netstack/fdbased: add generic segmentation offload (GSO) support

The linux packet socket can handle GSO packets, so we can segment packets to
64K instead of the MTU which is usually 1500.

Here are numbers for the nginx-1m test:
runsc:		579330.01 [Kbytes/sec] received
runsc-gso:	1794121.66 [Kbytes/sec] received
runc:		2122139.06 [Kbytes/sec] received

and for tcp_benchmark:

$ tcp_benchmark  --duration 15   --ideal
[  4]  0.0-15.0 sec  86647 MBytes  48456 Mbits/sec

$ tcp_benchmark --client --duration 15   --ideal
[  4]  0.0-15.0 sec  2173 MBytes  1214 Mbits/sec

$ tcp_benchmark --client --duration 15   --ideal --gso 65536
[  4]  0.0-15.0 sec  19357 MBytes  10825 Mbits/sec

PiperOrigin-RevId: 240809103
Change-Id: I2637f104db28b5d4c64e1e766c610162a195775a

2 files changed, 6 insertions, 6 deletions

diff --git a/pkg/tcpip/link/waitable/waitable.go b/pkg/tcpip/link/waitable/waitable.go
index 39217e49c..bd9f9845b 100644
--- a/pkg/tcpip/link/waitable/waitable.go
+++ b/pkg/tcpip/link/waitable/waitable.go
@@ -100,12 +100,12 @@ func (e *Endpoint) LinkAddress() tcpip.LinkAddress {
 // WritePacket implements stack.LinkEndpoint.WritePacket. It is called by
 // higher-level protocols to write packets. It only forwards packets to the
 // lower endpoint if Wait or WaitWrite haven't been called.
-func (e *Endpoint) WritePacket(r *stack.Route, hdr buffer.Prependable, payload buffer.VectorisedView, protocol tcpip.NetworkProtocolNumber) *tcpip.Error {
+func (e *Endpoint) WritePacket(r *stack.Route, gso *stack.GSO, hdr buffer.Prependable, payload buffer.VectorisedView, protocol tcpip.NetworkProtocolNumber) *tcpip.Error {
 	if !e.writeGate.Enter() {
 		return nil
 	}
 
-	err := e.lower.WritePacket(r, hdr, payload, protocol)
+	err := e.lower.WritePacket(r, gso, hdr, payload, protocol)
 	e.writeGate.Leave()
 	return err
 }
diff --git a/pkg/tcpip/link/waitable/waitable_test.go b/pkg/tcpip/link/waitable/waitable_test.go
index 6c57e597a..a2df6be95 100644
--- a/pkg/tcpip/link/waitable/waitable_test.go
+++ b/pkg/tcpip/link/waitable/waitable_test.go
@@ -65,7 +65,7 @@ func (e *countedEndpoint) LinkAddress() tcpip.LinkAddress {
 	return e.linkAddr
 }
 
-func (e *countedEndpoint) WritePacket(r *stack.Route, hdr buffer.Prependable, payload buffer.VectorisedView, protocol tcpip.NetworkProtocolNumber) *tcpip.Error {
+func (e *countedEndpoint) WritePacket(r *stack.Route, _ *stack.GSO, hdr buffer.Prependable, payload buffer.VectorisedView, protocol tcpip.NetworkProtocolNumber) *tcpip.Error {
 	e.writeCount++
 	return nil
 }
@@ -75,21 +75,21 @@ func TestWaitWrite(t *testing.T) {
 	_, wep := New(stack.RegisterLinkEndpoint(ep))
 
 	// Write and check that it goes through.
-	wep.WritePacket(nil, buffer.Prependable{}, buffer.VectorisedView{}, 0)
+	wep.WritePacket(nil, nil /* gso */, buffer.Prependable{}, buffer.VectorisedView{}, 0)
 	if want := 1; ep.writeCount != want {
 		t.Fatalf("Unexpected writeCount: got=%v, want=%v", ep.writeCount, want)
 	}
 
 	// Wait on dispatches, then try to write. It must go through.
 	wep.WaitDispatch()
-	wep.WritePacket(nil, buffer.Prependable{}, buffer.VectorisedView{}, 0)
+	wep.WritePacket(nil, nil /* gso */, buffer.Prependable{}, buffer.VectorisedView{}, 0)
 	if want := 2; ep.writeCount != want {
 		t.Fatalf("Unexpected writeCount: got=%v, want=%v", ep.writeCount, want)
 	}
 
 	// Wait on writes, then try to write. It must not go through.
 	wep.WaitWrite()
-	wep.WritePacket(nil, buffer.Prependable{}, buffer.VectorisedView{}, 0)
+	wep.WritePacket(nil, nil /* gso */, buffer.Prependable{}, buffer.VectorisedView{}, 0)
 	if want := 2; ep.writeCount != want {
 		t.Fatalf("Unexpected writeCount: got=%v, want=%v", ep.writeCount, want)
 	}