/* SPDX-License-Identifier: LGPL-2.1-or-later */ #include #include #include #include #include #include #include #include #include "sd-json.h" #include "sd-path.h" #include "sd-varlink.h" #include "alloc-util.h" #include "blockdev-util.h" #include "btrfs-util.h" #include "bus-get-properties.h" #include "chase.h" #include "chattr-util.h" #include "copy.h" #include "devnum-util.h" #include "dirent-util.h" #include "discover-image.h" #include "dissect-image.h" #include "env-file.h" #include "env-util.h" #include "extension-util.h" #include "fd-util.h" #include "fs-util.h" #include "hashmap.h" #include "hostname-setup.h" #include "id128-util.h" #include "label-util.h" #include "lock-util.h" #include "log.h" #include "loop-util.h" #include "mountpoint-util.h" #include "mstack.h" #include "namespace-util.h" #include "nsresource.h" #include "nulstr-util.h" #include "os-util.h" #include "path-lookup.h" #include "path-util.h" #include "process-util.h" #include "recurse-dir.h" #include "rm-rf.h" #include "runtime-scope.h" #include "stat-util.h" #include "string-table.h" #include "string-util.h" #include "strv.h" #include "time-util.h" #include "uid-classification.h" #include "vpick.h" #include "xattr-util.h" const char* const image_search_path[_IMAGE_CLASS_MAX] = { [IMAGE_MACHINE] = "/etc/machines\0" /* only place symlinks here */ "/run/machines\0" /* and here too */ "/var/lib/machines\0" /* the main place for images */ "/var/lib/container\0" /* legacy */ "/usr/local/lib/machines\0" "/usr/lib/machines\0", [IMAGE_PORTABLE] = "/etc/portables\0" /* only place symlinks here */ "/run/portables\0" /* and here too */ "/var/lib/portables\0" /* the main place for images */ "/usr/local/lib/portables\0" "/usr/lib/portables\0", /* Note that we don't allow storing extensions under /usr/, unlike with other image types. That's * because extension images are supposed to extend /usr/, so you get into recursive races, especially * with directory-based extensions, as the kernel's OverlayFS explicitly checks for this and errors * out with -ELOOP if it finds that a lowerdir= is a child of another lowerdir=. */ [IMAGE_SYSEXT] = "/etc/extensions\0" /* only place symlinks here */ "/run/extensions\0" /* and here too */ "/var/lib/extensions\0", /* the main place for images */ [IMAGE_CONFEXT] = "/run/confexts\0" /* only place symlinks here */ "/var/lib/confexts\0" /* the main place for images */ "/usr/local/lib/confexts\0" "/usr/lib/confexts\0", }; /* Inside the initrd, use a slightly different set of search path (i.e. include .extra/sysext/, * /.extra/global_sysext, .extra/confext/, and /.extra/global_confext in extension search dir) */ static const char* const image_search_path_initrd[_IMAGE_CLASS_MAX] = { /* (entries that aren't listed here will get the same search path as for the non initrd-case) */ [IMAGE_SYSEXT] = "/etc/extensions\0" /* only place symlinks here */ "/run/extensions\0" /* and here too */ "/var/lib/extensions\0" /* the main place for images */ "/.extra/sysext\0" /* put sysext (per-UKI and global) picked up by systemd-stub */ "/.extra/global_sysext\0", /* last, since not trusted */ [IMAGE_CONFEXT] = "/run/confexts\0" /* only place symlinks here */ "/var/lib/confexts\0" /* the main place for images */ "/usr/local/lib/confexts\0" "/.extra/confext\0" /* put confext (per-UKI and global) picked up by systemd-stub */ "/.extra/global_confext\0", /* last, since not trusted. */ }; static const char* image_class_suffix_table[_IMAGE_CLASS_MAX] = { [IMAGE_SYSEXT] = ".sysext", [IMAGE_CONFEXT] = ".confext", }; DEFINE_PRIVATE_STRING_TABLE_LOOKUP_TO_STRING(image_class_suffix, ImageClass); static const char *const image_root_table[_IMAGE_CLASS_MAX] = { [IMAGE_MACHINE] = "/var/lib/machines", [IMAGE_PORTABLE] = "/var/lib/portables", [IMAGE_SYSEXT] = "/var/lib/extensions", [IMAGE_CONFEXT] = "/var/lib/confexts", }; DEFINE_PRIVATE_STRING_TABLE_LOOKUP_TO_STRING(image_root, ImageClass); static const char *const image_root_runtime_table[_IMAGE_CLASS_MAX] = { [IMAGE_MACHINE] = "/run/machines", [IMAGE_PORTABLE] = "/run/portables", [IMAGE_SYSEXT] = "/run/extensions", [IMAGE_CONFEXT] = "/run/confexts", }; DEFINE_PRIVATE_STRING_TABLE_LOOKUP_TO_STRING(image_root_runtime, ImageClass); static const char *const image_dirname_table[_IMAGE_CLASS_MAX] = { [IMAGE_MACHINE] = "machines", [IMAGE_PORTABLE] = "portables", [IMAGE_SYSEXT] = "extensions", [IMAGE_CONFEXT] = "confexts", }; static const char auxiliary_suffixes_nulstr[] = ".nspawn\0" ".oci-config\0" ".roothash\0" ".roothash.p7s\0" ".usrhash\0" ".usrhash.p7s\0" ".verity\0" ".raw.tpmstate\0" ".raw.efinvramstate\0"; DEFINE_PRIVATE_STRING_TABLE_LOOKUP_TO_STRING(image_dirname, ImageClass); static Image* image_free(Image *i) { assert(i); free(i->name); free(i->path); free(i->fh); free(i->hostname); strv_free(i->machine_info); strv_free(i->os_release); strv_free(i->sysext_release); strv_free(i->confext_release); return mfree(i); } DEFINE_TRIVIAL_REF_UNREF_FUNC(Image, image, image_free); DEFINE_HASH_OPS_WITH_VALUE_DESTRUCTOR(image_hash_ops, char, string_hash_func, string_compare_func, Image, image_unref); static char** image_settings_path(Image *image, RuntimeScope scope) { _cleanup_strv_free_ char **l = NULL; _cleanup_free_ char *fn = NULL; size_t i = 0; int r; assert(image); l = new0(char*, 5); if (!l) return