[Fix] amend neceassary changes and helm e2e test coverage

yansun1996 · yansun1996 · commit ec2fc0db1fb5 · 2026-03-11T23:38:14.000Z
Signed-off-by: yansun1996 &lt;Yan.Sun3@amd.com&gt;
diff --git a/api/v1alpha1/zz_generated.deepcopy.go b/api/v1alpha1/zz_generated.deepcopy.go
diff --git a/bundle/manifests/amd-gpu-operator.clusterserviceversion.yaml b/bundle/manifests/amd-gpu-operator.clusterserviceversion.yaml
@@ -210,6 +210,12 @@ spec:
         path: devicePlugin.enableNodeLabeller
         x-descriptors:
         - urn:alm:descriptor:com.amd.deviceconfigs:enableNodeLabeller
+      - description: HostNetwork specifies whether to use host network namespace for
+          the device plugin DaemonSet pods.
+        displayName: HostNetwork
+        path: devicePlugin.hostNetwork
+        x-descriptors:
+        - urn:alm:descriptor:com.amd.deviceconfigs:hostNetwork
       - description: node labeller image registry secret used to pull/push images
         displayName: ImageRegistrySecret
         path: devicePlugin.imageRegistrySecret
@@ -572,6 +578,12 @@ spec:
         path: metricsExporter.enable
         x-descriptors:
         - urn:alm:descriptor:com.amd.deviceconfigs:enable
+      - description: HostNetwork specifies whether to use host network namespace for
+          the metrics exporter DaemonSet pods.
+        displayName: HostNetwork
+        path: metricsExporter.hostNetwork
+        x-descriptors:
+        - urn:alm:descriptor:com.amd.deviceconfigs:hostNetwork
       - description: metrics exporter image
         displayName: Image
         path: metricsExporter.image
diff --git a/bundle/manifests/amd.com_deviceconfigs.yaml b/bundle/manifests/amd.com_deviceconfigs.yaml
@@ -264,6 +264,10 @@ spec:
                     default: true
                     description: enable or disable the node labeller
                     type: boolean
+                  hostNetwork:
+                    description: HostNetwork specifies whether to use host network
+                      namespace for the device plugin DaemonSet pods.
+                    type: boolean
                   imageRegistrySecret:
                     description: node labeller image registry secret used to pull/push
                       images
@@ -799,6 +803,10 @@ spec:
                   enable:
                     description: enable metrics exporter, disabled by default
                     type: boolean
+                  hostNetwork:
+                    description: HostNetwork specifies whether to use host network
+                      namespace for the metrics exporter DaemonSet pods.
+                    type: boolean
                   image:
                     description: metrics exporter image
                     pattern: ^([a-z0-9]+(?:[._-][a-z0-9]+)*(:[0-9]+)?)(/[a-z0-9]+(?:[._-][a-z0-9]+)*)*(?::[a-z0-9._-]+)?(?:@[a-zA-Z0-9]+:[a-f0-9]+)?$
diff --git a/config/crd/bases/amd.com_deviceconfigs.yaml b/config/crd/bases/amd.com_deviceconfigs.yaml
@@ -260,6 +260,10 @@ spec:
                     default: true
                     description: enable or disable the node labeller
                     type: boolean
+                  hostNetwork:
+                    description: HostNetwork specifies whether to use host network
+                      namespace for the device plugin DaemonSet pods.
+                    type: boolean
                   imageRegistrySecret:
                     description: node labeller image registry secret used to pull/push
                       images
@@ -795,6 +799,10 @@ spec:
                   enable:
                     description: enable metrics exporter, disabled by default
                     type: boolean
+                  hostNetwork:
+                    description: HostNetwork specifies whether to use host network
+                      namespace for the metrics exporter DaemonSet pods.
+                    type: boolean
                   image:
                     description: metrics exporter image
                     pattern: ^([a-z0-9]+(?:[._-][a-z0-9]+)*(:[0-9]+)?)(/[a-z0-9]+(?:[._-][a-z0-9]+)*)*(?::[a-z0-9._-]+)?(?:@[a-zA-Z0-9]+:[a-f0-9]+)?$
diff --git a/config/manifests/bases/amd-gpu-operator.clusterserviceversion.yaml b/config/manifests/bases/amd-gpu-operator.clusterserviceversion.yaml
@@ -177,6 +177,12 @@ spec:
         path: devicePlugin.enableNodeLabeller
         x-descriptors:
         - urn:alm:descriptor:com.amd.deviceconfigs:enableNodeLabeller
+      - description: HostNetwork specifies whether to use host network namespace for
+          the device plugin DaemonSet pods.
+        displayName: HostNetwork
+        path: devicePlugin.hostNetwork
+        x-descriptors:
+        - urn:alm:descriptor:com.amd.deviceconfigs:hostNetwork
       - description: node labeller image registry secret used to pull/push images
         displayName: ImageRegistrySecret
         path: devicePlugin.imageRegistrySecret
@@ -539,6 +545,12 @@ spec:
         path: metricsExporter.enable
         x-descriptors:
         - urn:alm:descriptor:com.amd.deviceconfigs:enable
+      - description: HostNetwork specifies whether to use host network namespace for
+          the metrics exporter DaemonSet pods.
+        displayName: HostNetwork
+        path: metricsExporter.hostNetwork
+        x-descriptors:
+        - urn:alm:descriptor:com.amd.deviceconfigs:hostNetwork
       - description: metrics exporter image
         displayName: Image
         path: metricsExporter.image
diff --git a/hack/k8s-patch/metadata-patch/values.yaml b/hack/k8s-patch/metadata-patch/values.yaml
@@ -121,6 +121,10 @@ deviceConfig:
       nodeLabellerArguments: []
       # -- image pull secret for device plugin and node labeller, e.g. {"name": "mySecretName"}
       imageRegistrySecret: {}
+      # -- specify the kubelet device plugins directory path on the host, default is "/var/lib/kubelet/device-plugins", make sure this path is consistent with kubelet configuration on the cluster for device plugin to work properly
+      kubeletSocketPath: "/var/lib/kubelet/device-plugins"
+      # -- specify whether to use host network for device plugin daemonset
+      hostNetwork: false
       upgradePolicy:
         # -- the type of daemonset upgrade, RollingUpdate or OnDelete
         upgradeStrategy: RollingUpdate
@@ -145,6 +149,8 @@ deviceConfig:
       tolerations: []
       # -- metrics exporter image pull secret, e.g. {"name": "pullSecretName"}
       imageRegistrySecret: {}
+      # -- specify whether to use host network for metrics exporter daemonset
+      hostNetwork: false
       # -- metrics exporter node selector, if not specified it will reuse spec.selector
       selector: {}
       upgradePolicy:
diff --git a/hack/k8s-patch/template-patch/default-deviceconfig.yaml b/hack/k8s-patch/template-patch/default-deviceconfig.yaml
@@ -183,6 +183,10 @@ spec:
     {{- with .kubeletSocketPath }}
     kubeletSocketPath: {{ . }}
     {{- end }}
+
+    {{- if (hasKey . "hostNetwork") }}
+    hostNetwork: {{ .hostNetwork }}
+    {{- end }}
   {{- end }}
 
   {{- with .Values.deviceConfig.spec.metricsExporter }}
@@ -255,6 +259,10 @@ spec:
       {{- toYaml . | nindent 6 }}
     {{- end }}
 
+    {{- if (hasKey . "hostNetwork") }}
+    hostNetwork: {{ .hostNetwork }}
+    {{- end }}
+
     {{- with .rbacConfig }}
     rbacConfig:
       {{- if (hasKey . "enable") }}
diff --git a/helm-charts-k8s/Chart.lock b/helm-charts-k8s/Chart.lock
@@ -9,4 +9,4 @@ dependencies:
   repository: file://./charts/remediation-crds
   version: v1.0.0
 digest: sha256:4c6b1f3224839e54d1523759be597d20ca2fc6508eb17fda2992a95a00e1fd70
