runtime: only poll network from one P at a time in findRunnable

This change reintroduces CL 564197. It was reverted due to a failing
benchmark. That failure has been resolved.

For #65064

Change-Id: Ic88841d2bc24c2717ad324873f0f52699f21dc66
Reviewed-on: https://go-review.googlesource.com/c/go/+/669235
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Mauri de Souza Meneguzzo <mauri870@gmail.com>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
Reviewed-by: Michael Pratt <mpratt@google.com>

src/runtime/proc.go

@@ -3396,8 +3396,12 @@ top:
 	// blocked thread (e.g. it has already returned from netpoll, but does
 	// not set lastpoll yet), this thread will do blocking netpoll below
 	// anyway.
-	if netpollinited() && netpollAnyWaiters() && sched.lastpoll.Load() != 0 {
-		if list, delta := netpoll(0); !list.empty() { // non-blocking
+	// We only poll from one thread at a time to avoid kernel contention
+	// on machines with many cores.
+	if netpollinited() && netpollAnyWaiters() && sched.lastpoll.Load() != 0 && sched.pollingNet.Swap(1) == 0 {
+		list, delta := netpoll(0)
+		sched.pollingNet.Store(0)
+		if !list.empty() { // non-blocking
 			gp := list.pop()
 			injectglist(&list)
 			netpollAdjustWaiters(delta)

src/runtime/runtime2.go

@@ -760,9 +760,10 @@ type p struct {
 }
 
 type schedt struct {
-	goidgen   atomic.Uint64
-	lastpoll  atomic.Int64 // time of last network poll, 0 if currently polling
-	pollUntil atomic.Int64 // time to which current poll is sleeping
+	goidgen    atomic.Uint64
+	lastpoll   atomic.Int64 // time of last network poll, 0 if currently polling
+	pollUntil  atomic.Int64 // time to which current poll is sleeping
+	pollingNet atomic.Int32 // 1 if some P doing non-blocking network poll
 
 	lock mutex