NULL; fn = strjoin(image->name, ".nspawn"); if (!fn) return NULL; static const uint64_t system_locations[] = { SD_PATH_SYSTEM_CONFIGURATION, SD_PATH_SYSTEM_RUNTIME, SD_PATH_SYSTEM_LIBRARY_PRIVATE, _SD_PATH_INVALID }; static const uint64_t user_locations[] = { SD_PATH_USER_CONFIGURATION, SD_PATH_USER_RUNTIME, SD_PATH_USER_LIBRARY_PRIVATE, _SD_PATH_INVALID }; const uint64_t *locations; switch (scope) { case RUNTIME_SCOPE_SYSTEM: locations = system_locations; break; case RUNTIME_SCOPE_USER: locations = user_locations; break; default: assert_not_reached(); } for (size_t k = 0; locations[k] != _SD_PATH_INVALID; k++) { _cleanup_free_ char *s = NULL; r = sd_path_lookup(locations[k], "systemd/nspawn", &s); if (r == -ENXIO) continue; if (r < 0) return NULL; l[i] = path_join(s, fn); if (!l[i]) return NULL; i++; } r = file_in_same_dir(image->path, fn, l + i); if (r == -ENOMEM) return NULL; if (r < 0) log_debug_errno(r, "Failed to generate .nspawn settings path from image path, ignoring: %m"); strv_uniq(l); return TAKE_PTR(l); } static int image_auxiliary_path(Image *image, const char *suffix, char **ret) { assert(image); assert(suffix); _cleanup_free_ char *fn = strjoin(image->name, suffix); if (!fn) return -ENOMEM; return file_in_same_dir(image->path, fn, ret); } static int image_new( ImageType t, ImageClass c, const char *pretty, const char *path, bool read_only, usec_t crtime, usec_t mtime, struct file_handle *fh, uint64_t on_mount_id, uint64_t inode, Image **ret) { _cleanup_(image_unrefp) Image *i = NULL; assert(t >= 0); assert(t < _IMAGE_TYPE_MAX); assert(pretty); assert(path); assert(ret); i = new(Image, 1); if (!i) return -ENOMEM; *i = (Image) { .n_ref = 1, .type = t, .class = c, .read_only = read_only, .crtime = crtime, .mtime = mtime, .on_mount_id = on_mount_id, .inode = inode, .usage = UINT64_MAX, .usage_exclusive = UINT64_MAX, .limit = UINT64_MAX, .limit_exclusive = UINT64_MAX, }; if (fh) { i->fh = file_handle_dup(fh); if (!i->fh) return -ENOMEM; } i->name = strdup(pretty); if (!i->name) return -ENOMEM; i->path = strdup(path); if (!i->path) return -ENOMEM; path_simplify(i->path); *ret = TAKE_PTR(i); return 0; } static int extract_image_basename( const char *path, const char *class_suffix, /* e.g. ".sysext" (this is an optional suffix) */ char **format_suffixes, /* e.g. ".raw" (one of these will be required) */ char **ret_basename, char **ret_suffix) { _cleanup_free_ char *name = NULL, *suffix = NULL; int r; assert(path); r = path_extract_filename(path, &name); if (r < 0) return r; if (format_suffixes) { char *e = endswith_strv(name, format_suffixes); if (!e) /* Format suffix is required */ return -EINVAL; if (ret_suffix) { suffix = strdup(e); if (!suffix) return -ENOMEM; } *e = 0; } if (class_suffix) { char *e = endswith(name, class_suffix); if (e) { /* Class suffix is optional */ if (ret_suffix) { _cleanup_free_ char *j = strjoin(e, suffix); if (!j) return -ENOMEM; free_and_replace(suffix, j); } *e = 0; } } if (!image_name_is_valid(name)) return -EINVAL; if (ret_suffix) *ret_suffix = TAKE_PTR(suffix); if (ret_basename) *ret_basename = TAKE_PTR(name); return 0; } static int image_update_quota(Image *i, int fd) { _cleanup_close_ int fd_close = -EBADF; int r; assert(i); if (image_is_vendor(i) || image_is_host(i)) return -EROFS; if (i->type != IMAGE_SUBVOLUME) return -EOPNOTSUPP; if (fd < 0) { fd_close = open(i->path, O_CLOEXEC|O_DIRECTORY); if (fd_close < 0) return -errno; fd = fd_close; } else { /* Convert from O_PATH to proper fd, if needed */ fd = fd_reopen_condition(fd, O_CLOEXEC|O_DIRECTORY, O_PATH, &fd_close); if (fd < 0) return fd; } r = btrfs_quota_scan_ongoing(fd); if (r < 0) return r; if (r > 0) return 0; BtrfsQuotaInfo quota; r = btrfs_subvol_get_subtree_quota_fd(fd, 0, "a); if (r < 0) return r; i->usage = quota.referenced; i->usage_exclusive = quota.exclusive; i->limit = quota.referenced_max; i->limit_exclusive = quota.exclusive_max; return 1; } static int image_make( ImageClass c, const char *pretty, int fd, /* O_PATH fd */ const char *path, const struct stat *st, Image **ret) { _cleanup_free_ char *pretty_buffer = NULL; bool read_only; int r; assert(path); assert(path_is_absolute(path)); /* We explicitly *do* follow symlinks here, since we want to allow symlinking trees, raw files and block * devices into /var/lib/machines/, and treat them normally. * Note that if the caller does not want to follow symlinks (and does not care about symlink races) * then the caller should pass in a resolved path and an fd. * * This function returns -ENOENT if we can't find the image after all, and -EMEDIUMTYPE if it's not a file we * recognize. */ _cleanup_close_ int _fd = -EBADF; if (fd < 0) { /* If we didn't get an fd passed in, then let's pin it via O_PATH now */ _fd = open(path, O_PATH|O_CLOEXEC); if (_fd < 0) return -errno; fd = _fd; st = NULL; /* refresh stat() data now that we have the inode pinned */ } struct stat stbuf; if (!st) { if (fstat(fd, &stbuf) < 0) return -errno; st = &stbuf; } read_only = path_startswith(path, "/usr") || (faccessat(fd, "", W_OK, AT_EACCESS|AT_EMPTY_PATH) < 0 && errno == EROFS); _cleanup_free_ struct file_handle *fh = NULL; uint64_t on_mount_id; int _mnt_id; /* The fallback is required for CentOS 9 compatibility when working on a directory located on an * overlayfs. */ r = name_to_handle_at_try_fid(fd, /* path= */ NULL, &fh, &_mnt_id, &on_mount_id, AT_EMPTY_PATH); if (r < 0) { if (is_name_to_handle_at_fatal_error(r)) return r; r = path_get_unique_mnt_id_at(fd, /* path= */ NULL, &on_mount_id); if (r < 0) { if (!ERRNO_IS_NEG_NOT_SUPPORTED(r) && r != -EUNATCH) return r; int on_mount_id_fallback = -1; r = path_get_mnt_id_at(fd, /* path= */ NULL, &on_mount_id_fallback); if (r < 0) return r; on_mount_id = on_mount_id_fallback; } } else if (r == 0) on_mount_id = _mnt_id; if (S_ISDIR(st->st_mode)) { if (!ret) return 0; if (endswith(path, ".mstack")) { usec_t crtime = 0; r = fd_getcrtime(fd, &crtime); if (r < 0) log_debug_errno(r, "Unable to read creation time of '%s', ignoring: %m", path); if (!pretty) { r = extract_image_basename( path, image_class_suffix_to_string(c), STRV_MAKE(".mstack"), &pretty_buffer, /* ret_suffix= */ NULL); if (r < 0) return r; pretty = pretty_buffer; } _cleanup_(mstack_freep) MStack *mstack = NULL; r = mstack_load(path, fd, &mstack); if (r < 0) { log_debug_errno(r, "Failed to load mstack '%s', ignoring: %m", path); read_only = true; } else if (!read_only) { r = mstack_is_read_only(mstack); if (r < 0) log_debug_errno(r, "Failed to determine if mstack '%s' is read-only, assuming it is: %m", path); read_only = r != 0; } r = image_new(IMAGE_MSTACK, c, pretty, path, read_only, crtime, /* mtime= */ 0, fh, on_mount_id, (uint64_t) st->st_ino, ret); if (r < 0) return r; if (mstack) { r = mstack_is_foreign_uid_owned(mstack); if (r < 0) log_debug_errno(r, "Failed to determine if mstack '%s' is foreign UID owned, assuming it is not: %m", path); if (r > 0) (*ret)->foreign_uid_owned = true; } return 0; } if (!pretty) { r = extract_image_basename( path, image_class_suffix_to_string(c), /* format_suffixes= */ NULL, &pretty_buffer, /* ret_suffix= */ NULL); if (r < 0) return r; pretty = pretty_buffer; } if (btrfs_might_be_subvol(st)) { r = fd_is_fs_type(fd, BTRFS_SUPER_MAGIC); if (r < 0) return r; if (r > 0) { BtrfsSubvolInfo info; /* It's a btrfs subvolume */ r = btrfs_subvol_get_info_fd(fd, 0, &info); if (r < 0) return r; r = image_new(IMAGE_SUBVOLUME, c, pretty, path, info.read_only || read_only, info.otime, info.ctime, fh, on_mount_id, (uint64_t) st->st_ino, ret); if (r < 0) return r; (*ret)->foreign_uid_owned = uid_is_foreign(st->st_uid); (void) image_update_quota(*ret, fd); return 0; } } /* Get directory creation time (not available everywhere, but that's OK */ usec_t crtime = 0; (void) fd_getcrtime(fd, &crtime); /* If the IMMUTABLE bit is set, we consider the directory read-only. Since the ioctl is not * supported everywhere we ignore failures. */ unsigned file_attr = 0; (void) read_attr_fd(fd, &file_attr); /* It's just a normal directory. */ r = image_new(IMAGE_DIRECTORY, c, pretty, path, read_only || (file_attr & FS_IMMUTABLE_FL), crtime, 0, /* we don't use mtime of stat() here, since it's not the time of last change of the tree, but only of the top-level dir */ fh, on_mount_id, (uint64_t) st->st_ino, ret); if (r < 0) return r; (*ret)->foreign_uid_owned = uid_is_foreign(st->st_uid); return 0; } else if (S_ISREG(st->st_mode) && endswith(path, ".raw")) { usec_t crtime = 0; /* It's a RAW disk image */ if (!ret) return 0; (void) fd_getcrtime(fd, &crtime); if (!pretty) { r = extract_image_basename( path, image_class_suffix_to_string(c), STRV_MAKE(".raw"), &pretty_buffer, /* ret_suffix= */ NULL); if (r < 0) return r; pretty = pretty_buffer; } r = image_new(IMAGE_RAW, c, pretty, path, !(st->st_mode & 0222) || read_only, crtime, timespec_load(&st->st_mtim), fh, on_mount_id, (uint64_t) st->st_ino, ret); if (r < 0) return r; (*ret)->usage = (*ret)->usage_exclusive = st->st_blocks * 512; (*ret)->limit = (*ret)->limit_exclusive = st->st_size; return 0; } else if (S_ISBLK(st->st_mode)) { uint64_t size = UINT64_MAX; /* A block device */ if (!ret) return 0; if (!pretty) { r = extract_image_basename( path, /* class_suffix= */ NULL, /* format_suffix= */ NULL, &pretty_buffer, /* ret_suffix= */ NULL); if (r < 0) return r; pretty = pretty_buffer; } _cleanup_close_ int block_fd = fd_reopen(fd, O_RDONLY|O_NONBLOCK|O_CLOEXEC|O_NOCTTY); if (block_fd < 0) log_debug_errno(errno, "Failed to open block device '%s', ignoring: %m", path); else { if (!read_only) { int state = 0; if (ioctl(block_fd, BLKROGET, &state) < 0) log_debug_errno(errno, "Failed to issue BLKROGET on device '%s', ignoring: %m", path); else if (state) read_only = true; } r = blockdev_get_device_size(block_fd, &size); if (r < 0) log_debug_errno(r, "Failed to issue BLKGETSIZE64 on device '%s', ignoring: %m", path); block_fd = safe_close(block_fd); } r = image_new(IMAGE_BLOCK, c, pretty, path, !