20 @@ func (e *endpoint) Write(p tcpip.Payload, opts tcpip.WriteOptions) (uintptr, *tc
 	// Check if the connection has already been closed for sends.
 	if e.sndClosed {
 		e.sndBufMu.Unlock()
-		return 0, tcpip.ErrClosedForSend
+		return 0, nil, tcpip.ErrClosedForSend
 	}
 
 	// Check against the limit.
 	avail := e.sndBufSize - e.sndBufUsed
 	if avail <= 0 {
 		e.sndBufMu.Unlock()
-		return 0, tcpip.ErrWouldBlock
+		return 0, nil, tcpip.ErrWouldBlock
 	}
 
 	v, perr := p.Get(avail)
 	if perr != nil {
 		e.sndBufMu.Unlock()
-		return 0, perr
+		return 0, nil, perr
 	}
 
 	var err *tcpip.Error
@@ -558,7 +558,7 @@ func (e *endpoint) Write(p tcpip.Payload, opts tcpip.WriteOptions) (uintptr, *tc
 		// Let the protocol goroutine do the work.
 		e.sndWaker.Assert()
 	}
-	return uintptr(l), err
+	return uintptr(l), nil, err
 }
 
 // Peek reads data without consuming it from the endpoint.