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
|
package device
import (
"context"
"errors"
"fmt"
"net"
"os"
"os/exec"
"path/filepath"
"slices"
"sort"
"strings"
"time"
"golang.org/x/sys/unix"
"github.com/lxc/incus/v6/internal/linux"
"github.com/lxc/incus/v6/internal/server/instance"
storageDrivers "github.com/lxc/incus/v6/internal/server/storage/drivers"
"github.com/lxc/incus/v6/shared/idmap"
"github.com/lxc/incus/v6/shared/revert"
"github.com/lxc/incus/v6/shared/subprocess"
"github.com/lxc/incus/v6/shared/util"
)
// RBDFormatPrefix is the prefix used in disk paths to identify RBD.
const RBDFormatPrefix = "rbd"
// RBDFormatSeparator is the field separate used in disk paths for RBD devices.
const RBDFormatSeparator = " "
// DiskParseRBDFormat parses an rbd formatted string, and returns the pool name, volume name, and map of options.
func DiskParseRBDFormat(rbd string) (string, string, map[string]string, error) {
// Remove and check the prefix.
prefix, rbd, _ := strings.Cut(rbd, RBDFormatSeparator)
if prefix != RBDFormatPrefix {
return "", "", nil, fmt.Errorf("Invalid rbd format, wrong prefix: %q", prefix)
}
// Split the path and options.
path, rawOpts, _ := strings.Cut(rbd, RBDFormatSeparator)
// Check for valid RBD path.
pool, volume, validPath := strings.Cut(path, "/")
if !validPath {
return "", "", nil, fmt.Errorf("Invalid rbd format, missing pool and/or volume: %q", path)
}
// Parse options.
opts := make(map[string]string)
for _, o := range strings.Split(rawOpts, ":") {
k, v, isValid := strings.Cut(o, "=")
if !isValid {
return "", "", nil, fmt.Errorf("Invalid rbd format, bad option: %q", o)
}
opts[k] = v
}
return pool, volume, opts, nil
}
// DiskGetRBDFormat returns a rbd formatted string with the given values.
func DiskGetRBDFormat(clusterName string, userName string, poolName string, volumeName string) string {
// Resolve any symlinks to config path.
confPath := fmt.Sprintf("/etc/ceph/%s.conf", clusterName)
target, err := filepath.EvalSymlinks(confPath)
if err == nil {
confPath = target
}
// Configuration values containing :, @, or = can be escaped with a leading \ character.
// According to https://docs.ceph.com/docs/hammer/rbd/qemu-rbd/#usage
optEscaper := strings.NewReplacer(":", `\:`, "@", `\@`, "=", `\=`)
opts := []string{
fmt.Sprintf("id=%s", optEscaper.Replace(userName)),
fmt.Sprintf("pool=%s", optEscaper.Replace(poolName)),
fmt.Sprintf("cluster=%s", optEscaper.Replace(clusterName)),
fmt.Sprintf("conf=%s", optEscaper.Replace(confPath)),
}
return fmt.Sprintf("%s%s%s/%s%s%s", RBDFormatPrefix, RBDFormatSeparator, optEscaper.Replace(poolName), optEscaper.Replace(volumeName), RBDFormatSeparator, strings.Join(opts, ":"))
}
// BlockFsDetect detects the type of block device.
func BlockFsDetect(dev string) (string, error) {
out, err := subprocess.RunCommand("blkid", "-s", "TYPE", "-o", "value", dev)
if err != nil {
return "", err
}
return strings.TrimSpace(out), nil
}
// IsBlockdev returns boolean indicating whether device is block type.
func IsBlockdev(path string) bool {
// Get a stat struct from the provided path.
stat := unix.Stat_t{}
err := unix.Stat(path, &stat)
if err != nil {
return false
}
// Check if it's a block device
if stat.Mode&unix.S_IFMT == unix.S_IFBLK {
return true
}
// Not a device
return false
}
// DiskMount mounts a disk device.
