/*
Copyright 2017 The Kubernetes Authors.

Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at

    http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
*/

package node

import (
	"context"
	"fmt"
	"strings"

	v1 "k8s.io/api/core/v1"
	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
	"k8s.io/apimachinery/pkg/util/sets"
	"k8s.io/kubernetes/test/e2e/framework"
	e2enode "k8s.io/kubernetes/test/e2e/framework/node"
	e2epod "k8s.io/kubernetes/test/e2e/framework/pod"
	"k8s.io/kubernetes/test/e2e/storage/utils"
	imageutils "k8s.io/kubernetes/test/utils/image"
	admissionapi "k8s.io/pod-security-admission/api"

	"github.com/onsi/ginkgo/v2"
	"github.com/onsi/gomega"
)

func preparePod(name string, node *v1.Node, propagation *v1.MountPropagationMode, hostDir string) *v1.Pod {
	const containerName = "cntr"
	bTrue := true
	var oneSecond int64 = 1
	// The pod prepares /mnt/test/<podname> and sleeps.
	cmd := fmt.Sprintf("mkdir /mnt/test/%[1]s; sleep 3600", name)
	pod := &v1.Pod{
		ObjectMeta: metav1.ObjectMeta{
			Name: name,
		},
		Spec: v1.PodSpec{
			NodeName: node.Name,
			Containers: []v1.Container{
				{
					Name:    containerName,
					Image:   imageutils.GetE2EImage(imageutils.BusyBox),
					Command: []string{"sh", "-c", cmd},
					VolumeMounts: []v1.VolumeMount{
						{
							Name:             "host",
							MountPath:        "/mnt/test",
							MountPropagation: propagation,
						},
					},
					SecurityContext: &v1.SecurityContext{
						Privileged: &bTrue,
					},
				},
			},
			Volumes: []v1.Volume{
				{
					Name: "host",
					VolumeSource: v1.VolumeSource{
						HostPath: &v1.HostPathVolumeSource{
							Path: hostDir,
						},
					},
				},
			},
			// speed up termination of the pod
			TerminationGracePeriodSeconds: &oneSecond,
		},
	}
	return pod
}

var _ = SIGDescribe("Mount propagation", func() {
	f := framework.NewDefaultFramework("mount-propagation")
	f.NamespacePodSecurityLevel = admissionapi.LevelPrivileged

	ginkgo.It("should propagate mounts within defined scopes", func(ctx context.Context) {
		// This test runs two pods: master and slave with respective mount
		// propagation on common /var/lib/kubelet/XXXX directory. Both mount a
		// tmpfs to a subdirectory there. We check that these mounts are
		// propagated to the right places.

		hostExec := utils.NewHostExec(f)
		ginkgo.DeferCleanup(hostExec.Cleanup)

		// Pick a node where all pods will run.
		node, err := e2enode.GetRandomReadySchedulableNode(ctx, f.ClientSet)
		framework.ExpectNoError(err)

		// Fail the test if the namespace is not set. We expect that the
		// namespace is unique and we might delete user data if it's not.
		if len(f.Namespace.Name) == 0 {
			gomega.Expect(f.Namespace.Name).NotTo(gomega.BeEmpty())
			return
		}

		// hostDir is the directory that's shared via HostPath among all pods.
		// Make sure it's random enough so we don't clash with another test
		// running in parallel.
		hostDir := "/var/lib/kubelet/" + f.Namespace.Name
		ginkgo.DeferCleanup(func(ctx context.Context) error {
			cleanCmd := fmt.Sprintf("rm -rf %q", hostDir)
			return hostExec.IssueCommand(ctx, cleanCmd, node)
		})

		podClient := e2epod.NewPodClient(f)
		bidirectional := v1.MountPropagationBidirectional
		master := podClient.CreateSync(ctx, preparePod("master", node, &bidirectional, hostDir))

		hostToContainer := v1.MountPropagationHostToContainer
		slave := podClient.CreateSync(ctx, preparePod("slave", node, &hostToContainer, hostDir))

		none := v1.MountPropagationNone
		private := podClient.CreateSync(ctx, preparePod("private", node, &none, hostDir))
		defaultPropagation := podClient.CreateSync(ctx, preparePod("default", node, nil, hostDir))

		// Check that the pods sees directories of each other. This just checks
		// that they have the same HostPath, not the mount propagation.
		podNames := []string{master.Name, slave.Name, private.Name, defaultPropagation.Name}
		for _, podName := range podNames {
			for _, dirName := range podNames {
				cmd := fmt.Sprintf("test -d /mnt/test/%s", dirName)
				e2epod.ExecShellInPod(ctx, f, podName, cmd)
			}
		}