-generated: "2026-03-10T05:41:14.760044744Z"
+generated: "2026-03-11T23:36:30.598519294Z"
diff --git a/helm-charts-k8s/README.md b/helm-charts-k8s/README.md
@@ -181,7 +181,9 @@ Kubernetes: `>= 1.29.0-0`
 | deviceConfig.spec.devicePlugin.devicePluginTolerations | list | `[]` | device plugin tolerations |
 | deviceConfig.spec.devicePlugin.enableDevicePlugin | bool | `true` | enable/disable the device plugin. Cannot be enabled at the same time as draDriver. |
 | deviceConfig.spec.devicePlugin.enableNodeLabeller | bool | `true` | enable / disable node labeller |
+| deviceConfig.spec.devicePlugin.hostNetwork | bool | `false` | specify whether to use host network for device plugin daemonset |
 | deviceConfig.spec.devicePlugin.imageRegistrySecret | object | `{}` | image pull secret for device plugin and node labeller, e.g. {"name": "mySecretName"} |
+| deviceConfig.spec.devicePlugin.kubeletSocketPath | string | `"/var/lib/kubelet/device-plugins"` | specify the kubelet device plugins directory path on the host, default is "/var/lib/kubelet/device-plugins", make sure this path is consistent with kubelet configuration on the cluster for device plugin to work properly |
 | deviceConfig.spec.devicePlugin.nodeLabellerArguments | list | `[]` | pass supported labels while starting node labeller daemonset, default ["vram", "cu-count", "simd-count", "device-id", "family", "product-name", "driver-version"], also support ["compute-memory-partition", "compute-partitioning-supported", "memory-partitioning-supported"] |
 | deviceConfig.spec.devicePlugin.nodeLabellerImage | string | `"rocm/k8s-device-plugin:labeller-latest"` | node labeller image |
 | deviceConfig.spec.devicePlugin.nodeLabellerImagePullPolicy | string | `"IfNotPresent"` | node labeller image pull policy |
@@ -219,6 +221,7 @@ Kubernetes: `>= 1.29.0-0`
 | deviceConfig.spec.driver.version | string | `"30.20.1"` | specify an out-of-tree driver version to install |
 | deviceConfig.spec.metricsExporter.config | object | `{}` | name of the metrics exporter config map, e.g. {"name": "metricConfigMapName"} |
 | deviceConfig.spec.metricsExporter.enable | bool | `true` | enable / disable device metrics exporter |
+| deviceConfig.spec.metricsExporter.hostNetwork | bool | `false` | specify whether to use host network for metrics exporter daemonset |
 | deviceConfig.spec.metricsExporter.image | string | `"docker.io/rocm/device-metrics-exporter:latest"` | metrics exporter image |
 | deviceConfig.spec.metricsExporter.imagePullPolicy | string | `"IfNotPresent"` | metrics exporter image pull policy |
 | deviceConfig.spec.metricsExporter.imageRegistrySecret | object | `{}` | metrics exporter image pull secret, e.g. {"name": "pullSecretName"} |
diff --git a/helm-charts-k8s/crds/deviceconfig-crd.yaml b/helm-charts-k8s/crds/deviceconfig-crd.yaml
@@ -269,6 +269,10 @@ spec:
                     default: true
                     description: enable or disable the node labeller
                     type: boolean
+                  hostNetwork:
+                    description: HostNetwork specifies whether to use host network namespace
+                      for the device plugin DaemonSet pods.
+                    type: boolean
                   imageRegistrySecret:
                     description: node labeller image registry secret used to pull/push
                       images
@@ -803,6 +807,10 @@ spec:
                   enable:
                     description: enable metrics exporter, disabled by default
                     type: boolean
+                  hostNetwork:
+                    description: HostNetwork specifies whether to use host network namespace
+                      for the metrics exporter DaemonSet pods.
