Create new controller for all node labeling operations

api/nvidia/v1/clusterpolicy_types.go

@@ -2205,6 +2205,14 @@ func (d *DriverSpec) IsVGPULicensingEnabled() bool {
 	return d.LicensingConfig.ConfigMapName != "" || d.LicensingConfig.SecretName != ""
 }
 
+// IsAutoUpgradeEnabled returns true if auto upgrade is enabled
+func (d *DriverSpec) IsAutoUpgradeEnabled() bool {
+	if d.UpgradePolicy == nil {
+		return false
+	}
+	return d.UpgradePolicy.AutoUpgrade
+}
+
 // IsEnabled returns true if device-plugin is enabled(default) through gpu-operator
 func (p *DevicePluginSpec) IsEnabled() bool {
 	if p.Enabled == nil {

cmd/gpu-operator/main.go

@@ -216,6 +216,16 @@ func main() {
 		setupLog.Error(err, "unable to create controller", "controller", "NVIDIADriver")
 		os.Exit(1)
 	}
+
+	if err = (&controllers.NodeLabelingReconciler{
+		Namespace: operatorNamespace,
+		Client:    mgr.GetClient(),
+		Scheme:    mgr.GetScheme(),
+		Log:       ctrl.Log.WithName("controllers").WithName("NodeLabeling"),
+	}).SetupWithManager(ctx, mgr); err != nil {
+		setupLog.Error(err, "unable to create controller", "controller", "NodeLabeling")
+		os.Exit(1)
+	}
 	// +kubebuilder:scaffold:builder
 	if err := mgr.AddHealthzCheck("health", healthz.Ping); err != nil {
 		setupLog.Error(err, "unable to set up health check")

controllers/clusterpolicy_controller.go

@@ -280,9 +280,8 @@ func addWatchNewGPUNode(r *ClusterPolicyReconciler, c controller.Controller, mgr
 			oldLabels := e.ObjectOld.GetLabels()
 			nodeName := e.ObjectNew.GetName()
 
-			gpuCommonLabelMissing := hasGPULabels(newLabels) && !hasCommonGPULabel(newLabels)
-			gpuCommonLabelOutdated := !hasGPULabels(newLabels) && hasCommonGPULabel(newLabels)
-			migManagerLabelMissing := hasMIGCapableGPU(newLabels) && !hasMIGManagerLabel(newLabels)
+			// Trigger when NodeLabelingReconciler sets gpu.present=true on a new GPU node.
+			gpuCommonLabelAdded := !hasCommonGPULabel(oldLabels) && hasCommonGPULabel(newLabels)
 			commonOperandsLabelChanged := hasOperandsDisabled(oldLabels) != hasOperandsDisabled(newLabels)
 
 			oldGPUWorkloadConfig, _ := getWorkloadConfig(oldLabels, true)
@@ -293,19 +292,15 @@ func addWatchNewGPUNode(r *ClusterPolicyReconciler, c controller.Controller, mgr
 			newOSTreeLabel := newLabels[nfdOSTreeVersionLabelKey]
 			osTreeLabelChanged := oldOSTreeLabel != newOSTreeLabel
 
-			needsUpdate := gpuCommonLabelMissing ||
-				gpuCommonLabelOutdated ||
-				migManagerLabelMissing ||
+			needsUpdate := gpuCommonLabelAdded ||
 				commonOperandsLabelChanged ||
 				gpuWorkloadConfigLabelChanged ||
 				osTreeLabelChanged
 
 			if needsUpdate {
 				r.Log.Info("Node needs an update",
 					"name", nodeName,
-					"gpuCommonLabelMissing", gpuCommonLabelMissing,
-					"gpuCommonLabelOutdated", gpuCommonLabelOutdated,
-					"migManagerLabelMissing", migManagerLabelMissing,
+					"gpuCommonLabelAdded", gpuCommonLabelAdded,
 					"commonOperandsLabelChanged", commonOperandsLabelChanged,
 					"gpuWorkloadConfigLabelChanged", gpuWorkloadConfigLabelChanged,
 					"osTreeLabelChanged", osTreeLabelChanged,