Add kubelet configuration flag 'topology-manager-scope'

add kubelet config option.
* --topology-manager-scope=[ container | pod ]
* default=container

Signed-off-by: Krzysztof Wiatrzyk <k.wiatrzyk@samsung.com>

pkg/kubelet/apis/config/fuzzer/fuzzer.go

@@ -70,6 +70,7 @@ func Funcs(codecs runtimeserializer.CodecFactory) []interface{} {
 			obj.CPUManagerReconcilePeriod = obj.NodeStatusUpdateFrequency
 			obj.NodeStatusMaxImages = 50
 			obj.TopologyManagerPolicy = kubeletconfig.NoneTopologyManagerPolicy
+			obj.TopologyManagerScope = kubeletconfig.ContainerTopologyManagerScope
 			obj.QOSReserved = map[string]string{
 				"memory": "50%",
 			}

pkg/kubelet/apis/config/helpers_test.go

@@ -152,6 +152,7 @@ var (
 		"CPUManagerPolicy",
 		"CPUManagerReconcilePeriod.Duration",
 		"TopologyManagerPolicy",
+		"TopologyManagerScope",
 		"QOSReserved[*]",
 		"CgroupDriver",
 		"CgroupRoot",

pkg/kubelet/apis/config/scheme/testdata/KubeletConfiguration/after/v1beta1.yaml

@@ -70,5 +70,6 @@ serializeImagePulls: true
 streamingConnectionIdleTimeout: 4h0m0s
 syncFrequency: 1m0s
 topologyManagerPolicy: none
+topologyManagerScope: container
 volumePluginDir: /usr/libexec/kubernetes/kubelet-plugins/volume/exec/
 volumeStatsAggPeriod: 1m0s

pkg/kubelet/apis/config/scheme/testdata/KubeletConfiguration/roundtrip/default/v1beta1.yaml

@@ -70,5 +70,6 @@ serializeImagePulls: true
 streamingConnectionIdleTimeout: 4h0m0s
 syncFrequency: 1m0s
 topologyManagerPolicy: none
+topologyManagerScope: container
 volumePluginDir: /usr/libexec/kubernetes/kubelet-plugins/volume/exec/
 volumeStatsAggPeriod: 1m0s

pkg/kubelet/apis/config/types.go

@@ -61,12 +61,18 @@ const (
 	// BestEffortTopologyManagerPolicy is a mode in which kubelet will favour
 	// pods with NUMA alignment of CPU and device resources.
 	BestEffortTopologyManagerPolicy = "best-effort"
-	// NoneTopologyManager Policy is a mode in which kubelet has no knowledge
+	// NoneTopologyManagerPolicy is a mode in which kubelet has no knowledge
 	// of NUMA alignment of a pod's CPU and device resources.
 	NoneTopologyManagerPolicy = "none"
-	// SingleNumaNodeTopologyManager Policy iis a mode in which kubelet only allows
+	// SingleNumaNodeTopologyManagerPolicy is a mode in which kubelet only allows
 	// pods with a single NUMA alignment of CPU and device resources.
-	SingleNumaNodeTopologyManager = "single-numa-node"
+	SingleNumaNodeTopologyManagerPolicy = "single-numa-node"
+	// ContainerTopologyManagerScope represents that
+	// topology policy is applied on a per-container basis.
+	ContainerTopologyManagerScope = "container"
+	// PodTopologyManagerScope represents that
+	// topology policy is applied on a per-pod basis.
+	PodTopologyManagerScope = "pod"
 )
 
 // +k8s:deepcopy-gen:interfaces=k8s.io/apimachinery/pkg/runtime.Object
@@ -221,6 +227,12 @@ type KubeletConfiguration struct {
 	// TopologyManagerPolicy is the name of the policy to use.
 	// Policies other than "none" require the TopologyManager feature gate to be enabled.
 	TopologyManagerPolicy string
+	// TopologyManagerScope represents the scope of topology hint generation
+	// that topology manager requests and hint providers generate.
+	// "pod" scope requires the TopologyManager feature gate to be enabled.
+	// Default: "container"
+	// +optional
+	TopologyManagerScope string
 	// Map of QoS resource reservation percentages (memory only for now).
 	// Requires the QOSReserved feature gate to be enabled.
 	QOSReserved map[string]string

pkg/kubelet/apis/config/v1beta1/defaults.go

@@ -157,6 +157,9 @@ func SetDefaults_KubeletConfiguration(obj *kubeletconfigv1beta1.KubeletConfigura
 	if obj.TopologyManagerPolicy == "" {
 		obj.TopologyManagerPolicy = kubeletconfigv1beta1.NoneTopologyManagerPolicy
 	}
+	if obj.TopologyManagerScope == "" {
+		obj.TopologyManagerScope = kubeletconfigv1beta1.ContainerTopologyManagerScope
+	}
 	if obj.RuntimeRequestTimeout == zeroDuration {
 		obj.RuntimeRequestTimeout = metav1.Duration{Duration: 2 * time.Minute}
 	}

pkg/kubelet/apis/config/v1beta1/zz_generated.conversion.go

@@ -275,6 +275,7 @@ func autoConvert_v1beta1_KubeletConfiguration_To_config_KubeletConfiguration(in
 	out.CPUManagerPolicy = in.CPUManagerPolicy
 	out.CPUManagerReconcilePeriod = in.CPUManagerReconcilePeriod
 	out.TopologyManagerPolicy = in.TopologyManagerPolicy
+	out.TopologyManagerScope = in.TopologyManagerScope
 	out.QOSReserved = *(*map[string]string)(unsafe.Pointer(&in.QOSReserved))
 	out.RuntimeRequestTimeout = in.RuntimeRequestTimeout
 	out.HairpinMode = in.HairpinMode
@@ -427,6 +428,7 @@ func autoConvert_config_KubeletConfiguration_To_v1beta1_KubeletConfiguration(in
 	out.CPUManagerPolicy = in.CPUManagerPolicy
 	out.CPUManagerReconcilePeriod = in.CPUManagerReconcilePeriod
 	out.TopologyManagerPolicy = in.TopologyManagerPolicy
+	out.TopologyManagerScope = in.TopologyManagerScope
 	out.QOSReserved = *(*map[string]string)(unsafe.Pointer(&in.QOSReserved))
 	out.RuntimeRequestTimeout = in.RuntimeRequestTimeout
 	out.HairpinMode = in.HairpinMode

pkg/kubelet/apis/config/validation/validation.go

@@ -125,6 +125,9 @@ func ValidateKubeletConfiguration(kc *kubeletconfig.KubeletConfiguration) error
 	if kc.TopologyManagerPolicy != kubeletconfig.NoneTopologyManagerPolicy && !localFeatureGate.Enabled(features.TopologyManager) {
 		allErrors = append(allErrors, fmt.Errorf("invalid configuration: topologyManager %v requires feature gate TopologyManager", kc.TopologyManagerPolicy))
 	}
+	if kc.TopologyManagerScope != kubeletconfig.ContainerTopologyManagerScope && !localFeatureGate.Enabled(features.TopologyManager) {
+		allErrors = append(allErrors, fmt.Errorf("invalid configuration: TopologyManagerScope %v requires feature gate TopologyManager", kc.TopologyManagerScope))
+	}
 	for _, val := range kc.EnforceNodeAllocatable {
 		switch val {
 		case kubetypes.NodeAllocatableEnforcementKey: