diff --git a/libcontainer/cgroups/fs/fs.go b/libcontainer/cgroups/fs/fs.go
index a35f628a8..b574d3a05 100644
--- a/libcontainer/cgroups/fs/fs.go
+++ b/libcontainer/cgroups/fs/fs.go
@@ -35,6 +35,14 @@ var (
 
 var errSubsystemDoesNotExist = errors.New("cgroup: subsystem does not exist")
 
+func init() {
+	// If using cgroups-hybrid mode then add a "" controller indicating
+	// it should join the cgroups v2.
+	if cgroups.IsCgroup2HybridMode() {
+		subsystems = append(subsystems, &NameGroup{GroupName: "", Join: true})
+	}
+}
+
 type subsystem interface {
 	// Name returns the name of the subsystem.
 	Name() string
diff --git a/libcontainer/cgroups/systemd/v1.go b/libcontainer/cgroups/systemd/v1.go
index 1e54cc42c..a74a05a5c 100644
--- a/libcontainer/cgroups/systemd/v1.go
+++ b/libcontainer/cgroups/systemd/v1.go
@@ -143,6 +143,19 @@ func initPaths(c *configs.Cgroup) (map[string]string, error) {
 		}
 		paths[s.Name()] = subsystemPath
 	}
+
+	// If systemd is using cgroups-hybrid mode then add the slice path of
+	// this container to the paths so the following process executed with
+	// "runc exec" joins that cgroup as well.
+	if cgroups.IsCgroup2HybridMode() {
+		// "" means cgroup-hybrid path
+		cgroupsHybridPath, err := getSubsystemPath(slice, unit, "")
+		if err != nil && cgroups.IsNotFound(err) {
+			return nil, err
+		}
+		paths[""] = cgroupsHybridPath
+	}
+
 	return paths, nil
 }
 
diff --git a/libcontainer/cgroups/utils.go b/libcontainer/cgroups/utils.go
index c4c952946..3bab0e75a 100644
--- a/libcontainer/cgroups/utils.go
+++ b/libcontainer/cgroups/utils.go
@@ -21,11 +21,14 @@ import (
 const (
 	CgroupProcesses   = "cgroup.procs"
 	unifiedMountpoint = "/sys/fs/cgroup"
+	hybridMountpoint  = "/sys/fs/cgroup/unified"
 )
 
 var (
 	isUnifiedOnce sync.Once
 	isUnified     bool
+	isHybridOnce  sync.Once
+	isHybrid      bool
 )
 
 // IsCgroup2UnifiedMode returns whether we are running in cgroup v2 unified mode.
@@ -47,6 +50,24 @@ func IsCgroup2UnifiedMode() bool {
 	return isUnified
 }
 
+// IsCgroup2HybridMode returns whether we are running in cgroup v2 hybrid mode.
+func IsCgroup2HybridMode() bool {
+	isHybridOnce.Do(func() {
+		var st unix.Statfs_t
+		err := unix.Statfs(hybridMountpoint, &st)
+		if err != nil {
+			if os.IsNotExist(err) {
+				// ignore the "not found" error
+				isHybrid = false
+				return
+			}
+			panic(fmt.Sprintf("cannot statfs cgroup root: %s", err))
+		}
+		isHybrid = st.Type == unix.CGROUP2_SUPER_MAGIC
+	})
+	return isHybrid
+}
+
 type Mount struct {
 	Mountpoint string
 	Root       string
diff --git a/libcontainer/cgroups/v1_utils.go b/libcontainer/cgroups/v1_utils.go
index d850a36b3..47c75f22b 100644
--- a/libcontainer/cgroups/v1_utils.go
+++ b/libcontainer/cgroups/v1_utils.go
@@ -113,6 +113,11 @@ func FindCgroupMountpoint(cgroupPath, subsystem string) (string, error) {
 		return "", errUnified
 	}
 
+	// If subsystem is empty, we look for the cgroupv2 hybrid path.
+	if len(subsystem) == 0 {
+		return hybridMountpoint, nil
+	}
+
 	// Avoid parsing mountinfo by trying the default path first, if possible.
 	if path := tryDefaultPath(cgroupPath, subsystem); path != "" {
 		return path, nil
@@ -223,6 +228,11 @@ func GetOwnCgroupPath(subsystem string) (string, error) {
 		return "", err
 	}
 
+	// If subsystem is empty, we look for the cgroupv2 hybrid path.
+	if len(subsystem) == 0 {
+		return hybridMountpoint, nil
+	}
+
 	return getCgroupPathHelper(subsystem, cgroup)
 }