(st->st_mode & 0222) || read_only, 0, 0, fh, on_mount_id, (uint64_t) st->st_ino, ret); if (r < 0) return r; if (!IN_SET(size, 0, UINT64_MAX)) (*ret)->usage = (*ret)->usage_exclusive = (*ret)->limit = (*ret)->limit_exclusive = size; return 0; } return -EMEDIUMTYPE; } static int pick_image_search_path( RuntimeScope scope, ImageClass class, const char *root, char ***ret) { int r; assert(scope < _RUNTIME_SCOPE_MAX && scope != RUNTIME_SCOPE_GLOBAL); assert(class < _IMAGE_CLASS_MAX); assert(ret); if (class < 0) { *ret = NULL; return 0; } if (scope < 0) { _cleanup_strv_free_ char **a = NULL, **b = NULL; r = pick_image_search_path(RUNTIME_SCOPE_USER, class, root, &a); if (r < 0) return r; r = pick_image_search_path(RUNTIME_SCOPE_SYSTEM, class, root, &b); if (r < 0) return r; r = strv_extend_strv(&a, b, /* filter_duplicates= */ false); if (r < 0) return r; *ret = TAKE_PTR(a); return 0; } switch (scope) { case RUNTIME_SCOPE_SYSTEM: { const char *ns; bool is_initrd; r = chase_and_access("/etc/initrd-release", root, CHASE_PREFIX_ROOT, F_OK, /* ret_path= */ NULL); if (r < 0 && r != -ENOENT) return r; is_initrd = r >= 0; /* Use the initrd search path if there is one, otherwise use the common one */ ns = is_initrd && image_search_path_initrd[class] ? image_search_path_initrd[class] : image_search_path[class]; if (!ns) break; _cleanup_strv_free_ char **search = strv_split_nulstr(ns); if (!search) return -ENOMEM; *ret = TAKE_PTR(search); return 0; } case RUNTIME_SCOPE_USER: { if (!IN_SET(class, IMAGE_MACHINE, IMAGE_PORTABLE)) break; static const uint64_t dirs[] = { SD_PATH_USER_RUNTIME, SD_PATH_USER_STATE_PRIVATE, SD_PATH_USER_LIBRARY_PRIVATE, }; _cleanup_strv_free_ char **search = NULL; FOREACH_ELEMENT(d, dirs) { _cleanup_free_ char *p = NULL; r = sd_path_lookup(*d, image_dirname_to_string(class), &p); if (r == -ENXIO) /* No XDG_RUNTIME_DIR set */ continue; if (r < 0) return r; r = strv_consume(&search, TAKE_PTR(p)); if (r < 0) return r; } *ret = TAKE_PTR(search); return 0; } default: assert_not_reached(); } *ret = NULL; return 0; } static char** make_possible_filenames(ImageClass class, const char *image_name) { _cleanup_strv_free_ char **l = NULL; assert(image_name); FOREACH_STRING(v_suffix, "", ".v") FOREACH_STRING(format_suffix, "", ".raw", ".mstack") { _cleanup_free_ char *j = NULL; const char *class_suffix; class_suffix = image_class_suffix_to_string(class); if (class_suffix) { j = strjoin(image_name, class_suffix, format_suffix, v_suffix); if (!j) return NULL; if (strv_consume(&l, TAKE_PTR(j)) < 0) return NULL; } j = strjoin(image_name, format_suffix, v_suffix); if (!j) return NULL; if (strv_consume(&l, TAKE_PTR(j)) < 0) return NULL; } return TAKE_PTR(l); } int image_find(RuntimeScope scope, ImageClass class, const char *name, const char *root, Image **ret) { int r; assert(scope < _RUNTIME_SCOPE_MAX && scope != RUNTIME_SCOPE_GLOBAL); assert(class >= 0); assert(class < _IMAGE_CLASS_MAX); assert(name); /* There are no images with invalid names */ if (!image_name_is_valid(name)) return -ENOENT; _cleanup_strv_free_ char **names = make_possible_filenames(class, name); if (!names) return -ENOMEM; _cleanup_close_ int rfd = XAT_FDROOT; /* We only expect absolute paths */ if (root) { rfd = open(root, O_CLOEXEC|O_DIRECTORY|O_PATH); if (rfd < 0) return log_debug_errno(errno, "Failed to open root directory '%s': %m", root); } _cleanup_strv_free_ char **search = NULL; r = pick_image_search_path(scope, class, root, &search); if (r < 0) return r; STRV_FOREACH(s, search) { _cleanup_closedir_ DIR *d = NULL; _cleanup_free_ char *search_path = NULL; r = chase_and_opendirat(rfd, rfd, *s, /* chase_flags= */ 0, &search_path, &d); if (r == -ENOENT) continue; if (r < 0) return r; STRV_FOREACH(n, names) { const char *fname = *n; _cleanup_free_ char *fname_path = NULL, *chased_path = NULL, *resolved_file = NULL; _cleanup_close_ int fd = -EBADF; fname_path = path_join(search_path, fname); if (!fname_path) return -ENOMEM; /* Follow symlinks only inside given root */ r = chaseat(rfd, rfd, fname_path, /* flags= */ 0, &chased_path, &fd); if (r == -ENOENT) continue; if (r < 0) return r; r = chaseat_prefix_root(chased_path, root, &resolved_file); if (r < 0) return r; struct stat st; if (fstat(fd, &st) < 0) return -errno; if (endswith(fname, ".raw")) { if (!S_ISREG(st.st_mode)) { log_debug("Ignoring non-regular file '%s' with .raw suffix.", fname); continue; } } else if (endswith(fname, ".mstack")) { if (!S_ISDIR(st.st_mode)) { log_debug("Ignoring non-directory '%s' with .mstack suffix.", fname); continue; } } else if (endswith(fname, ".v")) { if (!S_ISDIR(st.st_mode)) { log_debug("Ignoring non-directory file '%s' with .v suffix.", fname); continue; } _cleanup_free_ char *suffix = NULL; suffix = strdup(ASSERT_PTR(startswith(fname, name))); if (!suffix) return -ENOMEM; *ASSERT_PTR(endswith(suffix, ".v")) = 0; PickFilter filter = { .type_mask = endswith(suffix, ".raw") ? (UINT32_C(1) << DT_REG) | (UINT32_C(1) << DT_BLK) : (UINT32_C(1) << DT_DIR), .basename = name, .architecture = _ARCHITECTURE_INVALID, .suffix = suffix, }; _cleanup_(pick_result_done) PickResult result = PICK_RESULT_NULL; r = path_pick(root, rfd, fname_path, /* This has to be the unresolved entry with the .v suffix */ &filter, /* n_filters= */ 1, PICK_ARCHITECTURE|PICK_TRIES|PICK_RESOLVE, &result); if (r < 0) { log_debug_errno(r, "Failed to pick versioned image on '%s%s', skipping: %m", empty_to_root(root), skip_leading_slash(fname_path)); continue; } if (!result.path) { log_debug("Found versioned directory '%s%s', without matching entry, skipping.", empty_to_root(root), skip_leading_slash(fname_path)); continue; } /* Refresh the stat data for the discovered target */ st = result.st; close_and_replace(fd, result.fd); free(resolved_file); resolved_file = path_join(root, result.path); if (!resolved_file) return -ENOMEM; /* fname and fname_path are invalid now because they would need to be set * from result.path by