k/k#135686: DeviceTaint と ReservedFor

func (tc *Controller) handlePod(pod *v1.Pod) {
	eviction := tc.podEvictionTime(pod)
	podRef := newObject(pod)
	if eviction == nil {
		if tc.cancelWorkWithEvent(podRef) {
			tc.logger.V(3).Info("Canceled pod eviction", "pod", podRef)
		}
		return
	}

	tc.logger.V(3).Info("Going to evict pod", "pod", podRef, "eviction", eviction)
	tc.evictPod(podRef, *eviction)

	// If any reason is because of a taint, then eviction is in progress and the status may need to be updated.
	for _, reason := range eviction.reason {
		if reason.rule != nil {
			tc.workqueue.Add(workItemForRule(reason.rule))
		}
	}
}

func (tc *Controller) handlePods(claim *resourceapi.ResourceClaim) {
	for _, consumer := range claim.Status.ReservedFor {
		if consumer.APIGroup == "" && consumer.Resource == "pods" {
			pod, err := tc.podLister.Pods(claim.Namespace).Get(consumer.Name)
			if err != nil {
				if apierrors.IsNotFound(err) {
					return
				}
				// Should not happen.
				utilruntime.HandleErrorWithLogger(tc.logger, err, "retrieve pod from cache")
				return
			}
			if pod.UID != consumer.UID {
				// Not the pod we were looking for.
				return
			}
			tc.handlePod(pod)
		}
	}
}

// ResourceClaimStatus tracks whether the resource has been allocated and what
// the result of that was.
type ResourceClaimStatus struct {
	// Allocation is set once the claim has been allocated successfully.
	//
	// +optional
	Allocation *AllocationResult

	// ReservedFor indicates which entities are currently allowed to use
	// the claim. A Pod which references a ResourceClaim which is not
	// reserved for that Pod will not be started. A claim that is in
	// use or might be in use because it has been reserved must not get
	// deallocated.
	//
	// In a cluster with multiple scheduler instances, two pods might get
	// scheduled concurrently by different schedulers. When they reference
	// the same ResourceClaim which already has reached its maximum number
	// of consumers, only one pod can be scheduled.
	//
	// Both schedulers try to add their pod to the claim.status.reservedFor
	// field, but only the update that reaches the API server first gets
	// stored. The other one fails with an error and the scheduler
	// which issued it knows that it must put the pod back into the queue,
	// waiting for the ResourceClaim to become usable again.
	//
	// There can be at most 256 such reservations. This may get increased in
	// the future, but not reduced.
	//
	// +optional
	// +listType=map
	// +listMapKey=uid
	// +patchStrategy=merge
	// +patchMergeKey=uid
	ReservedFor []ResourceClaimConsumerReference

	// DeallocationRequested is tombstoned since Kubernetes 1.32 where
	// it got removed. May be reused once decoding v1alpha3 is no longer
	// supported.
	// DeallocationRequested bool

	// Devices contains the status of each device allocated for this
	// claim, as reported by the driver. This can include driver-specific
	// information. Entries are owned by their respective drivers.
	//
	// +optional
	// +listType=map
	// +listMapKey=driver
	// +listMapKey=device
	// +listMapKey=pool
	// +listMapKey=shareID
	// +featureGate=DRAResourceClaimDeviceStatus
	Devices []AllocatedDeviceStatus
}