1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544
|
//go:build linux
// +build linux
package fuseoverlayfs // import "github.com/docker/docker/daemon/graphdriver/fuse-overlayfs"
import (
"bytes"
"context"
"fmt"
"io"
"os"
"os/exec"
"path"
"path/filepath"
"strings"
"github.com/containerd/containerd/sys"
"github.com/docker/docker/daemon/graphdriver"
"github.com/docker/docker/daemon/graphdriver/overlayutils"
"github.com/docker/docker/pkg/archive"
"github.com/docker/docker/pkg/chrootarchive"
"github.com/docker/docker/pkg/containerfs"
"github.com/docker/docker/pkg/directory"
"github.com/docker/docker/pkg/idtools"
"github.com/docker/docker/pkg/parsers/kernel"
"github.com/docker/docker/pkg/system"
"github.com/moby/locker"
"github.com/moby/sys/mount"
"github.com/opencontainers/selinux/go-selinux/label"
"github.com/pkg/errors"
"github.com/sirupsen/logrus"
"golang.org/x/sys/unix"
)
var (
// untar defines the untar method
untar = chrootarchive.UntarUncompressed
)
const (
driverName = "fuse-overlayfs"
binary = "fuse-overlayfs"
linkDir = "l"
diffDirName = "diff"
workDirName = "work"
mergedDirName = "merged"
lowerFile = "lower"
maxDepth = 128
// idLength represents the number of random characters
// which can be used to create the unique link identifier
// for every layer. If this value is too long then the
// page size limit for the mount command may be exceeded.
// The idLength should be selected such that following equation
// is true (512 is a buffer for label metadata).
// ((idLength + len(linkDir) + 1) * maxDepth) <= (pageSize - 512)
idLength = 26
)
// Driver contains information about the home directory and the list of active
// mounts that are created using this driver.
type Driver struct {
home string
uidMaps []idtools.IDMap
gidMaps []idtools.IDMap
ctr *graphdriver.RefCounter
naiveDiff graphdriver.DiffDriver
locker *locker.Locker
}
var (
logger = logrus.WithField("storage-driver", driverName)
)
func init() {
graphdriver.Register(driverName, Init)
}
// Init returns the naive diff driver for fuse-overlayfs.
// If fuse-overlayfs is not supported on the host, the error
// graphdriver.ErrNotSupported is returned.
func Init(home string, options []string, uidMaps, gidMaps []idtools.IDMap) (graphdriver.Driver, error) {
if _, err := exec.LookPath(binary); err != nil {
logger.Error(err)
return nil, graphdriver.ErrNotSupported
}
if !kernel.CheckKernelVersion(4, 18, 0) {
return nil, graphdriver.ErrNotSupported
}
remappedRoot := idtools.NewIDMappingsFromMaps(uidMaps, gidMaps)
currentID := idtools.CurrentIdentity()
dirID := idtools.Identity{
UID: currentID.UID,
GID: remappedRoot.RootPair().GID,
}
if err := idtools.MkdirAllAndChown(home, 0710, dirID); err != nil {
return nil, err
}
if err := idtools.MkdirAllAndChown(path.Join(home, linkDir), 0700, currentID); err != nil {
return nil, err
}
d := &Driver{
home: home,
uidMaps: uidMaps,
gidMaps: gidMaps,
ctr: graphdriver.NewRefCounter(graphdriver.NewFsChecker(graphdriver.FsMagicFUSE)),
locker: locker.New(),
}
d.naiveDiff = graphdriver.NewNaiveDiffDriver(d, uidMaps, gidMaps)
return d, nil
}
func (d *Driver) String() string {
return driverName
}
// Status returns current driver information in a two dimensional string array.
func (d *Driver) Status() [][2]string {
return [][2]string{}
}
// GetMetadata returns metadata about the overlay driver such as the LowerDir,
// UpperDir, WorkDir, and MergeDir used to store data.
func (d *Driver) GetMetadata(id string) (map[string]string, error) {
dir := d.dir(id)
if _, err := os.Stat(dir); err != nil {
return nil, err
}
metadata := map[string]string{
"WorkDir": path.Join(dir, workDirName),
"MergedDir": path.Join(dir, mergedDirName),
"UpperDir": path.Join(dir, diffDirName),
}
lowerDirs, err := d.getLowerDirs(id)
if err != nil {
return nil, err
}
if len(lowerDirs) > 0 {
metadata["LowerDir"] = strings.Join(lowerDirs, ":")
}
return metadata, nil
}
// Cleanup any state created by overlay which should be cleaned when daemon
// is being shutdown. For now, we just have to unmount the bind mounted
// we had created.
