Vendor containers/storage, update containers/image

Vendor containers/storage, and its dependencies github.com/pborman/uuid
and github.com/mistifyio/go-zfs, which we didn't already use.

Update the build Dockerfile to install their dependencies.

Add scriptlets that try to detect whether or not we need to use the
"libdm_no_deferred_remove" and/or "btrfs_noversion" build tags.

Signed-off-by: Nalin Dahyabhai <nalin@redhat.com>

vendor/github.com/opencontainers/runc/libcontainer/label/label.go

@@ -0,0 +1,84 @@
+// +build !selinux !linux
+
+package label
+
+// InitLabels returns the process label and file labels to be used within
+// the container.  A list of options can be passed into this function to alter
+// the labels.
+func InitLabels(options []string) (string, string, error) {
+	return "", "", nil
+}
+
+func GetROMountLabel() string {
+	return ""
+}
+
+func GenLabels(options string) (string, string, error) {
+	return "", "", nil
+}
+
+func FormatMountLabel(src string, mountLabel string) string {
+	return src
+}
+
+func SetProcessLabel(processLabel string) error {
+	return nil
+}
+
+func GetFileLabel(path string) (string, error) {
+	return "", nil
+}
+
+func SetFileLabel(path string, fileLabel string) error {
+	return nil
+}
+
+func SetFileCreateLabel(fileLabel string) error {
+	return nil
+}
+
+func Relabel(path string, fileLabel string, shared bool) error {
+	return nil
+}
+
+func GetPidLabel(pid int) (string, error) {
+	return "", nil
+}
+
+func Init() {
+}
+
+func ReserveLabel(label string) error {
+	return nil
+}
+
+func UnreserveLabel(label string) error {
+	return nil
+}
+
+// DupSecOpt takes a process label and returns security options that
+// can be used to set duplicate labels on future container processes
+func DupSecOpt(src string) []string {
+	return nil
+}
+
+// DisableSecOpt returns a security opt that can disable labeling
+// support for future container processes
+func DisableSecOpt() []string {
+	return nil
+}
+
+// Validate checks that the label does not include unexpected options
+func Validate(label string) error {
+	return nil
+}
+
+// RelabelNeeded checks whether the user requested a relabel
+func RelabelNeeded(label string) bool {
+	return false
+}
+
+// IsShared checks that the label includes a "shared" mark
+func IsShared(label string) bool {
+	return false
+}

