Merge pull request #2100 from tkatila/gpu-rm-removal

Remove deprecated and unused functionality: GPU RM, QAT Kerneldrv, xpumanager-sidecar

Author: mythi

.github/workflows/lib-build.yaml

@@ -26,7 +26,6 @@ jobs:
           - intel-dsa-plugin
           - intel-iaa-plugin
           - intel-idxd-config-initcontainer
-          - intel-xpumanager-sidecar
 
           # # Demo images
           - crypto-perf

.github/workflows/lib-publish.yaml

@@ -56,7 +56,6 @@ jobs:
           - intel-dsa-plugin
           - intel-iaa-plugin
           - intel-idxd-config-initcontainer
-          - intel-xpumanager-sidecar
     steps:
       - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4
       - uses: actions/setup-go@d35c59abb061a4a6fb18e82ac0862c26744d6ab5 # v5

.trivyignore.yaml

@@ -19,17 +19,13 @@ misconfigurations:
   - id: AVD-KSV-0047
     statement: gpu plugin in kubelet mode requires "nodes/proxy" resource access
     paths:
-      - gpu_plugin/overlays/fractional_resources/gpu-manager-role.yaml
       - operator/rbac/gpu_manager_role.yaml
       - operator/rbac/role.yaml
 
   - id: AVD-KSV-0014
     statement: These are false detections for not setting "readOnlyFilesystem"
     paths:
       - fpga_plugin/overlays/region/mode-region.yaml
-      - gpu_plugin/overlays/fractional_resources/add-mounts.yaml
-      - gpu_plugin/overlays/fractional_resources/add-args.yaml
-      - gpu_plugin/overlays/fractional_resources/gpu-manager-role.yaml
       - gpu_plugin/overlays/monitoring_shared-dev_nfd/add-args.yaml
       - gpu_plugin/overlays/nfd_labeled_nodes/add-args.yaml
       - iaa_plugin/overlays/iaa_initcontainer/iaa_initcontainer.yaml

Makefile

@@ -177,7 +177,7 @@ endif
 
 dockerlib = build/docker/lib
 dockertemplates = build/docker/templates
-images = $(shell basename -s .Dockerfile.in -a $(dockertemplates)/*.Dockerfile.in | grep -v -e dlb -e fpga -e kerneldrv)
+images = $(shell basename -s .Dockerfile.in -a $(dockertemplates)/*.Dockerfile.in | grep -v -e dlb -e fpga -e xpumanager-sidecar)
 dockerfiles = $(shell basename -s .in -a $(dockertemplates)/*.Dockerfile.in | xargs -I"{}" echo build/docker/{})
 
 test-image-base-layer:

README.md

@@ -196,12 +196,6 @@ The [Device plugins operator README](cmd/operator/README.md) gives the installat
 
 The [Device plugins Operator for OpenShift](https://github.com/intel/intel-technology-enabling-for-openshift) gives the installation and usage details for the operator available on [Red Hat OpenShift Container Platform](https://catalog.redhat.com/software/operators/detail/61e9f2d7b9cdd99018fc5736).
 
-## XeLink XPU Manager Sidecar
-
-To support interconnected GPUs in Kubernetes, XeLink sidecar is needed.
-
-The [XeLink XPU Manager sidecar README](cmd/xpumanager_sidecar/README.md) gives information how the sidecar functions and how to use it.
-
 ## Intel GPU Level-Zero sidecar
 
 Sidecar uses Level-Zero API to provide additional GPU information for the GPU plugin that it cannot get through sysfs interfaces.

build/docker/intel-qat-plugin-kerneldrv.Dockerfile

@@ -47,20 +47,17 @@ Intel GPU plugin may register four node resources to the Kubernetes cluster:
 | gpu.intel.com/xe | GPU instance running new `xe` KMD |
 | gpu.intel.com/xe_monitoring | Monitoring resource for the new `xe` KMD devices |
 
-While GPU plugin basic operations support nodes having both (`i915` and `xe`) KMDs on the same node, its resource management (=GAS) does not, for that node needs to have only one of the KMDs present.
-
 For workloads on different KMDs, see [KMD and UMD](#kmd-and-umd).
 
 ## Modes and Configuration Options
 
 | Flag | Argument | Default | Meaning |
 |:---- |:-------- |:------- |:------- |
 | -enable-monitoring | - | disabled | Enable '*_monitoring' resource that provides access to all Intel GPU devices on the node, [see use](./monitoring.md) |
-| -resource-manager | - | disabled | Deprecated. Enable fractional resource management, [see use](./fractional.md) |
 | -health-management | - | disabled | Enable health management by requesting data from oneAPI/Level-Zero interface. Requires [GPU Level-Zero](../gpu_levelzero/) sidecar. See [health management](#health-management) |
 | -wsl | - | disabled | Adapt plugin to run in the WSL environment. Requires [GPU Level-Zero](../gpu_levelzero/) sidecar. |
 | -shared-dev-num | int | 1 | Number of containers that can share the same GPU device |
-| -allocation-policy | string | none | 3 possible values: balanced, packed, none. For shared-dev-num > 1: _balanced_ mode spreads workloads among GPU devices, _packed_ mode fills one GPU fully before moving to next, and _none_ selects first available device from kubelet. Default is _none_. Allocation policy does not have an effect when resource manager is enabled. |
+| -allocation-policy | string | none | 3 possible values: balanced, packed, none. For shared-dev-num > 1: _balanced_ mode spreads workloads among GPU devices, _packed_ mode fills one GPU fully before moving to next, and _none_ selects first available device from kubelet. Default is _none_. |
 
