pkg/netns: ensure makeNetnsDir is race free

There are some rather bad problems when we bind mount over multiple
times, this is a rather small race but can happen. In order to avoid
this take an exclusive lock like ip netns add does because they create
the same bind mount setup. As such we will not race against other podman
process or ip netns add which is a good thing.

Signed-off-by: Paul Holzinger <pholzing@redhat.com>

pkg/netns/netns_linux.go

@@ -105,6 +105,25 @@ func makeNetnsDir(nsRunDir string) error {
 	if err != nil {
 		return err
 	}
+	// Important, the bind mount setup is racy if two process try to set it up in parallel.
+	// This can have very bad consequences because we end up with two duplicated mounts
+	// for the netns file that then might have a different parent mounts.
+	// Also because as root netns dir is also created by ip netns we should not race against them.
+	// Use a lock on the netns dir like they do, compare the iproute2 ip netns add code.
+	// https://github.com/iproute2/iproute2/blob/8b9d9ea42759c91d950356ca43930a975d0c352b/ip/ipnetns.c#L806-L815
+
+	dirFD, err := unix.Open(nsRunDir, unix.O_RDONLY|unix.O_DIRECTORY|unix.O_CLOEXEC, 0)
+	if err != nil {
+		return &os.PathError{Op: "open", Path: nsRunDir, Err: err}
+	}
+	// closing the fd will also unlock so we do not have to call flock(fd,LOCK_UN)
+	defer unix.Close(dirFD)
+
+	err = unix.Flock(dirFD, unix.LOCK_EX)
+	if err != nil {
+		return fmt.Errorf("failed to lock %s dir: %w", nsRunDir, err)
+	}
+
 	// Remount the namespace directory shared. This will fail with EINVAL
 	// if it is not already a mountpoint, so bind-mount it on to itself
 	// to "upgrade" it to a mountpoint.