*: introduce pidfd-socket flag

The container manager like containerd-shim can't use cgroup.kill feature or
freeze all the processes in cgroup to terminate the exec init process.
It's unsafe to call kill(2) since the pid can be recycled. It's good to
provide the pidfd of init process through the pidfd-socket. It's similar to
the console-socket. With the pidfd, the container manager like containerd-shim
can send the signal to target process safely.

And for the standard init process, we can have polling support to get
exit event instead of blocking on wait4.

Signed-off-by: Wei Fu <fuweid89@gmail.com>

utils_linux.go

@@ -18,6 +18,7 @@ import (
 	"github.com/opencontainers/runc/libcontainer"
 	"github.com/opencontainers/runc/libcontainer/configs"
 	"github.com/opencontainers/runc/libcontainer/specconv"
+	"github.com/opencontainers/runc/libcontainer/system/kernelversion"
 	"github.com/opencontainers/runc/libcontainer/utils"
 )
 
@@ -199,6 +200,7 @@ type runner struct {
 	preserveFDs     int
 	pidFile         string
 	consoleSocket   string
+	pidfdSocket     string
 	container       *libcontainer.Container
 	action          CtAct
 	notifySocket    *notifySocket
@@ -255,6 +257,14 @@ func (r *runner) run(config *specs.Process) (int, error) {
 	}
 	defer tty.Close()
 
+	if r.pidfdSocket != "" {
+		connClose, err := setupPidfdSocket(process, r.pidfdSocket)
+		if err != nil {
+			return -1, err
+		}
+		defer connClose()
+	}
+
 	switch r.action {
 	case CT_ACT_CREATE:
 		err = r.container.Start(process)
@@ -390,6 +400,7 @@ func startContainer(context *cli.Context, action CtAct, criuOpts *libcontainer.C
 		listenFDs:       listenFDs,
 		notifySocket:    notifySocket,
 		consoleSocket:   context.String("console-socket"),
+		pidfdSocket:     context.String("pidfd-socket"),
 		detach:          context.Bool("detach"),
 		pidFile:         context.String("pid-file"),
 		preserveFDs:     context.Int("preserve-fds"),
@@ -399,3 +410,36 @@ func startContainer(context *cli.Context, action CtAct, criuOpts *libcontainer.C
 	}
 	return r.run(spec.Process)
 }
+
+func setupPidfdSocket(process *libcontainer.Process, sockpath string) (_clean func(), _ error) {
+	linux530 := kernelversion.KernelVersion{Kernel: 5, Major: 3}
+	ok, err := kernelversion.GreaterEqualThan(linux530)
+	if err != nil {
+		return nil, err
+	}
+	if !ok {
+		return nil, fmt.Errorf("--pidfd-socket requires >= v5.3 kernel")
+	}
+
+	conn, err := net.Dial("unix", sockpath)
+	if err != nil {
+		return nil, fmt.Errorf("failed to dail %s: %w", sockpath, err)
+	}
+
+	uc, ok := conn.(*net.UnixConn)
+	if !ok {
+		conn.Close()
+		return nil, errors.New("failed to cast to UnixConn")
+	}
+
+	socket, err := uc.File()
+	if err != nil {
+		conn.Close()
+		return nil, fmt.Errorf("failed to dup socket: %w", err)
+	}
+
+	process.PidfdSocket = socket
+	return func() {
+		conn.Close()
+	}, nil
+}