func DiskMount(srcPath string, dstPath string, recursive bool, propagation string, mountOptions []string, fsName string) error {
var err error
flags, mountOptionsStr := linux.ResolveMountOptions(mountOptions)
var readonly bool
if slices.Contains(mountOptions, "ro") {
readonly = true
}
// Detect the filesystem
if fsName == "none" {
flags |= unix.MS_BIND
}
if propagation != "" {
switch propagation {
case "private":
flags |= unix.MS_PRIVATE
case "shared":
flags |= unix.MS_SHARED
case "slave":
flags |= unix.MS_SLAVE
case "unbindable":
flags |= unix.MS_UNBINDABLE
case "rprivate":
flags |= unix.MS_PRIVATE | unix.MS_REC
case "rshared":
flags |= unix.MS_SHARED | unix.MS_REC
case "rslave":
flags |= unix.MS_SLAVE | unix.MS_REC
case "runbindable":
flags |= unix.MS_UNBINDABLE | unix.MS_REC
default:
return fmt.Errorf("Invalid propagation mode %q", propagation)
}
}
if recursive {
flags |= unix.MS_REC
}
// Mount the filesystem
err = unix.Mount(srcPath, dstPath, fsName, uintptr(flags), mountOptionsStr)
if err != nil {
return fmt.Errorf("Unable to mount %q at %q with filesystem %q: %w", srcPath, dstPath, fsName, err)
}
// Remount bind mounts in readonly mode if requested
if readonly && flags&unix.MS_BIND == unix.MS_BIND {
flags = unix.MS_RDONLY | unix.MS_BIND | unix.MS_REMOUNT
err = unix.Mount("", dstPath, fsName, uintptr(flags), "")
if err != nil {
return fmt.Errorf("Unable to mount %q in readonly mode: %w", dstPath, err)
}
}
flags = unix.MS_REC | unix.MS_SLAVE
err = unix.Mount("", dstPath, "", uintptr(flags), "")
if err != nil {
return fmt.Errorf("Unable to make mount %q private: %w", dstPath, err)
}
return nil
}
// DiskMountClear unmounts and removes the mount path used for disk shares.
func DiskMountClear(mntPath string) error {
if util.PathExists(mntPath) {
if linux.IsMountPoint(mntPath) {
err := storageDrivers.TryUnmount(mntPath, 0)
if err != nil {
return fmt.Errorf("Failed unmounting %q: %w", mntPath, err)
}
}
err := os.Remove(mntPath)
if err != nil {
return fmt.Errorf("Failed removing %q: %w", mntPath, err)
}
}
return nil
}
func diskCephRbdMap(clusterName string, userName string, poolName string, volumeName string) (string, error) {
devPath, err := subprocess.RunCommand(
"rbd",
"--id", userName,
"--cluster", clusterName,
"--pool", poolName,
"map",
volumeName)
if err != nil {
return "", err
}
idx := strings.Index(devPath, "/dev/rbd")
if idx < 0 {
return "", errors.New("Failed to detect mapped device path")
}
devPath = devPath[idx:]
return strings.TrimSpace(devPath), nil
}
func diskCephRbdUnmap(deviceName string) error {
unmapImageName := deviceName
busyCount := 0
again:
_, err := subprocess.RunCommand(
"rbd",
"unmap",
unmapImageName)
if err != nil {
var runError subprocess.RunError
if errors.As(err, &runError) {
var exitError *exec.ExitError
if errors.As(runError.Unwrap(), &exitError) {
if exitError.ExitCode() == 22 {
// EINVAL (already unmapped)
return nil
}
if exitError.ExitCode() == 16 {
// EBUSY (currently in use)
busyCount++
if busyCount == 10 {
return err
}
// Wait a second an try again
time.Sleep(time.Second)
goto again
}
}
}
return err
}
goto again
}
// diskCephfsOptions returns the mntSrcPath and fsOptions to use for mounting a cephfs share.
func diskCephfsOptions(clusterName string, userName string, fsName string, fsPath string) (string, []string, error) {
// Get the FSID.
fsid, err := storageDrivers.CephFsid(clusterName, userName)
if err != nil {
return "", nil, err
}
// Get the monitor list.
monAddresses, err := storageDrivers.CephMonitors(clusterName, userName)
if err != nil {
return "", nil, err
}
// Get the keyring entry.
secret, err := storageDrivers.CephKeyring(clusterName, userName)
if err != nil {
return "", nil, err
}
srcPath, fsOptions := storageDrivers.CephBuildMount(
userName,
secret,
fsid,
monAddresses,
fsName,
fsPath,
)
return srcPath, fsOptions, nil
}
// DiskVMVirtiofsdStart starts a new virtiofsd process.
// If the idmaps slice is supplied then the proxy process is run inside a user namespace using the supplied maps.
// Returns UnsupportedError error if the host system or instance does not support virtiosfd, returns normal error
// type if process cannot be started for other reasons.
// Returns revert function and listener file handle on success.
func DiskVMVirtiofsdStart(execPath string, inst instance.Instance, socketPath string, pidPath string, logPath string, sharePath string, idmaps []idmap.Entry, cacheOption string) (func(), net.Listener, error) {
reverter := revert.New()
defer reverter.Fail()
if !filepath.IsAbs(sharePath) {
return nil, nil, fmt.Errorf("Share path not absolute: %q", sharePath)
}
// Remove old socket if needed.