extracting the filename to set * fname = path_join(fname, filename) and then * fname_path = path_join(*s, fname) but since they are unused we don't do it */ fname = NULL; fname_path = mfree(fname_path); } else if (!S_ISDIR(st.st_mode) && !S_ISBLK(st.st_mode)) { log_debug("Ignoring non-directory and non-block device file '%s' without suffix.", fname); continue; } /* Only put resolved paths into the image entry (incl. --root=). * Defending against symlink races is not done * and would be a TODO. */ r = image_make(class, name, fd, resolved_file, &st, ret); if (IN_SET(r, -ENOENT, -EMEDIUMTYPE)) continue; if (r < 0) return r; if (ret) (*ret)->discoverable = true; return 1; } } if (scope == RUNTIME_SCOPE_SYSTEM && class == IMAGE_MACHINE && streq(name, ".host")) { r = image_make(class, ".host", /* fd= */ -EBADF, /* path= */ empty_to_root(root), /* st= */ NULL, ret); if (r < 0) return r; if (ret) (*ret)->discoverable = true; return 1; } return -ENOENT; }; int image_from_path(const char *path, Image **ret) { int r; /* Note that we don't set the 'discoverable' field of the returned object, because we don't check here whether * the image is in the image search path. And if it is we don't know if the path we used is actually not * overridden by another, different image earlier in the search path */ if (path_equal(path, "/")) return image_make( IMAGE_MACHINE, ".host", /* fd= */ -EBADF, /* path= */ "/", /* st= */ NULL, ret); _cleanup_free_ char *absolute = NULL; r = path_make_absolute_cwd(path, &absolute); if (r < 0) return r; return image_make( _IMAGE_CLASS_INVALID, /* pretty= */ NULL, /* fd= */ -EBADF, absolute, /* st= */ NULL, ret); } int image_find_harder( RuntimeScope scope, ImageClass class, const char *name_or_path, const char *root, Image **ret) { if (image_name_is_valid(name_or_path)) return image_find(scope, class, name_or_path, root, ret); return image_from_path(name_or_path, ret); } int image_discover( RuntimeScope scope, ImageClass class, const char *root, Hashmap **images) { int r; assert(scope < _RUNTIME_SCOPE_MAX && scope != RUNTIME_SCOPE_GLOBAL); assert(class >= 0); assert(class < _IMAGE_CLASS_MAX); assert(images); _cleanup_close_ int rfd = XAT_FDROOT; /* We only expect absolute paths */ if (root) { rfd = open(root, O_CLOEXEC|O_DIRECTORY|O_PATH); if (rfd < 0) return log_debug_errno(errno, "Failed to open root directory '%s': %m", root); } _cleanup_strv_free_ char **search = NULL; r = pick_image_search_path(scope, class, root, &search); if (r < 0) return r; STRV_FOREACH(s, search) { _cleanup_closedir_ DIR *d = NULL; _cleanup_free_ char *search_path = NULL; r = chase_and_opendirat(rfd, rfd, *s, /* chase_flags= */ 0, &search_path, &d); if (r == -ENOENT) continue; if (r < 0) return r; FOREACH_DIRENT_ALL(de, d, return -errno) { _cleanup_free_ char *pretty = NULL, *fname_path = NULL, *chased_path = NULL, *resolved_file = NULL; _cleanup_(image_unrefp) Image *image = NULL; const char *fname = de->d_name; _cleanup_close_ int fd = -EBADF; if (dot_or_dot_dot(fname)) continue; /* Ignore sysupdate temporary files */ if (startswith(fname, ".sysupdate.")) continue; fname_path = path_join(search_path, fname); if (!fname_path) return -ENOMEM; /* Follow symlinks only inside given root */ r = chaseat(rfd, rfd, fname_path, /* flags= */ 0, &chased_path, &fd); if (r == -ENOENT) continue; if (r < 0) return r; r = chaseat_prefix_root(chased_path, root, &resolved_file); if (r < 0) return r; struct stat st; if (fstat(fd, &st) < 0) return -errno; if (S_ISREG(st.st_mode)) { r = extract_image_basename( fname, image_class_suffix_to_string(class), STRV_MAKE(".raw"), &pretty, /* ret_suffix= */ NULL); if (r < 0) { log_debug_errno(r, "Skipping directory entry '%s', which doesn't look like an image.", fname); continue; } } else if (S_ISDIR(st.st_mode)) { const char *v; v = endswith(fname, ".v"); if (v) { _cleanup_free_ char *suffix = NULL, *nov = NULL; nov = strndup(fname, v - fname); /* Chop off the .v */ if (!nov) return -ENOMEM; r = extract_image_basename( nov, image_class_suffix_to_string(class), STRV_MAKE(".raw", ".mstack", ""), &pretty, &suffix); if (r < 0) { log_debug_errno(r, "Skipping directory entry '%s', which doesn't look like a versioned image.", fname); continue; } PickFilter filter = { .type_mask = endswith(suffix, ".raw") ? (UINT32_C(1) << DT_REG) | (UINT32_C(1) << DT_BLK) : (UINT32_C(1) << DT_DIR), .basename = pretty, .architecture = _ARCHITECTURE_INVALID, .suffix = suffix, }; _cleanup_(pick_result_done) PickResult result = PICK_RESULT_NULL; r = path_pick(root, rfd, fname_path, /* This has to be the unresolved entry with the .v suffix */ &filter, /* n_filters= */ 1, PICK_ARCHITECTURE|PICK_TRIES|PICK_RESOLVE, &result); if (r < 0) { log_debug_errno(r, "Failed to pick versioned image on '%s%s', skipping: %m", empty_to_root(root), skip_leading_slash(fname_path)); continue; } if (!result.path) { log_debug("Found versioned directory '%s%s', without matching entry, skipping.", empty_to_root(root), skip_leading_slash(fname_path)); continue; } /* Refresh the stat data for the discovered target */ st = result.st; close_and_replace(fd, result.fd); free(resolved_file); resolved_file = path_join(root, result.path); if (!resolved_file) return -ENOMEM; /* fname and fname_path are invalid now because they would need to * be set from