podman cgroup enhancement

currently, setting any sort of resource limit in a pod does nothing. With the newly refactored creation process in c/common, podman ca now set resources at a pod level
meaning that resource related flags can now be exposed to podman pod create.

cgroupfs and systemd are both supported with varying completion. cgroupfs is a much simpler process and one that is virtually complete for all resource types, the flags now just need to be added. systemd on the other hand
has to be handeled via the dbus api meaning that the limits need to be passed as recognized properties to systemd. The properties added so far are the ones that podman pod create supports as well as `cpuset-mems` as this will
be the next flag I work on.

Signed-off-by: Charlie Doern <cdoern@redhat.com>

vendor/github.com/opencontainers/runc/libcontainer/cgroups/fs/paths.go

@@ -0,0 +1,186 @@
+package fs
+
+import (
+	"errors"
+	"os"
+	"path/filepath"
+	"sync"
+
+	"golang.org/x/sys/unix"
+
+	"github.com/opencontainers/runc/libcontainer/cgroups"
+	"github.com/opencontainers/runc/libcontainer/configs"
+	"github.com/opencontainers/runc/libcontainer/utils"
+)
+
+// The absolute path to the root of the cgroup hierarchies.
+var (
+	cgroupRootLock sync.Mutex
+	cgroupRoot     string
+)
+
+const defaultCgroupRoot = "/sys/fs/cgroup"
+
+func initPaths(cg *configs.Cgroup) (map[string]string, error) {
+	root, err := rootPath()
+	if err != nil {
+		return nil, err
+	}
+
+	inner, err := innerPath(cg)
+	if err != nil {
+		return nil, err
+	}
+
+	paths := make(map[string]string)
+	for _, sys := range subsystems {
+		name := sys.Name()
+		path, err := subsysPath(root, inner, name)
+		if err != nil {
+			// The non-presence of the devices subsystem
+			// is considered fatal for security reasons.
+			if cgroups.IsNotFound(err) && (cg.SkipDevices || name != "devices") {
+				continue
+			}
+
+			return nil, err
+		}
+		paths[name] = path
+	}
+
+	return paths, nil
+}
+
+func tryDefaultCgroupRoot() string {
+	var st, pst unix.Stat_t
+
+	// (1) it should be a directory...
+	err := unix.Lstat(defaultCgroupRoot, &st)
+	if err != nil || st.Mode&unix.S_IFDIR == 0 {
+		return ""
+	}
+
+	// (2) ... and a mount point ...
+	err = unix.Lstat(filepath.Dir(defaultCgroupRoot), &pst)
+	if err != nil {
+		return ""
+	}
+
+	if st.Dev == pst.Dev {
+		// parent dir has the same dev -- not a mount point
+		return ""
+	}
+
+	// (3) ... of 'tmpfs' fs type.
+	var fst unix.Statfs_t
+	err = unix.Statfs(defaultCgroupRoot, &fst)
+	if err != nil || fst.Type != unix.TMPFS_MAGIC {
+		return ""
+	}
+
+	// (4) it should have at least 1 entry ...
+	dir, err := os.Open(defaultCgroupRoot)
+	if err != nil {
+		return ""
+	}
+	names, err := dir.Readdirnames(1)
+	if err != nil {
+		return ""
+	}
+	if len(names) < 1 {
+		return ""
+	}
+	// ... which is a cgroup mount point.
+	err = unix.Statfs(filepath.Join(defaultCgroupRoot, names[0]), &fst)
+	if err != nil || fst.Type != unix.CGROUP_SUPER_MAGIC {
+		return ""
+	}
+
+	return defaultCgroupRoot
+}
+
+// rootPath finds and returns path to the root of the cgroup hierarchies.
+func rootPath() (string, error) {
+	cgroupRootLock.Lock()
+	defer cgroupRootLock.Unlock()
+
+	if cgroupRoot != "" {
+		return cgroupRoot, nil
+	}
+
+	// fast path
+	cgroupRoot = tryDefaultCgroupRoot()
+	if cgroupRoot != "" {
+		return cgroupRoot, nil
+	}
+
+	// slow path: parse mountinfo
+	mi, err := cgroups.GetCgroupMounts(false)
+	if err != nil {
+		return "", err
+	}
+	if len(mi) < 1 {
+		return "", errors.New("no cgroup mount found in mountinfo")
+	}
+
+	// Get the first cgroup mount (e.g. "/sys/fs/cgroup/memory"),
+	// use its parent directory.
+	root := filepath.Dir(mi[0].Mountpoint)
+
+	if _, err := os.Stat(root); err != nil {
+		return "", err
+	}
+
+	cgroupRoot = root
+	return cgroupRoot, nil
+}
+
+func innerPath(c *configs.Cgroup) (string, error) {
+	if (c.Name != "" || c.Parent != "") && c.Path != "" {
+		return "", errors.New("cgroup: either Path or Name and Parent should be used")
+	}
+
+	// XXX: Do not remove CleanPath. Path safety is important! -- cyphar
+	innerPath := utils.CleanPath(c.Path)
+	if innerPath == "" {
+		cgParent := utils.CleanPath(c.Parent)
+		cgName := utils.CleanPath(c.Name)
+		innerPath = filepath.Join(cgParent, cgName)
+	}
+
+	return innerPath, nil
+}
+
+func subsysPath(root, inner, subsystem string) (string, error) {
+	// If the cgroup name/path is absolute do not look relative to the cgroup of the init process.
+	if filepath.IsAbs(inner) {
+		mnt, err := cgroups.FindCgroupMountpoint(root, subsystem)
+		// If we didn't mount the subsystem, there is no point we make the path.
+		if err != nil {
+			return "", err
+		}
+
+		// Sometimes subsystems can be mounted together as 'cpu,cpuacct'.
+		return filepath.Join(root, filepath.Base(mnt), inner), nil
+	}
+
+	// Use GetOwnCgroupPath instead of GetInitCgroupPath, because the creating
+	// process could in container and shared pid namespace with host, and
+	// /proc/1/cgroup could point to whole other world of cgroups.
+	parentPath, err := cgroups.GetOwnCgroupPath(subsystem)
+	if err != nil {
+		return "", err
+	}
+
+	return filepath.Join(parentPath, inner), nil
+}
+
+func apply(path string, pid int) error {
+	if path == "" {
+		return nil
+	}
+	if err := os.MkdirAll(path, 0o755); err != nil {
+		return err
+	}
+	return cgroups.WriteCgroupProc(path, pid)
+}