Eliminate one allocation per send/recv for non-flipcall transport.

Ported from https://github.com/hugelgupf/p9/pull/44.

name               old time/op    new time/op    delta
SendRecvLegacy-6     61.5µs ± 6%    60.1µs ±11%     ~     (p=0.063 n=9+9)
SendRecv-6           40.7µs ± 2%    39.8µs ± 5%   -2.27%  (p=0.035 n=10+10)

name               old alloc/op   new alloc/op   delta
SendRecvLegacy-6       769B ± 0%      705B ± 0%   -8.37%  (p=0.000 n=8+10)
SendRecv-6             320B ± 0%      256B ± 0%  -20.00%  (p=0.000 n=10+10)

name               old allocs/op  new allocs/op  delta
SendRecvLegacy-6       25.0 ± 0%      23.0 ± 0%   -8.00%  (p=0.000 n=10+10)
SendRecv-6             14.0 ± 0%      12.0 ± 0%  -14.29%  (p=0.000 n=10+10)

PiperOrigin-RevId: 326127979

pkg/p9/client_test.go

@@ -62,6 +62,8 @@ func TestVersion(t *testing.T) {
 }
 
 func benchmarkSendRecv(b *testing.B, fn func(c *Client) func(message, message) error) {
+	b.ReportAllocs()
+
 	// See above.
 	serverSocket, clientSocket, err := unet.SocketPair(false)
 	if err != nil {

pkg/p9/transport.go

@@ -66,14 +66,17 @@ const (
 var dataPool = sync.Pool{
 	New: func() interface{} {
 		// These buffers are used for decoding without a payload.
-		return make([]byte, initialBufferLength)
+		// We need to return a pointer to avoid unnecessary allocations
+		// (see https://staticcheck.io/docs/checks#SA6002).
+		b := make([]byte, initialBufferLength)
+		return &b
 	},
 }
 
 // send sends the given message over the socket.
 func send(s *unet.Socket, tag Tag, m message) error {
-	data := dataPool.Get().([]byte)
-	dataBuf := buffer{data: data[:0]}
+	data := dataPool.Get().(*[]byte)
+	dataBuf := buffer{data: (*data)[:0]}
 
 	if log.IsLogging(log.Debug) {
 		log.Debugf("send [FD %d] [Tag %06d] %s", s.FD(), tag, m.String())
@@ -141,7 +144,7 @@ func send(s *unet.Socket, tag Tag, m message) error {
 	}
 
 	// All set.
-	dataPool.Put(dataBuf.data)
+	dataPool.Put(&dataBuf.data)
 	return nil
 }
 
@@ -227,12 +230,29 @@ func recv(s *unet.Socket, msize uint32, lookup lookupTagAndType) (Tag, message,
 
 	// Not yet initialized.
 	var dataBuf buffer
+	var vecs [][]byte
+
+	appendBuffer := func(size int) *[]byte {
+		// Pull a data buffer from the pool.
+		datap := dataPool.Get().(*[]byte)
+		data := *datap
+		if size > len(data) {
+			// Create a larger data buffer.
+			data = make([]byte, size)
+			datap = &data
+		} else {
+			// Limit the data buffer.
+			data = data[:size]
+		}
+		dataBuf = buffer{data: data}
+		vecs = append(vecs, data)
+		return datap
+	}
 
 	// Read the rest of the payload.
 	//
 	// This requires some special care to ensure that the vectors all line
 	// up the way they should. We do this to minimize copying data around.
-	var vecs [][]byte
 	if payloader, ok := m.(payloader); ok {
 		fixedSize := payloader.FixedSize()
 
@@ -246,22 +266,8 @@ func recv(s *unet.Socket, msize uint32, lookup lookupTagAndType) (Tag, message,
 		}
 
 		if fixedSize != 0 {
-			// Pull a data buffer from the pool.
-			data := dataPool.Get().([]byte)
-			if int(fixedSize) > len(data) {
-				// Create a larger data buffer, ensuring
-				// sufficient capicity for the message.
-				data = make([]byte, fixedSize)
-				defer dataPool.Put(data)
-				dataBuf = buffer{data: data}
-				vecs = append(vecs, data)
-			} else {
-				// Limit the data buffer, and make sure it
-				// gets filled before the payload buffer.
-				defer dataPool.Put(data)
-				dataBuf = buffer{data: data[:fixedSize]}
-				vecs = append(vecs, data[:fixedSize])
-			}
+			datap := appendBuffer(int(fixedSize))
+			defer dataPool.Put(datap)
 		}
 
 		// Include the payload.
@@ -274,20 +280,8 @@ func recv(s *unet.Socket, msize uint32, lookup lookupTagAndType) (Tag, message,
 			vecs = append(vecs, p)
 		}
 	} else if remaining != 0 {
-		// Pull a data buffer from the pool.
-		data := dataPool.Get().([]byte)
-		if int(remaining) > len(data) {
-			// Create a larger data buffer.
-			data = make([]byte, remaining)
-			defer dataPool.Put(data)
-			dataBuf = buffer{data: data}
-			vecs = append(vecs, data)
-		} else {
-			// Limit the data buffer.
-			defer dataPool.Put(data)
-			dataBuf = buffer{data: data[:remaining]}
-			vecs = append(vecs, data[:remaining])
-		}
+		datap := appendBuffer(int(remaining))
+		defer dataPool.Put(datap)
 	}
 
 	if len(vecs) > 0 {

pkg/p9/transport_test.go

@@ -182,6 +182,8 @@ func TestSendClosed(t *testing.T) {
 }
 
 func BenchmarkSendRecv(b *testing.B) {
+	b.ReportAllocs()
+
 	server, client, err := unet.SocketPair(false)
 	if err != nil {
 		b.Fatalf("socketpair got err %v expected nil", err)