func (d *Driver) Cleanup() error {
return mount.RecursiveUnmount(d.home)
}
// CreateReadWrite creates a layer that is writable for use as a container
// file system.
func (d *Driver) CreateReadWrite(id, parent string, opts *graphdriver.CreateOpts) error {
if opts != nil && len(opts.StorageOpt) != 0 {
return fmt.Errorf("--storage-opt is not supported")
}
return d.create(id, parent, opts)
}
// Create is used to create the upper, lower, and merge directories required for overlay fs for a given id.
// The parent filesystem is used to configure these directories for the overlay.
func (d *Driver) Create(id, parent string, opts *graphdriver.CreateOpts) (retErr error) {
if opts != nil && len(opts.StorageOpt) != 0 {
return fmt.Errorf("--storage-opt is not supported")
}
return d.create(id, parent, opts)
}
func (d *Driver) create(id, parent string, opts *graphdriver.CreateOpts) (retErr error) {
dir := d.dir(id)
rootUID, rootGID, err := idtools.GetRootUIDGID(d.uidMaps, d.gidMaps)
if err != nil {
return err
}
root := idtools.Identity{UID: rootUID, GID: rootGID}
dirID := idtools.Identity{
UID: rootUID,
GID: rootGID,
}
if err := idtools.MkdirAllAndChown(path.Dir(dir), 0710, dirID); err != nil {
return err
}
if err := idtools.MkdirAndChown(dir, 0710, dirID); err != nil {
return err
}
defer func() {
// Clean up on failure
if retErr != nil {
os.RemoveAll(dir)
}
}()
if opts != nil && len(opts.StorageOpt) > 0 {
return fmt.Errorf("--storage-opt is not supported")
}
if err := idtools.MkdirAndChown(path.Join(dir, diffDirName), 0755, root); err != nil {
return err
}
lid := overlayutils.GenerateID(idLength, logger)
if err := os.Symlink(path.Join("..", id, diffDirName), path.Join(d.home, linkDir, lid)); err != nil {
return err
}
// Write link id to link file
if err := os.WriteFile(path.Join(dir, "link"), []byte(lid), 0644); err != nil {
return err
}
// if no parent directory, done
if parent == "" {
return nil
}
if err := idtools.MkdirAndChown(path.Join(dir, workDirName), 0710, dirID); err != nil {
return err
}
if err := os.WriteFile(path.Join(d.dir(parent), "committed"), []byte{}, 0600); err != nil {
return err
}
lower, err := d.getLower(parent)
if err != nil {
return err
}
if lower != "" {
if err := os.WriteFile(path.Join(dir, lowerFile), []byte(lower), 0666); err != nil {
return err
}
}
return nil
}
func (d *Driver) getLower(parent string) (string, error) {
parentDir := d.dir(parent)
// Ensure parent exists
if _, err := os.Lstat(parentDir); err != nil {
return "", err
}
// Read Parent link fileA
parentLink, err := os.ReadFile(path.Join(parentDir, "link"))
if err != nil {
return "", err
}
lowers := []string{path.Join(linkDir, string(parentLink))}
parentLower, err := os.ReadFile(path.Join(parentDir, lowerFile))
if err == nil {
parentLowers := strings.Split(string(parentLower), ":")
lowers = append(lowers, parentLowers...)
}
if len(lowers) > maxDepth {
return "", errors.New("max depth exceeded")
}
return strings.Join(lowers, ":"), nil
}
func (d *Driver) dir(id string) string {
return path.Join(d.home, id)
}
func (d *Driver) getLowerDirs(id string) ([]string, error) {
var lowersArray []string
lowers, err := os.ReadFile(path.Join(d.dir(id), lowerFile))
if err == nil {
for _, s := range strings.Split(string(lowers), ":") {
lp, err := os.Readlink(path.Join(d.home, s))
if err != nil {
return nil, err
}
lowersArray = append(lowersArray, path.Clean(path.Join(d.home, linkDir, lp)))
}
} else if !os.IsNotExist(err) {
return nil, err
}
return lowersArray, nil
}
// Remove cleans the directories that are created for this id.