vendor/github.com/opencontainers/runc/libcontainer/label/label_selinux.go

@@ -0,0 +1,208 @@
+// +build selinux,linux
+
+package label
+
+import (
+	"fmt"
+	"strings"
+
+	"github.com/opencontainers/runc/libcontainer/selinux"
+)
+
+// Valid Label Options
+var validOptions = map[string]bool{
+	"disable": true,
+	"type":    true,
+	"user":    true,
+	"role":    true,
+	"level":   true,
+}
+
+var ErrIncompatibleLabel = fmt.Errorf("Bad SELinux option z and Z can not be used together")
+
+// InitLabels returns the process label and file labels to be used within
+// the container.  A list of options can be passed into this function to alter
+// the labels.  The labels returned will include a random MCS String, that is
+// guaranteed to be unique.
+func InitLabels(options []string) (string, string, error) {
+	if !selinux.SelinuxEnabled() {
+		return "", "", nil
+	}
+	processLabel, mountLabel := selinux.GetLxcContexts()
+	if processLabel != "" {
+		pcon := selinux.NewContext(processLabel)
+		mcon := selinux.NewContext(mountLabel)
+		for _, opt := range options {
+			val := strings.SplitN(opt, "=", 2)
+			if val[0] != "label" {
+				continue
+			}
+			if len(val) < 2 {
+				return "", "", fmt.Errorf("bad label option %q, valid options 'disable' or \n'user, role, level, type' followed by ':' and a value", opt)
+			}
+			if val[1] == "disable" {
+				return "", "", nil
+			}
+			con := strings.SplitN(val[1], ":", 2)
+			if len(con) < 2 || !validOptions[con[0]] {
+				return "", "", fmt.Errorf("bad label option %q, valid options 'disable, user, role, level, type'", con[0])
+
+			}
+			pcon[con[0]] = con[1]
+			if con[0] == "level" || con[0] == "user" {
+				mcon[con[0]] = con[1]
+			}
+		}
+		processLabel = pcon.Get()
+		mountLabel = mcon.Get()
+	}
+	return processLabel, mountLabel, nil
+}
+
+func GetROMountLabel() string {
+	return selinux.GetROFileLabel()
+}
+
+// DEPRECATED: The GenLabels function is only to be used during the transition to the official API.
+func GenLabels(options string) (string, string, error) {
+	return InitLabels(strings.Fields(options))
+}
+
+// FormatMountLabel returns a string to be used by the mount command.
+// The format of this string will be used to alter the labeling of the mountpoint.
+// The string returned is suitable to be used as the options field of the mount command.
+// If you need to have additional mount point options, you can pass them in as
+// the first parameter.  Second parameter is the label that you wish to apply
+// to all content in the mount point.
+func FormatMountLabel(src, mountLabel string) string {
+	if mountLabel != "" {
+		switch src {
+		case "":
+			src = fmt.Sprintf("context=%q", mountLabel)
+		default:
+			src = fmt.Sprintf("%s,context=%q", src, mountLabel)
+		}
+	}
+	return src
+}
+
+// SetProcessLabel takes a process label and tells the kernel to assign the
+// label to the next program executed by the current process.
+func SetProcessLabel(processLabel string) error {
+	if processLabel == "" {
+		return nil
+	}
+	return selinux.Setexeccon(processLabel)
+}
+
+// GetProcessLabel returns the process label that the kernel will assign
+// to the next program executed by the current process.  If "" is returned
+// this indicates that the default labeling will happen for the process.
+func GetProcessLabel() (string, error) {
+	return selinux.Getexeccon()
+}
+
+// GetFileLabel returns the label for specified path
+func GetFileLabel(path string) (string, error) {
+	return selinux.Getfilecon(path)
+}
+
+// SetFileLabel modifies the "path" label to the specified file label
+func SetFileLabel(path string, fileLabel string) error {
+	if selinux.SelinuxEnabled() && fileLabel != "" {
+		return selinux.Setfilecon(path, fileLabel)
+	}
+	return nil
+}
+
+// SetFileCreateLabel tells the kernel the label for all files to be created
+func SetFileCreateLabel(fileLabel string) error {
+	if selinux.SelinuxEnabled() {
+		return selinux.Setfscreatecon(fileLabel)
+	}
+	return nil
+}
+
+// Relabel changes the label of path to the filelabel string.
+// It changes the MCS label to s0 if shared is true.
+// This will allow all containers to share the content.
+func Relabel(path string, fileLabel string, shared bool) error {
+	if !selinux.SelinuxEnabled() {
+		return nil
+	}
+
+	if fileLabel == "" {
+		return nil
+	}
+
+	exclude_paths := map[string]bool{"/": true, "/usr": true, "/etc": true}
+	if exclude_paths[path] {
+		return fmt.Errorf("SELinux relabeling of %s is not allowed", path)
+	}
+
+	if shared {
+		c := selinux.NewContext(fileLabel)
+		c["level"] = "s0"
+		fileLabel = c.Get()
+	}
+	if err := selinux.Chcon(path, fileLabel, true); err != nil {
+		return fmt.Errorf("SELinux relabeling of %s is not allowed: %q", path, err)
+	}
+	return nil
+}
+
+// GetPidLabel will return the label of the process running with the specified pid
+func GetPidLabel(pid int) (string, error) {
+	return selinux.Getpidcon(pid)
+}
+
+// Init initialises the labeling system
+func Init() {
+	selinux.SelinuxEnabled()
+}
+
+// ReserveLabel will record the fact that the MCS label has already been used.
+// This will prevent InitLabels from using the MCS label in a newly created
+// container
+func ReserveLabel(label string) error {
+	selinux.ReserveLabel(label)
+	return nil
+}
+
+// UnreserveLabel will remove the reservation of the MCS label.
+// This will allow InitLabels to use the MCS label in a newly created
+// containers
+func UnreserveLabel(label string) error {
+	selinux.FreeLxcContexts(label)
+	return nil
+}
+
+// DupSecOpt takes a process label and returns security options that
+// can be used to set duplicate labels on future container processes
+func DupSecOpt(src string) []string {
+	return selinux.DupSecOpt(src)
+}
+
+// DisableSecOpt returns a security opt that can disable labeling
+// support for future container processes
+func DisableSecOpt() []string {
+	return selinux.DisableSecOpt()
+}
+
+// Validate checks that the label does not include unexpected options
+func Validate(label string) error {
+	if strings.Contains(label, "z") && strings.Contains(label, "Z") {
+		return ErrIncompatibleLabel
+	}
+	return nil
+}
+
+// RelabelNeeded checks whether the user requested a relabel
+func RelabelNeeded(label string) bool {
+	return strings.Contains(label, "z") || strings.Contains(label, "Z")
+}
+
+// IsShared checks that the label includes a "shared" mark
+func IsShared(label string) bool {
+	return strings.Contains(label, "z")
+}