		// Each pod mounts one tmpfs to /mnt/test/<podname> and puts a file there.
		for _, podName := range podNames {
			cmd := fmt.Sprintf("mount -t tmpfs e2e-mount-propagation-%[1]s /mnt/test/%[1]s; echo %[1]s > /mnt/test/%[1]s/file", podName)
			e2epod.ExecShellInPod(ctx, f, podName, cmd)

			// unmount tmpfs when the test finishes
			cmd = fmt.Sprintf("umount /mnt/test/%s", podName)
			ginkgo.DeferCleanup(e2epod.ExecShellInPod, f, podName, cmd)
		}

		// The host mounts one tmpfs to testdir/host and puts a file there so we
		// can check mount propagation from the host to pods.
		cmd := fmt.Sprintf("mkdir %[1]q/host; mount -t tmpfs e2e-mount-propagation-host %[1]q/host; echo host > %[1]q/host/file", hostDir)
		err = hostExec.IssueCommand(ctx, cmd, node)
		framework.ExpectNoError(err)

		ginkgo.DeferCleanup(func(ctx context.Context) error {
			cmd := fmt.Sprintf("umount %q/host", hostDir)
			return hostExec.IssueCommand(ctx, cmd, node)
		})

		// Now check that mounts are propagated to the right containers.
		// expectedMounts is map of pod name -> expected mounts visible in the
		// pod.
		expectedMounts := map[string]sets.String{
			// Master sees only its own mount and not the slave's one.
			"master": sets.NewString("master", "host"),
			// Slave sees master's mount + itself.
			"slave": sets.NewString("master", "slave", "host"),
			// Private sees only its own mount
			"private": sets.NewString("private"),
			// Default (=private) sees only its own mount
			"default": sets.NewString("default"),
		}
		dirNames := append(podNames, "host")
		for podName, mounts := range expectedMounts {
			for _, mountName := range dirNames {
				cmd := fmt.Sprintf("cat /mnt/test/%s/file", mountName)
				stdout, stderr, err := e2epod.ExecShellInPodWithFullOutput(ctx, f, podName, cmd)
				framework.Logf("pod %s mount %s: stdout: %q, stderr: %q error: %v", podName, mountName, stdout, stderr, err)
				msg := fmt.Sprintf("When checking pod %s and directory %s", podName, mountName)
				shouldBeVisible := mounts.Has(mountName)
				if shouldBeVisible {
					framework.ExpectNoError(err, "%s: failed to run %q", msg, cmd)
					gomega.Expect(stdout).To(gomega.Equal(mountName), msg)
				} else {
					// We *expect* cat to return error here
					gomega.Expect(err).To(gomega.HaveOccurred(), msg)
				}
			}
		}

		// Find the kubelet PID to ensure we're working with the kubelet's mount namespace
		cmd = "pidof kubelet"
		kubeletPid, err := hostExec.IssueCommandWithResult(ctx, cmd, node)
		framework.ExpectNoError(err, "Checking kubelet pid")
		kubeletPid = strings.TrimSuffix(kubeletPid, "\n")
		gomega.Expect(strings.Count(kubeletPid, " ")).To(gomega.Equal(0), "kubelet should only have a single PID in the system (pidof returned %q)", kubeletPid)
		enterKubeletMountNS := fmt.Sprintf("nsenter -t %s -m", kubeletPid)

		// Check that the master and host mounts are propagated to the container runtime's mount namespace
		for _, mountName := range []string{"host", master.Name} {
			cmd := fmt.Sprintf("%s cat \"%s/%s/file\"", enterKubeletMountNS, hostDir, mountName)
			output, err := hostExec.IssueCommandWithResult(ctx, cmd, node)
			framework.ExpectNoError(err, "host container namespace should see mount from %s: %s", mountName, output)
			output = strings.TrimSuffix(output, "\n")
			gomega.Expect(output).To(gomega.Equal(mountName), "host container namespace should see mount contents from %s", mountName)
		}

		// Check that the slave, private, and default mounts are not propagated to the container runtime's mount namespace
		for _, podName := range []string{slave.Name, private.Name, defaultPropagation.Name} {
			cmd := fmt.Sprintf("%s test ! -e \"%s/%s/file\"", enterKubeletMountNS, hostDir, podName)
			output, err := hostExec.IssueCommandWithResult(ctx, cmd, node)
			framework.ExpectNoError(err, "host container namespace shouldn't see mount from %s: %s", podName, output)
		}
	})
})