+                    type: boolean
                   image:
                     description: metrics exporter image
                     pattern: ^([a-z0-9]+(?:[._-][a-z0-9]+)*(:[0-9]+)?)(/[a-z0-9]+(?:[._-][a-z0-9]+)*)*(?::[a-z0-9._-]+)?(?:@[a-zA-Z0-9]+:[a-f0-9]+)?$
diff --git a/helm-charts-k8s/templates/default-deviceconfig.yaml b/helm-charts-k8s/templates/default-deviceconfig.yaml
@@ -183,6 +183,10 @@ spec:
     {{- with .kubeletSocketPath }}
     kubeletSocketPath: {{ . }}
     {{- end }}
+
+    {{- if (hasKey . "hostNetwork") }}
+    hostNetwork: {{ .hostNetwork }}
+    {{- end }}
   {{- end }}
 
   {{- with .Values.deviceConfig.spec.metricsExporter }}
@@ -255,6 +259,10 @@ spec:
       {{- toYaml . | nindent 6 }}
     {{- end }}
 
+    {{- if (hasKey . "hostNetwork") }}
+    hostNetwork: {{ .hostNetwork }}
+    {{- end }}
+
     {{- with .rbacConfig }}
     rbacConfig:
       {{- if (hasKey . "enable") }}
diff --git a/helm-charts-k8s/values.yaml b/helm-charts-k8s/values.yaml
@@ -121,6 +121,10 @@ deviceConfig:
       nodeLabellerArguments: []
       # -- image pull secret for device plugin and node labeller, e.g. {"name": "mySecretName"}
       imageRegistrySecret: {}
+      # -- specify the kubelet device plugins directory path on the host, default is "/var/lib/kubelet/device-plugins", make sure this path is consistent with kubelet configuration on the cluster for device plugin to work properly
+      kubeletSocketPath: "/var/lib/kubelet/device-plugins"
+      # -- specify whether to use host network for device plugin daemonset
+      hostNetwork: false
       upgradePolicy:
         # -- the type of daemonset upgrade, RollingUpdate or OnDelete
         upgradeStrategy: RollingUpdate
@@ -145,6 +149,8 @@ deviceConfig:
       tolerations: []
       # -- metrics exporter image pull secret, e.g. {"name": "pullSecretName"}
       imageRegistrySecret: {}
+      # -- specify whether to use host network for metrics exporter daemonset
+      hostNetwork: false
       # -- metrics exporter node selector, if not specified it will reuse spec.selector
       selector: {}
       upgradePolicy:
diff --git a/tests/helm-e2e/helm_e2e_test.go b/tests/helm-e2e/helm_e2e_test.go
@@ -516,6 +516,7 @@ deviceConfig:
       imageRegistrySecret:
         name: mySecretName
       kubeletSocketPath: /var/lib/kubelet/device-plugins-custom
+      hostNetwork: true
       upgradePolicy:
         # -- the type of daemonset upgrade, RollingUpdate or OnDelete
         upgradeStrategy: OnDelete
@@ -564,6 +565,7 @@ deviceConfig:
 						Name: "mySecretName",
 					},
 					KubeletSocketPath: "/var/lib/kubelet/device-plugins-custom",
+					HostNetwork:       &boolTrue,
 					UpgradePolicy: &v1alpha1.DaemonSetUpgradeSpec{
 						UpgradeStrategy: "OnDelete",
 						MaxUnavailable:  5,
@@ -797,6 +799,7 @@ deviceConfig:
         prometheus.io/port: "5001"
       serviceAnnotations:
         service.beta.kubernetes.io/aws-load-balancer-type: "nlb"
+      hostNetwork: true
       tolerations:
         - key: "example-key"
           operator: "Equal"
@@ -880,6 +883,7 @@ deviceConfig:
 					ServiceAnnotations: map[string]string{
 						"service.beta.kubernetes.io/aws-load-balancer-type": "nlb",
 					},
+					HostNetwork: &boolTrue,
 					Tolerations: []corev1.Toleration{
 						{
 							Key:      "example-key",