vendor/github.com/opencontainers/runc/libcontainer/system/linux.go

@@ -0,0 +1,143 @@
+// +build linux
+
+package system
+
+import (
+	"bufio"
+	"fmt"
+	"os"
+	"os/exec"
+	"syscall"
+	"unsafe"
+)
+
+// If arg2 is nonzero, set the "child subreaper" attribute of the
+// calling process; if arg2 is zero, unset the attribute.  When a
+// process is marked as a child subreaper, all of the children
+// that it creates, and their descendants, will be marked as
+// having a subreaper.  In effect, a subreaper fulfills the role
+// of init(1) for its descendant processes.  Upon termination of
+// a process that is orphaned (i.e., its immediate parent has
+// already terminated) and marked as having a subreaper, the
+// nearest still living ancestor subreaper will receive a SIGCHLD
+// signal and be able to wait(2) on the process to discover its
+// termination status.
+const PR_SET_CHILD_SUBREAPER = 36
+
+type ParentDeathSignal int
+
+func (p ParentDeathSignal) Restore() error {
+	if p == 0 {
+		return nil
+	}
+	current, err := GetParentDeathSignal()
+	if err != nil {
+		return err
+	}
+	if p == current {
+		return nil
+	}
+	return p.Set()
+}
+
+func (p ParentDeathSignal) Set() error {
+	return SetParentDeathSignal(uintptr(p))
+}
+
+func Execv(cmd string, args []string, env []string) error {
+	name, err := exec.LookPath(cmd)
+	if err != nil {
+		return err
+	}
+
+	return syscall.Exec(name, args, env)
+}
+
+func Prlimit(pid, resource int, limit syscall.Rlimit) error {
+	_, _, err := syscall.RawSyscall6(syscall.SYS_PRLIMIT64, uintptr(pid), uintptr(resource), uintptr(unsafe.Pointer(&limit)), uintptr(unsafe.Pointer(&limit)), 0, 0)
+	if err != 0 {
+		return err
+	}
+	return nil
+}
+
+func SetParentDeathSignal(sig uintptr) error {
+	if _, _, err := syscall.RawSyscall(syscall.SYS_PRCTL, syscall.PR_SET_PDEATHSIG, sig, 0); err != 0 {
+		return err
+	}
+	return nil
+}
+
+func GetParentDeathSignal() (ParentDeathSignal, error) {
+	var sig int
+	_, _, err := syscall.RawSyscall(syscall.SYS_PRCTL, syscall.PR_GET_PDEATHSIG, uintptr(unsafe.Pointer(&sig)), 0)
+	if err != 0 {
+		return -1, err
+	}
+	return ParentDeathSignal(sig), nil
+}
+
+func SetKeepCaps() error {
+	if _, _, err := syscall.RawSyscall(syscall.SYS_PRCTL, syscall.PR_SET_KEEPCAPS, 1, 0); err != 0 {
+		return err
+	}
+
+	return nil
+}
+
+func ClearKeepCaps() error {
+	if _, _, err := syscall.RawSyscall(syscall.SYS_PRCTL, syscall.PR_SET_KEEPCAPS, 0, 0); err != 0 {
+		return err
+	}
+
+	return nil
+}
+
+func Setctty() error {
+	if _, _, err := syscall.RawSyscall(syscall.SYS_IOCTL, 0, uintptr(syscall.TIOCSCTTY), 0); err != 0 {
+		return err
+	}
+	return nil
+}
+
+// RunningInUserNS detects whether we are currently running in a user namespace.
+// Copied from github.com/lxc/lxd/shared/util.go
+func RunningInUserNS() bool {
+	file, err := os.Open("/proc/self/uid_map")
+	if err != nil {
+		// This kernel-provided file only exists if user namespaces are supported
+		return false
+	}
+	defer file.Close()
+
+	buf := bufio.NewReader(file)
+	l, _, err := buf.ReadLine()
+	if err != nil {
+		return false
+	}
+
+	line := string(l)
+	var a, b, c int64
+	fmt.Sscanf(line, "%d %d %d", &a, &b, &c)
+	/*
+	 * We assume we are in the initial user namespace if we have a full
+	 * range - 4294967295 uids starting at uid 0.
+	 */
+	if a == 0 && b == 0 && c == 4294967295 {
+		return false
+	}
+	return true
+}
+
+// SetSubreaper sets the value i as the subreaper setting for the calling process
+func SetSubreaper(i int) error {
+	return Prctl(PR_SET_CHILD_SUBREAPER, uintptr(i), 0, 0, 0)
+}
+
+func Prctl(option int, arg2, arg3, arg4, arg5 uintptr) (err error) {
+	_, _, e1 := syscall.Syscall6(syscall.SYS_PRCTL, uintptr(option), arg2, arg3, arg4, arg5, 0)
+	if e1 != 0 {
+		err = e1
+	}
+	return
+}