func (d *Driver) Remove(id string) error {
if id == "" {
return fmt.Errorf("refusing to remove the directories: id is empty")
}
d.locker.Lock(id)
defer d.locker.Unlock(id)
dir := d.dir(id)
lid, err := os.ReadFile(path.Join(dir, "link"))
if err == nil {
if len(lid) == 0 {
logger.Errorf("refusing to remove empty link for layer %v", id)
} else if err := os.RemoveAll(path.Join(d.home, linkDir, string(lid))); err != nil {
logger.Debugf("Failed to remove link: %v", err)
}
}
if err := system.EnsureRemoveAll(dir); err != nil && !os.IsNotExist(err) {
return err
}
return nil
}
// Get creates and mounts the required file system for the given id and returns the mount path.
func (d *Driver) Get(id, mountLabel string) (_ containerfs.ContainerFS, retErr error) {
d.locker.Lock(id)
defer d.locker.Unlock(id)
dir := d.dir(id)
if _, err := os.Stat(dir); err != nil {
return nil, err
}
diffDir := path.Join(dir, diffDirName)
lowers, err := os.ReadFile(path.Join(dir, lowerFile))
if err != nil {
// If no lower, just return diff directory
if os.IsNotExist(err) {
return containerfs.NewLocalContainerFS(diffDir), nil
}
return nil, err
}
mergedDir := path.Join(dir, mergedDirName)
if count := d.ctr.Increment(mergedDir); count > 1 {
return containerfs.NewLocalContainerFS(mergedDir), nil
}
defer func() {
if retErr != nil {
if c := d.ctr.Decrement(mergedDir); c <= 0 {
if unmounted := fusermountU(mergedDir); !unmounted {
if mntErr := unix.Unmount(mergedDir, 0); mntErr != nil {
logger.Errorf("error unmounting %v: %v", mergedDir, mntErr)
}
}
// Cleanup the created merged directory; see the comment in Put's rmdir
if rmErr := unix.Rmdir(mergedDir); rmErr != nil && !os.IsNotExist(rmErr) {
logger.Debugf("Failed to remove %s: %v: %v", id, rmErr, err)
}
}
}
}()
workDir := path.Join(dir, workDirName)
splitLowers := strings.Split(string(lowers), ":")
absLowers := make([]string, len(splitLowers))
for i, s := range splitLowers {
absLowers[i] = path.Join(d.home, s)
}
var readonly bool
if _, err := os.Stat(path.Join(dir, "committed")); err == nil {
readonly = true
} else if !os.IsNotExist(err) {
return nil, err
}
var opts string
if readonly {
opts = "lowerdir=" + diffDir + ":" + strings.Join(absLowers, ":")
} else {
opts = "lowerdir=" + strings.Join(absLowers, ":") + ",upperdir=" + diffDir + ",workdir=" + workDir
}
mountData := label.FormatMountLabel(opts, mountLabel)
mountTarget := mergedDir
rootUID, rootGID, err := idtools.GetRootUIDGID(d.uidMaps, d.gidMaps)
if err != nil {
return nil, err
}
if err := idtools.MkdirAndChown(mergedDir, 0700, idtools.Identity{UID: rootUID, GID: rootGID}); err != nil {
return nil, err
}
mountProgram := exec.Command(binary, "-o", mountData, mountTarget)
mountProgram.Dir = d.home
var b bytes.Buffer
mountProgram.Stderr = &b
if err = mountProgram.Run(); err != nil {
output := b.String()
if output == "" {
output = "<stderr empty>"
}
return nil, errors.Wrapf(err, "using mount program %s: %s", binary, output)
}
return containerfs.NewLocalContainerFS(mergedDir), nil
}
// Put unmounts the mount path created for the give id.
// It also removes the 'merged' directory to force the kernel to unmount the
// overlay mount in other namespaces.
