kube: plug HostUsers in the pod spec

map HostUsers=false to userns=auto.

One difference with the current implementation in the Kubelet is that
the podman default size is 1024 while the Kubelet uses 65536.

This is done on purpose, because 65536 is a problem for rootless as
the entire IDs space would be allocated to a single pod.

Signed-off-by: Giuseppe Scrivano <gscrivan@redhat.com>

3 files changed, 48 insertions, 2 deletions

diff --git a/pkg/domain/infra/abi/play.go b/pkg/domain/infra/abi/play.go
index 12786afcd..57d795682 100644
--- a/pkg/domain/infra/abi/play.go
+++ b/pkg/domain/infra/abi/play.go
@@ -355,6 +355,11 @@ func (ic *ContainerEngine) playKubePod(ctx context.Context, podName string, podY
 
 	if options.Userns == "" {
 		options.Userns = "host"
+		if podYAML.Spec.HostUsers != nil && !*podYAML.Spec.HostUsers {
+			options.Userns = "auto"
+		}
+	} else if podYAML.Spec.HostUsers != nil {
+		logrus.Info("overriding the user namespace mode in the pod spec")
 	}
 
 	// Validate the userns modes supported.
diff --git a/pkg/k8s.io/api/core/v1/types.go b/pkg/k8s.io/api/core/v1/types.go
index d47178878..6f20cd351 100644
--- a/pkg/k8s.io/api/core/v1/types.go
+++ b/pkg/k8s.io/api/core/v1/types.go
@@ -1984,6 +1984,18 @@ type PodSpec struct {
 	// Default to false.
 	// +optional
 	SetHostnameAsFQDN *bool `json:"setHostnameAsFQDN,omitempty"`
+	// Use the host's user namespace.
+	// Optional: Default to true.
+	// If set to true or not present, the pod will be run in the host user namespace, useful
+	// for when the pod needs a feature only available to the host user namespace, such as
+	// loading a kernel module with CAP_SYS_MODULE.
+	// When set to false, a new userns is created for the pod. Setting false is useful for
+	// mitigating container breakout vulnerabilities even allowing users to run their
+	// containers as root without actually having root privileges on the host.
+	// This field is alpha-level and is only honored by servers that enable the UserNamespacesSupport feature.
+	// +k8s:conversion-gen=false
+	// +optional
+	HostUsers *bool `json:"hostUsers,omitempty"`
 }
 
 type UnsatisfiableConstraintAction string
diff --git a/test/e2e/play_kube_test.go b/test/e2e/play_kube_test.go
index d1eb960cd..baa74cb51 100644
--- a/test/e2e/play_kube_test.go
+++ b/test/e2e/play_kube_test.go
@@ -380,6 +380,9 @@ spec:
   restartPolicy: {{ .RestartPolicy }}
   hostname: {{ .Hostname }}
   hostNetwork: {{ .HostNetwork }}
+{{ if .HostUsers }}
+  hostUsers: {{ .HostUsers }}
+{{ end }}
   hostAliases:
 {{ range .HostAliases }}
   - hostnames:
@@ -844,6 +847,7 @@ type Pod struct {
 	RestartPolicy   string
 	Hostname        string
 	HostNetwork     bool
+	HostUsers       *bool
 	HostAliases     []HostAlias
 	Ctrs            []*Ctr
 	InitCtrs        []*Ctr
@@ -968,6 +972,12 @@ func withHostNetwork() podOption {
 	}
 }
 
+func withHostUsers(val bool) podOption {
+	return func(pod *Pod) {
+		pod.HostUsers = &val
+	}
+}
+
 // Deployment describes the options a kube yaml can be configured at deployment level
 type Deployment struct {
 	Name        string
@@ -3783,8 +3793,7 @@ ENV OPENJ9_JAVA_OPTIONS=%q
 		Expect((inspect.InspectContainerToJSON()[0]).HostConfig.LogConfig.Tag).To(Equal("{{.ImageName}}"))
 	})
 
-	// Check that --userns=auto creates a user namespace
-	It("podman play kube --userns=auto", func() {
+	It("podman play kube using a user namespace", func() {
 		u, err := user.Current()
 		Expect(err).To(BeNil())
 		name := u.Name
@@ -3831,6 +3840,26 @@ ENV OPENJ9_JAVA_OPTIONS=%q
 		usernsInCtr.WaitWithDefaultTimeout()
 		Expect(usernsInCtr).Should(Exit(0))
 		Expect(string(usernsInCtr.Out.Contents())).To(Not(Equal(string(initialUsernsConfig))))
+
+		// Now try with hostUsers in the pod spec
+		for _, hostUsers := range []bool{true, false} {
+			pod = getPod(withHostUsers(hostUsers))
+			err = generateKubeYaml("pod", pod, kubeYaml)
+			Expect(err).To(BeNil())
+
+			kube = podmanTest.PodmanNoCache([]string{"play", "kube", "--replace", kubeYaml})
+			kube.WaitWithDefaultTimeout()
+			Expect(kube).Should(Exit(0))
+
+			usernsInCtr = podmanTest.Podman([]string{"exec", getCtrNameInPod(pod), "cat", "/proc/self/uid_map"})
+			usernsInCtr.WaitWithDefaultTimeout()
+			Expect(usernsInCtr).Should(Exit(0))
+			if hostUsers {
+				Expect(string(usernsInCtr.Out.Contents())).To(Equal(string(initialUsernsConfig)))
+			} else {
+				Expect(string(usernsInCtr.Out.Contents())).To(Not(Equal(string(initialUsernsConfig))))
+			}
+		}
 	})
 
 	// Check the block devices are exposed inside container