vendor/github.com/opencontainers/runc/libcontainer/system/proc.go

@@ -0,0 +1,43 @@
+package system
+
+import (
+	"io/ioutil"
+	"path/filepath"
+	"strconv"
+	"strings"
+)
+
+// look in /proc to find the process start time so that we can verify
+// that this pid has started after ourself
+func GetProcessStartTime(pid int) (string, error) {
+	data, err := ioutil.ReadFile(filepath.Join("/proc", strconv.Itoa(pid), "stat"))
+	if err != nil {
+		return "", err
+	}
+	return parseStartTime(string(data))
+}
+
+func parseStartTime(stat string) (string, error) {
+	// the starttime is located at pos 22
+	// from the man page
+	//
+	// starttime %llu (was %lu before Linux 2.6)
+	// (22)  The  time the process started after system boot.  In kernels before Linux 2.6, this
+	// value was expressed in jiffies.  Since Linux 2.6, the value is expressed in  clock  ticks
+	// (divide by sysconf(_SC_CLK_TCK)).
+	//
+	// NOTE:
+	// pos 2 could contain space and is inside `(` and `)`:
+	// (2) comm  %s
+	// The filename of the executable, in parentheses.
+	// This is visible whether or not the executable is
+	// swapped out.
+	//
+	// the following is an example:
+	// 89653 (gunicorn: maste) S 89630 89653 89653 0 -1 4194560 29689 28896 0 3 146 32 76 19 20 0 1 0 2971844 52965376 3920 18446744073709551615 1 1 0 0 0 0 0 16781312 137447943 0 0 0 17 1 0 0 0 0 0 0 0 0 0 0 0 0 0
+
+	// get parts after last `)`:
+	s := strings.Split(stat, ")")
+	parts := strings.Split(strings.TrimSpace(s[len(s)-1]), " ")
+	return parts[22-3], nil // starts at 3 (after the filename pos `2`)
+}

vendor/github.com/opencontainers/runc/libcontainer/system/setns_linux.go

@@ -0,0 +1,40 @@
+package system
+
+import (
+	"fmt"
+	"runtime"
+	"syscall"
+)
+
+// Via http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=7b21fddd087678a70ad64afc0f632e0f1071b092
+//
+// We need different setns values for the different platforms and arch
+// We are declaring the macro here because the SETNS syscall does not exist in th stdlib
+var setNsMap = map[string]uintptr{
+	"linux/386":     346,
+	"linux/arm64":   268,
+	"linux/amd64":   308,
+	"linux/arm":     375,
+	"linux/ppc":     350,
+	"linux/ppc64":   350,
+	"linux/ppc64le": 350,
+	"linux/s390x":   339,
+}
+
+var sysSetns = setNsMap[fmt.Sprintf("%s/%s", runtime.GOOS, runtime.GOARCH)]
+
+func SysSetns() uint32 {
+	return uint32(sysSetns)
+}
+
+func Setns(fd uintptr, flags uintptr) error {
+	ns, exists := setNsMap[fmt.Sprintf("%s/%s", runtime.GOOS, runtime.GOARCH)]
+	if !exists {
+		return fmt.Errorf("unsupported platform %s/%s", runtime.GOOS, runtime.GOARCH)
+	}
+	_, _, err := syscall.RawSyscall(ns, fd, flags, 0)
+	if err != 0 {
+		return err
+	}
+	return nil
+}