_ = os.Remove(socketPath)
// Locate virtiofsd.
cmd, err := exec.LookPath("virtiofsd")
if err != nil {
if util.PathExists("/usr/lib/qemu/virtiofsd") {
cmd = "/usr/lib/qemu/virtiofsd"
} else if util.PathExists("/usr/libexec/virtiofsd") {
cmd = "/usr/libexec/virtiofsd"
} else if util.PathExists("/usr/lib/virtiofsd") {
cmd = "/usr/lib/virtiofsd"
}
}
if cmd == "" {
return nil, nil, ErrMissingVirtiofsd
}
if util.IsTrue(inst.ExpandedConfig()["migration.stateful"]) {
return nil, nil, UnsupportedError{"Stateful migration unsupported"}
}
if util.IsTrue(inst.ExpandedConfig()["security.sev"]) || util.IsTrue(inst.ExpandedConfig()["security.sev.policy.es"]) {
return nil, nil, UnsupportedError{"SEV unsupported"}
}
// Trickery to handle paths > 107 chars.
socketFileDir, err := os.Open(filepath.Dir(socketPath))
if err != nil {
return nil, nil, err
}
defer func() { _ = socketFileDir.Close() }()
socketFile := fmt.Sprintf("/proc/self/fd/%d/%s", socketFileDir.Fd(), filepath.Base(socketPath))
listener, err := net.Listen("unix", socketFile)
if err != nil {
return nil, nil, fmt.Errorf("Failed to create unix listener for virtiofsd: %w", err)
}
reverter.Add(func() {
_ = listener.Close()
_ = os.Remove(socketPath)
})
unixListener, ok := listener.(*net.UnixListener)
if !ok {
return nil, nil, errors.New("Failed getting UnixListener for virtiofsd")
}
unixFile, err := unixListener.File()
if err != nil {
return nil, nil, fmt.Errorf("Failed to getting unix listener file for virtiofsd: %w", err)
}
defer func() { _ = unixFile.Close() }()
switch cacheOption {
case "metadata":
cacheOption = "metadata"
case "unsafe":
cacheOption = "always"
default:
cacheOption = "never"
}
// Start the virtiofsd process in non-daemon mode.
args := []string{"--fd=3", fmt.Sprintf("--cache=%s", cacheOption), fmt.Sprintf("--shared-dir=%s", sharePath)}
if len(idmaps) > 0 {
idmapSet := &idmap.Set{Entries: idmaps}
sort.Sort(idmapSet)
var lastUID int64
var lastGID int64
for _, entry := range idmapSet.Entries {
if entry.IsUID {
args = append(args, fmt.Sprintf("--translate-uid=map:%d:%d:%d", entry.NSID, entry.HostID, entry.MapRange))
args = append(args, fmt.Sprintf("--translate-uid=forbid-guest:%d:%d", lastUID, entry.NSID-lastUID))
lastUID = entry.NSID + entry.MapRange
}
if entry.IsGID {
args = append(args, fmt.Sprintf("--translate-gid=map:%d:%d:%d", entry.NSID, entry.HostID, entry.MapRange))
args = append(args, fmt.Sprintf("--translate-gid=forbid-guest:%d:%d", lastGID, entry.NSID-lastGID))
lastGID = entry.NSID + entry.MapRange
}
}
if lastUID < 4294967295 {
args = append(args, fmt.Sprintf("--translate-uid=forbid-guest:%d:%d", lastUID, 4294967295-lastUID))
}
if lastGID < 4294967295 {
args = append(args, fmt.Sprintf("--translate-gid=forbid-guest:%d:%d", lastGID, 4294967295-lastGID))
}
} else {
args = append(args, "--posix-acl")
}
proc, err := subprocess.NewProcess(cmd, args, logPath, logPath)
if err != nil {
return nil, nil, err
}
err = proc.StartWithFiles(context.Background(), []*os.File{unixFile})
if err != nil {
return nil, nil, fmt.Errorf("Failed to start virtiofsd: %w", err)
}
reverter.Add(func() { _ = proc.Stop() })
err = proc.Save(pidPath)
if err != nil {
return nil, nil, fmt.Errorf("Failed to save virtiofsd state: %w", err)
}
cleanup := reverter.Clone().Fail
reverter.Success()
return cleanup, listener, err
}
// DiskVMVirtiofsdStop stops an existing virtiofsd process and cleans up.
func DiskVMVirtiofsdStop(socketPath string, pidPath string) error {
if util.PathExists(pidPath) {
proc, err := subprocess.ImportProcess(pidPath)
if err != nil {
return err
}
err = proc.Stop()
// The virtiofsd process will terminate automatically once the VM has stopped.
// We therefore should only return an error if it's still running and fails to stop.
if err != nil && !errors.Is(err, subprocess.ErrNotRunning) {
return err
}
// Remove PID file if needed.
err = os.Remove(pidPath)
if err != nil && !errors.Is(err, os.ErrNotExist) {
return fmt.Errorf("Failed to remove PID file: %w", err)
}
}
// Remove socket file if needed.
err := os.Remove(socketPath)
if err != nil && !errors.Is(err, os.ErrNotExist) {
return fmt.Errorf("Failed to remove socket file: %w", err)
}
return nil
}
|