Make tests that wait for concurrent work wait on the event or a condition with a long deadline (issue 1814)

A test that slept a fixed time and then asserted what a goroutine, process or
watcher did failed under the checker's load. Each now waits for its positive
evidence (a started baseline, a call counted, a state reached) within 30 s, and
the concurrency-limit tests meet at a barrier instead of overlapping by sleep.

Claude-Session: https://claude.ai/code/session_01NcbgfJ9Gum47zkFKWRaovL

modules/zfs/cmd/zfs-tools/review_test.go

@@ -26,11 +26,11 @@ func TestATimedOutCommandIsEndedWithItsChildren(t *testing.T) {
 	KillGrace = 300 * time.Millisecond
 	defer func() { KillGrace = old }()
 	pidFile := filepath.Join(t.TempDir(), "child")
-	ctx, cancel := context.WithTimeout(context.Background(), 300*time.Millisecond)
+	ctx, cancel := context.WithTimeout(context.Background(), 3*time.Second) // long enough for a loaded machine to start the shell and write the pid
 	defer cancel()
 	start := time.Now()
 	r := ExecRunner(ctx, "sh", "-c", "sleep 30 & echo $! > "+pidFile+"; wait")
-	if took := time.Since(start); took > 5*time.Second {
+	if took := time.Since(start); took > 30*time.Second {
 		t.Fatalf("the call waited %s for a child holding its output", took)
 	}
 	if !r.TimedOut {
@@ -41,7 +41,9 @@ func TestATimedOutCommandIsEndedWithItsChildren(t *testing.T) {
 		t.Fatal(err)
 	}
 	pid, _ := strconv.Atoi(strings.TrimSpace(string(b)))
-	time.Sleep(100 * time.Millisecond)
+	// the kill is sent before the answer; give the system up to 30 s to take the child away
+	for end := time.Now().Add(30 * time.Second); time.Now().Before(end) && syscall.Kill(pid, 0) == nil; time.Sleep(10 * time.Millisecond) {
+	}
 	if err := syscall.Kill(pid, 0); err == nil {
 		syscall.Kill(pid, syscall.SIGKILL)
 		t.Errorf("the child %d outlived the timeout", pid)