vendor/github.com/opencontainers/runc/libcontainer/system/syscall_linux_386.go

@@ -0,0 +1,25 @@
+// +build linux,386
+
+package system
+
+import (
+	"syscall"
+)
+
+// Setuid sets the uid of the calling thread to the specified uid.
+func Setuid(uid int) (err error) {
+	_, _, e1 := syscall.RawSyscall(syscall.SYS_SETUID32, uintptr(uid), 0, 0)
+	if e1 != 0 {
+		err = e1
+	}
+	return
+}
+
+// Setgid sets the gid of the calling thread to the specified gid.
+func Setgid(gid int) (err error) {
+	_, _, e1 := syscall.RawSyscall(syscall.SYS_SETGID32, uintptr(gid), 0, 0)
+	if e1 != 0 {
+		err = e1
+	}
+	return
+}

vendor/github.com/opencontainers/runc/libcontainer/system/syscall_linux_64.go

@@ -0,0 +1,25 @@
+// +build linux,arm64 linux,amd64 linux,ppc linux,ppc64 linux,ppc64le linux,s390x
+
+package system
+
+import (
+	"syscall"
+)
+
+// Setuid sets the uid of the calling thread to the specified uid.
+func Setuid(uid int) (err error) {
+	_, _, e1 := syscall.RawSyscall(syscall.SYS_SETUID, uintptr(uid), 0, 0)
+	if e1 != 0 {
+		err = e1
+	}
+	return
+}
+
+// Setgid sets the gid of the calling thread to the specified gid.
+func Setgid(gid int) (err error) {
+	_, _, e1 := syscall.RawSyscall(syscall.SYS_SETGID, uintptr(gid), 0, 0)
+	if e1 != 0 {
+		err = e1
+	}
+	return
+}

vendor/github.com/opencontainers/runc/libcontainer/system/syscall_linux_arm.go

@@ -0,0 +1,25 @@
+// +build linux,arm
+
+package system
+
+import (
+	"syscall"
+)
+
+// Setuid sets the uid of the calling thread to the specified uid.
+func Setuid(uid int) (err error) {
+	_, _, e1 := syscall.RawSyscall(syscall.SYS_SETUID32, uintptr(uid), 0, 0)
+	if e1 != 0 {
+		err = e1
+	}
+	return
+}
+
+// Setgid sets the gid of the calling thread to the specified gid.
+func Setgid(gid int) (err error) {
+	_, _, e1 := syscall.RawSyscall(syscall.SYS_SETGID32, uintptr(gid), 0, 0)
+	if e1 != 0 {
+		err = e1
+	}
+	return
+}

vendor/github.com/opencontainers/runc/libcontainer/system/sysconfig.go

@@ -0,0 +1,12 @@
+// +build cgo,linux cgo,freebsd
+
+package system
+
+/*
+#include <unistd.h>
+*/
+import "C"
+
+func GetClockTicks() int {
+	return int(C.sysconf(C._SC_CLK_TCK))
+}

vendor/github.com/opencontainers/runc/libcontainer/system/sysconfig_notcgo.go

@@ -0,0 +1,15 @@
+// +build !cgo windows
+
+package system
+
+func GetClockTicks() int {
+	// TODO figure out a better alternative for platforms where we're missing cgo
+	//
+	// TODO Windows. This could be implemented using Win32 QueryPerformanceFrequency().
+	// https://msdn.microsoft.com/en-us/library/windows/desktop/ms644905(v=vs.85).aspx
+	//
+	// An example of its usage can be found here.
+	// https://msdn.microsoft.com/en-us/library/windows/desktop/dn553408(v=vs.85).aspx
+
+	return 100
+}

vendor/github.com/opencontainers/runc/libcontainer/system/unsupported.go

@@ -0,0 +1,9 @@
+// +build !linux
+
+package system
+
+// RunningInUserNS is a stub for non-Linux systems
+// Always returns false
+func RunningInUserNS() bool {
+	return false
+}

