Fix data race in executor with enabling the limit mode (#474)

The root cause is we write the limitModeFuncsRunning(a wait group)
in the new goroutine, so the Go data race detector will think them
have the data race due to reading and writing in different routines.

executor.go

@@ -96,7 +96,6 @@ func (jf *jobFunction) singletonRunner() {
 }
 
 func (e *executor) limitModeRunner() {
-	e.limitModeFuncWg.Add(1)
 	for {
 		select {
 		case <-e.ctx.Done():
@@ -169,6 +168,7 @@ func (e *executor) run() {
 				if countRunning < int64(e.limitModeMaxRunningJobs) {
 					diff := int64(e.limitModeMaxRunningJobs) - countRunning
 					for i := int64(0); i < diff; i++ {
+						e.limitModeFuncWg.Add(1)
 						go e.limitModeRunner()
 						e.limitModeFuncsRunning.Add(1)
 					}