 The plugin also accepts a number of other arguments (common to all plugins) related to logging.
 Please use the -h option to see the complete list of logging related options.
@@ -75,9 +72,6 @@ Intel GPU-plugin supports a few different operation modes. Depending on the work
 |:---- |:-------- |:------- |:------- |
 | shared-dev-num == 1 | No, 1 container per GPU | Workloads using all GPU capacity, e.g. AI training | Yes |
 | shared-dev-num > 1 | Yes, >1 containers per GPU | (Batch) workloads using only part of GPU resources, e.g. inference, media transcode/analytics, or CPU bound GPU workloads | No |
-| shared-dev-num > 1 && resource-management | Depends on resource requests | Any. For requirements and usage, see [fractional resource management](./fractional.md) | Yes. 1000 millicores = exclusive GPU usage. See note below. |
-
-> **Note**: Exclusive GPU usage with >=1000 millicores requires that also *all other GPU containers* specify (non-zero) millicores resource usage.
 
 ## Installing driver and firmware for Intel GPUs
 
@@ -122,10 +116,6 @@ $ kubectl apply -k 'https://github.com/intel/intel-device-plugins-for-kubernetes
 
 GPU plugin can be installed with the Intel Device Plugin Operator. It allows configuring GPU plugin's parameters without kustomizing the deployment files. The general installation is described in the [install documentation](../operator/README.md#installation). For configuring the GPU Custom Resource (CR), see the [configuration options](#modes-and-configuration-options) and [operation modes](#operation-modes-for-different-workload-types).
 
-### Install alongside with GPU Aware Scheduling (deprecated)
-
-GPU plugin can be installed alongside with GPU Aware Scheduling (GAS). It allows scheduling Pods which e.g. request only partial use of a GPU. The installation is described in [fractional resources](./fractional.md) page.
-
 ### Verify Plugin Installation
 
 You can verify that the plugin has been installed on the expected nodes by searching for the relevant
@@ -212,9 +202,9 @@ Furthermore, the deployments `securityContext` must be configured with appropria
 
 More info: https://kubernetes.io/blog/2021/11/09/non-root-containers-and-devices/
 
-### Labels created by GPU plugin
+### Labels created for Intel GPUs via NFD
 
-If installed with NFD and started with resource-management, plugin will export a set of labels for the node. For detailed info, see [labeling documentation](./labels.md).
+When NFD's NodeFeatureRules for Intel GPUs are installed, nodes are labeled with a variaty of GPU specific labels. For detailed info, see [labeling documentation](./labels.md).
 
 ### SR-IOV use with the plugin
 

build/docker/templates/intel-qat-plugin-kerneldrv.Dockerfile.in

@@ -15,35 +15,22 @@
 package main
 
 import (
-	"slices"
-
-	"github.com/intel/intel-device-plugins-for-kubernetes/cmd/internal/labeler"
 	"github.com/intel/intel-device-plugins-for-kubernetes/cmd/internal/pluginutils"
 	"k8s.io/klog/v2"
 )
 
 type DeviceProperties struct {
 	currentDriver string
-	drmDrivers    map[string]bool
-	tileCounts    []uint64
 	isPfWithVfs   bool
 }
 
-type invalidTileCountErr struct {
-	error
-}
-
 func newDeviceProperties() *DeviceProperties {
-	return &DeviceProperties{
-		drmDrivers: make(map[string]bool),
-	}
+	return &DeviceProperties{}
 }
 
 func (d *DeviceProperties) fetch(cardPath string) {
 	d.isPfWithVfs = pluginutils.IsSriovPFwithVFs(cardPath)
 
-	d.tileCounts = append(d.tileCounts, labeler.GetTileCount(cardPath))
-
 	driverName, err := pluginutils.ReadDeviceDriver(cardPath)
 	if err != nil {
 		klog.Warningf("card (%s) doesn't have driver, using default: %s", cardPath, deviceTypeDefault)
@@ -52,11 +39,6 @@ func (d *DeviceProperties) fetch(cardPath string) {
 	}
 
 	d.currentDriver = driverName
-	d.drmDrivers[d.currentDriver] = true
-}
-
-func (d *DeviceProperties) drmDriverCount() int {
-	return len(d.drmDrivers)
 }
 
 func (d *DeviceProperties) driver() string {
@@ -66,20 +48,3 @@ func (d *DeviceProperties) driver() string {
 func (d *DeviceProperties) monitorResource() string {
 	return d.currentDriver + monitorSuffix
 }
-
-func (d *DeviceProperties) maxTileCount() (uint64, error) {
-	if len(d.tileCounts) == 0 {
-		return 0, invalidTileCountErr{}
-	}
-
-	minCount := slices.Min(d.tileCounts)
-	maxCount := slices.Max(d.tileCounts)
-
-	if minCount != maxCount {
-		klog.Warningf("Node's GPUs are heterogenous (min: %d, max: %d tiles)", minCount, maxCount)
-
-		return 0, invalidTileCountErr{}
-	}
-
-	return maxCount, nil
-}