kill: wait for the container

Make sure to wait for the container to exit after kill. While the
cleanup process will take care eventually of transitioning the state, we
need to give a guarantee to the user to leave the container in the
expected state once the (kill) command has finished.

The issue could be observed in a flaking test (#16142) where
`podman rm -f -t0` failed because the preceding `podman kill`
left the container in "running" state which ultimately confused
the "stop" backend.

Note that we should only wait for the container to exit when SIGKILL is
being used.  Other signals have different semantics.

[NO NEW TESTS NEEDED] as I do not know how to reliably reproduce the
issue.  If #16142 stops flaking, we are good.

Fixes: #16142
Signed-off-by: Valentin Rothberg <vrothberg@redhat.com>

libpod/container_api.go

@@ -244,7 +244,12 @@ func (c *Container) Kill(signal uint) error {
 
 	c.newContainerEvent(events.Kill)
 
-	return c.save()
+	// Make sure to wait for the container to exit in case of SIGKILL.
+	if signal == uint(unix.SIGKILL) {
+		return c.waitForConmonToExitAndSave()
+	}
+
+	return nil
 }
 
 // Attach attaches to a container.

libpod/container_internal.go

@@ -1325,7 +1325,10 @@ func (c *Container) stop(timeout uint) error {
 
 	c.newContainerEvent(events.Stop)
 	c.state.StoppedByUser = true
+	return c.waitForConmonToExitAndSave()
+}
 
+func (c *Container) waitForConmonToExitAndSave() error {
 	conmonAlive, err := c.ociRuntime.CheckConmonRunning(c)
 	if err != nil {
 		return err

test/e2e/attach_test.go

@@ -82,6 +82,7 @@ var _ = Describe("Podman attach", func() {
 		Expect(results.OutputToString()).To(ContainSubstring("test"))
 		Expect(podmanTest.NumberOfContainersRunning()).To(Equal(1))
 	})
+
 	It("podman attach to the latest container", func() {
 		session := podmanTest.Podman([]string{"run", "-d", "--name", "test1", ALPINE, "/bin/sh", "-c", "while true; do echo test1; sleep 1; done"})
 		session.WaitWithDefaultTimeout()