vendor/github.com/opencontainers/runc/libcontainer/system/xattrs_linux.go

@@ -0,0 +1,99 @@
+package system
+
+import (
+	"syscall"
+	"unsafe"
+)
+
+var _zero uintptr
+
+// Returns the size of xattrs and nil error
+// Requires path, takes allocated []byte or nil as last argument
+func Llistxattr(path string, dest []byte) (size int, err error) {
+	pathBytes, err := syscall.BytePtrFromString(path)
+	if err != nil {
+		return -1, err
+	}
+	var newpathBytes unsafe.Pointer
+	if len(dest) > 0 {
+		newpathBytes = unsafe.Pointer(&dest[0])
+	} else {
+		newpathBytes = unsafe.Pointer(&_zero)
+	}
+
+	_size, _, errno := syscall.Syscall6(syscall.SYS_LLISTXATTR, uintptr(unsafe.Pointer(pathBytes)), uintptr(newpathBytes), uintptr(len(dest)), 0, 0, 0)
+	size = int(_size)
+	if errno != 0 {
+		return -1, errno
+	}
+
+	return size, nil
+}
+
+// Returns a []byte slice if the xattr is set and nil otherwise
+// Requires path and its attribute as arguments
+func Lgetxattr(path string, attr string) ([]byte, error) {
+	var sz int
+	pathBytes, err := syscall.BytePtrFromString(path)
+	if err != nil {
+		return nil, err
+	}
+	attrBytes, err := syscall.BytePtrFromString(attr)
+	if err != nil {
+		return nil, err
+	}
+
+	// Start with a 128 length byte array
+	sz = 128
+	dest := make([]byte, sz)
+	destBytes := unsafe.Pointer(&dest[0])
+	_sz, _, errno := syscall.Syscall6(syscall.SYS_LGETXATTR, uintptr(unsafe.Pointer(pathBytes)), uintptr(unsafe.Pointer(attrBytes)), uintptr(destBytes), uintptr(len(dest)), 0, 0)
+
+	switch {
+	case errno == syscall.ENODATA:
+		return nil, errno
+	case errno == syscall.ENOTSUP:
+		return nil, errno
+	case errno == syscall.ERANGE:
+		// 128 byte array might just not be good enough,
+		// A dummy buffer is used ``uintptr(0)`` to get real size
+		// of the xattrs on disk
+		_sz, _, errno = syscall.Syscall6(syscall.SYS_LGETXATTR, uintptr(unsafe.Pointer(pathBytes)), uintptr(unsafe.Pointer(attrBytes)), uintptr(unsafe.Pointer(nil)), uintptr(0), 0, 0)
+		sz = int(_sz)
+		if sz < 0 {
+			return nil, errno
+		}
+		dest = make([]byte, sz)
+		destBytes := unsafe.Pointer(&dest[0])
+		_sz, _, errno = syscall.Syscall6(syscall.SYS_LGETXATTR, uintptr(unsafe.Pointer(pathBytes)), uintptr(unsafe.Pointer(attrBytes)), uintptr(destBytes), uintptr(len(dest)), 0, 0)
+		if errno != 0 {
+			return nil, errno
+		}
+	case errno != 0:
+		return nil, errno
+	}
+	sz = int(_sz)
+	return dest[:sz], nil
+}
+
+func Lsetxattr(path string, attr string, data []byte, flags int) error {
+	pathBytes, err := syscall.BytePtrFromString(path)
+	if err != nil {
+		return err
+	}
+	attrBytes, err := syscall.BytePtrFromString(attr)
+	if err != nil {
+		return err
+	}
+	var dataBytes unsafe.Pointer
+	if len(data) > 0 {
+		dataBytes = unsafe.Pointer(&data[0])
+	} else {
+		dataBytes = unsafe.Pointer(&_zero)
+	}
+	_, _, errno := syscall.Syscall6(syscall.SYS_LSETXATTR, uintptr(unsafe.Pointer(pathBytes)), uintptr(unsafe.Pointer(attrBytes)), uintptr(dataBytes), uintptr(len(data)), uintptr(flags), 0)
+	if errno != 0 {
+		return errno
+	}
+	return nil
+}