func (d *Driver) Put(id string) error {
d.locker.Lock(id)
defer d.locker.Unlock(id)
dir := d.dir(id)
_, err := os.ReadFile(path.Join(dir, lowerFile))
if err != nil {
// If no lower, no mount happened and just return directly
if os.IsNotExist(err) {
return nil
}
return err
}
mountpoint := path.Join(dir, mergedDirName)
if count := d.ctr.Decrement(mountpoint); count > 0 {
return nil
}
if unmounted := fusermountU(mountpoint); !unmounted {
if err := unix.Unmount(mountpoint, unix.MNT_DETACH); err != nil {
logger.Debugf("Failed to unmount %s overlay: %s - %v", id, mountpoint, err)
}
}
// Remove the mountpoint here. Removing the mountpoint (in newer kernels)
// will cause all other instances of this mount in other mount namespaces
// to be unmounted. This is necessary to avoid cases where an overlay mount
// that is present in another namespace will cause subsequent mounts
// operations to fail with ebusy. We ignore any errors here because this may
// fail on older kernels which don't have
// torvalds/linux@8ed936b5671bfb33d89bc60bdcc7cf0470ba52fe applied.
if err := unix.Rmdir(mountpoint); err != nil && !os.IsNotExist(err) {
logger.Debugf("Failed to remove %s overlay: %v", id, err)
}
return nil
}
// Exists checks to see if the id is already mounted.
func (d *Driver) Exists(id string) bool {
_, err := os.Stat(d.dir(id))
return err == nil
}
// isParent determines whether the given parent is the direct parent of the
// given layer id
func (d *Driver) isParent(id, parent string) bool {
lowers, err := d.getLowerDirs(id)
if err != nil {
return false
}
if parent == "" && len(lowers) > 0 {
return false
}
parentDir := d.dir(parent)
var ld string
if len(lowers) > 0 {
ld = filepath.Dir(lowers[0])
}
if ld == "" && parent == "" {
return true
}
return ld == parentDir
}
// ApplyDiff applies the new layer into a root
func (d *Driver) ApplyDiff(id string, parent string, diff io.Reader) (size int64, err error) {
if !d.isParent(id, parent) {
return d.naiveDiff.ApplyDiff(id, parent, diff)
}
applyDir := d.getDiffPath(id)
logger.Debugf("Applying tar in %s", applyDir)
// Overlay doesn't need the parent id to apply the diff
if err := untar(diff, applyDir, &archive.TarOptions{
UIDMaps: d.uidMaps,
GIDMaps: d.gidMaps,
// Use AUFS whiteout format: https://github.com/containers/storage/blob/39a8d5ed9843844eafb5d2ba6e6a7510e0126f40/drivers/overlay/overlay.go#L1084-L1089
WhiteoutFormat: archive.AUFSWhiteoutFormat,
InUserNS: sys.RunningInUserNS(),
}); err != nil {
return 0, err
}
return directory.Size(context.TODO(), applyDir)
}
func (d *Driver) getDiffPath(id string) string {
dir := d.dir(id)
return path.Join(dir, diffDirName)
}
// DiffSize calculates the changes between the specified id
// and its parent and returns the size in bytes of the changes
// relative to its base filesystem directory.
func (d *Driver) DiffSize(id, parent string) (size int64, err error) {
return d.naiveDiff.DiffSize(id, parent)
}
// Diff produces an archive of the changes between the specified
// layer and its parent layer which may be "".
func (d *Driver) Diff(id, parent string) (io.ReadCloser, error) {
return d.naiveDiff.Diff(id, parent)
}
// Changes produces a list of changes between the specified layer and its
// parent layer. If parent is "", then all changes will be ADD changes.
func (d *Driver) Changes(id, parent string) ([]archive.Change, error) {
return d.naiveDiff.Changes(id, parent)
}
// fusermountU is from https://github.com/containers/storage/blob/39a8d5ed9843844eafb5d2ba6e6a7510e0126f40/drivers/overlay/overlay.go#L1016-L1040
func fusermountU(mountpoint string) (unmounted bool) {
// Attempt to unmount the FUSE mount using either fusermount or fusermount3.
// If they fail, fallback to unix.Unmount
for _, v := range []string{"fusermount3", "fusermount"} {
err := exec.Command(v, "-u", mountpoint).Run()
if err != nil && !os.IsNotExist(err) {
logrus.Debugf("Error unmounting %s with %s - %v", mountpoint, v, err)
}
if err == nil {
unmounted = true
break
}
}
// If fusermount|fusermount3 failed to unmount the FUSE file system, make sure all
// pending changes are propagated to the file system
if !unmounted {
fd, err := unix.Open(mountpoint, unix.O_DIRECTORY, 0)
if err == nil {
if err := unix.Syncfs(fd); err != nil {
logrus.Debugf("Error Syncfs(%s) - %v", mountpoint, err)
}
unix.Close(fd)
}
}
return
}
|