result.path by extracting the filename to set * fname = path_join(fname, filename) and then * fname_path = path_join(*s, fname) but since they are unused we * don't do it */ fname = NULL; fname_path = mfree(fname_path); } else { r = extract_image_basename( fname, image_class_suffix_to_string(class), STRV_MAKE(".mstack", ""), &pretty, /* ret_suffix= */ NULL); if (r < 0) { log_debug_errno(r, "Skipping directory entry '%s', which doesn't look like an image.", fname); continue; } } } else if (S_ISBLK(st.st_mode)) { r = extract_image_basename( fname, /* class_suffix= */ NULL, /* format_suffix= */ NULL, &pretty, /* ret_suffix= */ NULL); if (r < 0) { log_debug_errno(r, "Skipping directory entry '%s', which doesn't look like an image.", fname); continue; } } else { log_debug("Skipping directory entry '%s', which is neither regular file, directory nor block device.", fname); continue; } if (hashmap_contains(*images, pretty)) continue; /* Only put resolved paths into the image entry. * Defending against symlink races is not done * and would be a TODO. */ r = image_make(class, pretty, fd, resolved_file, &st, &image); if (IN_SET(r, -ENOENT, -EMEDIUMTYPE)) continue; if (r < 0) return r; image->discoverable = true; r = hashmap_ensure_put(images, &image_hash_ops, image->name, image); if (r < 0) return r; TAKE_PTR(image); } } if (scope == RUNTIME_SCOPE_SYSTEM && class == IMAGE_MACHINE && !hashmap_contains(*images, ".host")) { _cleanup_(image_unrefp) Image *image = NULL; r = image_make(IMAGE_MACHINE, ".host", /* fd= */ -EBADF, /* path= */ empty_to_root(root), /* st= */ NULL, &image); if (r < 0) return r; image->discoverable = true; r = hashmap_ensure_put(images, &image_hash_ops, image->name, image); if (r < 0) return r; image = NULL; } return 0; } static int unpriv_remove_cb( RecurseDirEvent event, const char *path, int dir_fd, int inode_fd, const struct dirent *de, const struct statx *sx, void *userdata) { int r, userns_fd = PTR_TO_FD(userdata); assert(sx); if (event == RECURSE_DIR_ENTER && S_ISDIR(sx->stx_mode) && uid_is_foreign(sx->stx_uid)) { /* This is owned by the foreign UID range, and a dir, let's remove it via mountfsd userns * shenanigans. */ _cleanup_close_ int tree_fd = -EBADF; r = mountfsd_mount_directory_fd( /* vl= */ NULL, inode_fd, userns_fd, DISSECT_IMAGE_FOREIGN_UID, &tree_fd); if (r < 0) return r; /* Fork off child that moves into userns and does the copying */ r = pidref_safe_fork_full( "rm-tree", /* stdio_fds= */ NULL, (int[]) { userns_fd, tree_fd, }, 2, FORK_RESET_SIGNALS|FORK_CLOSE_ALL_FDS|FORK_DEATHSIG_SIGTERM|FORK_WAIT|FORK_REOPEN_LOG, /* ret= */ NULL); if (r < 0) log_debug_errno(r, "Process that was supposed to remove subtree '%s' failed, ignoring: %m", empty_to_root(path)); else if (r == 0) { /* child */ r = namespace_enter( /* pidns_fd= */ -EBADF, /* mntns_fd= */ -EBADF, /* netns_fd= */ -EBADF, userns_fd, /* root_fd= */ -EBADF); if (r < 0) { log_debug_errno(r, "Failed to join user namespace: %m"); _exit(EXIT_FAILURE); } _cleanup_close_ int dfd = fd_reopen(tree_fd, O_DIRECTORY|O_CLOEXEC); if (dfd < 0) { log_error_errno(r, "Failed to reopen tree fd: %m"); _exit(EXIT_FAILURE); } r = rm_rf_children(dfd, REMOVE_PHYSICAL|REMOVE_SUBVOLUME|REMOVE_CHMOD, /* root_dev= */ NULL); if (r < 0) { log_error_errno(r, "Failed to empty '%s' directory in foreign UID mode: %m", empty_to_root(path)); _exit(EXIT_FAILURE); } _exit(EXIT_SUCCESS); } /* Don't descent further into this one, and delete it immediately */ return RECURSE_DIR_UNLINK_GRACEFUL; } /* Everything else try to remove */ if (event == RECURSE_DIR_LEAVE) return RECURSE_DIR_UNLINK_GRACEFUL; return RECURSE_DIR_CONTINUE; } static int unprivileged_remove(Image *i) { int r; assert(i); /* We want this to work in complex .mstack/ hierarchies, where the main directory (and maybe a .v/ * directory below or two) might be owned by the user themselves, but some subdirs might be owned by * the foreign UID range. We deal with this by recursively descending down the tree, and removing * foreign-owned ranges via userns shenanigans, and the rest just like that. */ _cleanup_close_ int userns_fd = nsresource_allocate_userns( /* vl= */ NULL, /* name= */ NULL, /* size= */ NSRESOURCE_UIDS_64K); if (userns_fd < 0) return log_debug_errno(userns_fd, "Failed to allocate transient user namespace: %m"); r = recurse_dir_at( AT_FDCWD, i->path, /* statx_mask= */ STATX_TYPE|STATX_UID, /* n_depth_max= */ UINT_MAX, RECURSE_DIR_IGNORE_DOT|RECURSE_DIR_ENSURE_TYPE|RECURSE_DIR_SAME_MOUNT|RECURSE_DIR_INODE_FD|RECURSE_DIR_TOPLEVEL, unpriv_remove_cb, FD_TO_PTR(userns_fd)); if (r < 0) return r; return 0; } int image_remove(Image *i, RuntimeScope scope) { _cleanup_(release_lock_file) LockFile global_lock = LOCK_FILE_INIT, local_lock = LOCK_FILE_INIT; _cleanup_strv_free_ char **settings = NULL; int r; assert(i); if (image_is_vendor(i) || image_is_host(i)) return -EROFS; settings = image_settings_path(i, scope); if (!settings) return -ENOMEM; /* Make sure we don't interfere with a running nspawn */ r = image_path_lock(scope, i->path, LOCK_EX|LOCK_NB, &global_lock, &local_lock); if (r < 0) return r; switch (i->type) { case IMAGE_SUBVOLUME: /* Let's unlink first, maybe it is a symlink? If that works we are happy. Otherwise, let's get out the * big guns */ if (unlink(i->path) < 0) { r = btrfs_subvol_remove(i->path, BTRFS_REMOVE_RECURSIVE|BTRFS_REMOVE_QUOTA); if (r < 0) return r; } break; case IMAGE_DIRECTORY: /* Allow deletion of read-only directories */ (void) chattr_path(i->path, 0, FS_IMMUTABLE_FL); _fallthrough_; case IMAGE_MSTACK: /* If this is foreign owned, try an unprivileged remove first, but accept if that doesn't work, and do it directly either way, maybe it works */ if (i->foreign_uid_owned) (void) unprivileged_remove(i); r = rm_rf(i->path, REMOVE_ROOT|REMOVE_PHYSICAL|REMOVE_SUBVOLUME); if (r < 0) return r; break; case IMAGE_BLOCK: /* If this is inside of /dev, then it's a real block device, hence let's not touch the device node * itself (but let's remove the stuff stored alongside it). If it's anywhere else, let's try to unlink * the thing (it's most likely a symlink after all). */ if (path_startswith(i->path, "/dev")) break; _fallthrough_; case IMAGE_RAW: if (unlink(i->path) < 0) return -errno; break; default: return -EOPNOTSUPP; } STRV_FOREACH(j, settings) if (unlink(*j) < 0 && errno != ENOENT) log_debug_errno(errno, "Failed to unlink %s, ignoring: %m", *j); NULSTR_FOREACH(suffix, auxiliary_suffixes_nulstr) { _cleanup_free_ char *aux = NULL; r = image_auxiliary_path(i, suffix, &aux); if (r < 0) return r; if (unlink(aux) < 0 && errno != ENOENT) log_debug_errno(errno, "Failed to unlink %s, ignoring: %m", aux); } return 0; } static int rename_auxiliary_file(const char *path, const char *new_name, const char *suffix) { int r; assert(path); assert(new_name); assert(suffix); _cleanup_free_ char *fn = strjoin(new_name, suffix); if (!fn) return -ENOMEM; _cleanup_free_ char *rs = NULL; r = file_in_same_dir(path, fn, &rs); if (r < 0) return r; return rename_noreplace(AT_FDCWD, path, AT_FDCWD, rs); } int image_rename(Image *i, const char *new_name, RuntimeScope scope) { _cleanup_(release_lock_file) LockFile global_lock = LOCK_FILE_INIT, local_lock = LOCK_FILE_INIT, name_lock = LOCK_FILE_INIT; _cleanup_free_ char *new_path = NULL, *nn = NULL; _cleanup_strv_free_ char **settings = NULL; unsigned file_attr = 0; int r; assert(i); if (!image_name_is_valid(new_name)) return -EINVAL; if (image_is_vendor(i) || image_is_host(i)) return -EROFS; settings = image_settings_path(i, scope); if (!settings) return -ENOMEM; /* Make sure we don't interfere with a running nspawn */ r = image_path_lock(scope, i->path, LOCK_EX|LOCK_NB, &global_lock, &local_lock); if (r < 0) return r; /* Make sure nobody takes the new name, between the time we * checked it is currently unused in all search paths, and the * time we take possession of it */ r = image_name_lock(scope, new_name, LOCK_EX|LOCK_NB, &name_lock); if (r < 0) return r; r = image_find(scope, IMAGE_MACHINE, new_name, NULL, NULL); if (r >= 0) return -EEXIST; if (r != -ENOENT) return r; switch (i->type) { case IMAGE_DIRECTORY: /* Turn of the immutable bit while we rename the image, so that we can rename it */ (void) read_attr_at(AT_FDCWD, i->path, &file_attr); if (file_attr & FS_IMMUTABLE_FL) (void) chattr_path(i->path, 0, FS_IMMUTABLE_FL); _fallthrough_; case IMAGE_SUBVOLUME: r = file_in_same_dir(i->path, new_name, &new_path); break; case IMAGE_BLOCK: /* Refuse renaming raw block devices in /dev, the names are picked by udev after all. */ if (path_startswith(i->path, "/dev")) return -EROFS; r = file_in_same_dir(i->path, new_name, &new_path); break; case IMAGE_RAW: { const char *fn; fn = strjoina(new_name, ".raw"); r = file_in_same_dir(i->path, fn, &new_path); break; } default: return -EOPNOTSUPP; } if (r < 0) return r; nn = strdup(new_name); if (!nn) return -ENOMEM; r = rename_noreplace(AT_FDCWD, i->path, AT_FDCWD, new_path); if (r < 0) return r; /* Restore the immutable bit, if it was set before */ if (file_attr & FS_IMMUTABLE_FL) (void) chattr_path(new_path, FS_IMMUTABLE_FL, FS_IMMUTABLE_FL); free_and_replace(i->path, new_path); free_and_replace(i->name, nn); STRV_FOREACH(j, settings) { r = rename_auxiliary_file(*j, new_name, ".nspawn"); if (r < 0 && r != -ENOENT) log_debug_errno(r, "Failed to rename settings file %s, ignoring: %m", *j); } NULSTR_FOREACH(suffix, auxiliary_suffixes_nulstr) { _cleanup_free_ char *aux = NULL; r = image_auxiliary_path(i, suffix, &aux); if (r < 0) return r; r = rename_auxiliary_file(aux, new_name, suffix); if (r < 0 && r != -ENOENT) log_debug_errno(r, "Failed to rename roothash file %s, ignoring: %m", aux); } return 0; } static int clone_auxiliary_file(const char *path, const char *new_name, const char *suffix) { int r; assert(path); assert(new_name); assert(suffix); _cleanup_free_ char *fn = strjoin(new_name, suffix); if (!fn) return -ENOMEM; _cleanup_free_ char *rs = NULL; r = file_in_same_dir(path, fn, &rs); if (r < 0) return r; return copy_file_atomic(path, rs, 0664, COPY_REFLINK); } static int get_pool_directory( RuntimeScope scope, ImageClass class, const char *fname, const char *suffix, char **ret) { int r; assert(scope >= 0); assert(scope < _RUNTIME_SCOPE_MAX); assert(class >= 0); assert(class < _IMAGE_CLASS_MAX); assert(ret); _cleanup_free_ char *root = NULL; switch (scope) { case RUNTIME_SCOPE_SYSTEM: r = sd_path_lookup(SD_PATH_SYSTEM_STATE_PRIVATE, /* suffix= */ NULL, &root); break; case RUNTIME_SCOPE_USER: r = sd_path_lookup(SD_PATH_USER_STATE_PRIVATE, /* suffix= */ NULL, &root); break; default: return -EOPNOTSUPP; } if (r < 0) return r; const char *n = image_dirname_to_string(class); if (!n) return -EOPNOTSUPP; _cleanup_free_ char *j = NULL; const char *fn; if (fname && suffix) { j = strjoin(fname, suffix); if (!j) return -ENOMEM; fn = j; } else fn = fname ?: suffix; _cleanup_free_ char *p = path_join(root, n, fn); if (!p) return -ENOMEM; *ret = TAKE_PTR(p); return 0; } static int unprivileged_clone(Image *i, const char *new_path) { int r; assert(i); assert(new_path); _cleanup_close_ int userns_fd = nsresource_allocate_userns( /* vl= */ NULL, /* name= */ NULL, /* size= */ NSRESOURCE_UIDS_64K); if (userns_fd < 0) return log_debug_errno(userns_fd, "Failed to allocate transient user namespace: %m"); _cleanup_(sd_varlink_unrefp) sd_varlink *link = NULL; r = mountfsd_connect(&link); if (r < 0) return r; /* Map original image */ _cleanup_close_ int tree_fd = -EBADF; r = mountfsd_mount_directory( link, i->path, userns_fd, DISSECT_IMAGE_FOREIGN_UID, &tree_fd); if (r < 0) return r; /* Make new image */ _cleanup_close_ int new_fd = -EBADF; r = mountfsd_make_directory( link, new_path, MODE_INVALID, /* flags= */ 0, &new_fd); if (r < 0) return 0; /* Mount new image */ _cleanup_close_ int target_fd = -EBADF; r = mountfsd_mount_directory_fd( link, new_fd, userns_fd, DISSECT_IMAGE_FOREIGN_UID, &target_fd); if (r < 0) return r; link = sd_varlink_unref(link); /* Fork off child that moves into userns and does the copying */ return copy_tree_at_foreign(tree_fd, target_fd, userns_fd); } int image_clone(Image *i, const char *new_name, bool read_only, RuntimeScope scope) { _cleanup_(release_lock_file) LockFile name_lock = LOCK_FILE_INIT; _cleanup_strv_free_ char **settings = NULL; int r; assert(i); if (!image_name_is_valid(new_name)) return -EINVAL; settings = image_settings_path(i, scope); if (!settings) return -ENOMEM; /* Make sure nobody takes the new name, between the time we * checked it is currently unused in all search paths, and the * time we take possession of it */ r = image_name_lock(scope, new_name, LOCK_EX|LOCK_NB, &name_lock); if (r < 0) return r; r = image_find(scope, i->class, new_name, NULL, NULL); if (r >= 0) return -EEXIST; if (r != -ENOENT) return r; switch (i->type) { case IMAGE_SUBVOLUME: case IMAGE_DIRECTORY: { /* If we can we'll always try to create a new btrfs subvolume here, even if the source is a plain * directory. */ _cleanup_free_ char *new_path = NULL; r = get_pool_directory(scope, i->class, new_name, /* suffix= */ NULL, &new_path); if (r < 0) return r; if (i->foreign_uid_owned) r = unprivileged_clone(i, new_path); else { r = btrfs_subvol_snapshot_at( AT_FDCWD, i->path, AT_FDCWD, new_path, (read_only ? BTRFS_SNAPSHOT_READ_ONLY : 0) | BTRFS_SNAPSHOT_FALLBACK_COPY | BTRFS_SNAPSHOT_FALLBACK_DIRECTORY | BTRFS_SNAPSHOT_FALLBACK_IMMUTABLE | BTRFS_SNAPSHOT_RECURSIVE | BTRFS_SNAPSHOT_QUOTA); if (r >= 0) /* Enable "subtree" quotas for the copy, if we didn't copy any quota from the source. */ (void) btrfs_subvol_auto_qgroup(new_path, /* subvol_id= */ 0, /* create_intermediary_qgroup= */ true); } break; } case IMAGE_RAW: { _cleanup_free_ char *new_path = NULL; r = get_pool_directory(scope, i->class, new_name, ".raw", &new_path); if (r < 0) return r; r = copy_file_atomic(i->path, new_path, read_only ? 0444 : 0644, COPY_REFLINK|COPY_CRTIME|COPY_NOCOW_AFTER); break; } case IMAGE_BLOCK: default: return -EOPNOTSUPP; } if (r < 0) return r; STRV_FOREACH(j, settings) { r = clone_auxiliary_file(*j, new_name, ".nspawn"); if (r < 0 && r != -ENOENT) log_debug_errno(r, "Failed to clone settings %s, ignoring: %m", *j); } NULSTR_FOREACH(suffix, auxiliary_suffixes_nulstr) { _cleanup_free_ char *aux = NULL; r = image_auxiliary_path(i, suffix, &aux); if (r < 0) return r; r = clone_auxiliary_file(aux, new_name, suffix); if (r < 0 && r != -ENOENT) log_debug_errno(r, "Failed to clone root hash file %s, ignoring: %m", aux); } return 0; } int image_read_only(Image *i, bool b, RuntimeScope scope) { _cleanup_(release_lock_file) LockFile global_lock = LOCK_FILE_INIT, local_lock = LOCK_FILE_INIT; int r; assert(i); if (image_is_vendor(i) || image_is_host(i) || image_is_hidden(i)) return -EROFS; /* Make sure we don't interfere with a running nspawn */ r = image_path_lock(scope, i->path, LOCK_EX|LOCK_NB, &global_lock, &local_lock); if (r < 0) return r; switch (i->type) { case IMAGE_SUBVOLUME: /* Note that we set the flag only on the top-level * subvolume of the image. */ r = btrfs_subvol_set_read_only(i->path, b); if (r < 0) return r; break; case IMAGE_DIRECTORY: /* For simple directory trees we cannot use the access mode of the top-level directory, since it has an effect on the container itself. However, we can use the "immutable" flag, to at least make the top-level directory read-only. It's not as good as a read-only subvolume, but at least something, and we can read the value back. */ r = chattr_path(i->path, b ? FS_IMMUTABLE_FL : 0, FS_IMMUTABLE_FL); if (r < 0) return r; break; case IMAGE_RAW: { struct stat st; if (stat(i->path, &st) < 0) return -errno; if (chmod(i->path, (st.st_mode & 0444) | (b ? 0000 : 0200)) < 0) return -errno; /* If the images is now read-only, it's a good time to * defrag it, given that no write patterns will * fragment it again. */ if (b) (void) btrfs_defrag(i->path); break; } case IMAGE_BLOCK: { _cleanup_close_ int fd = -EBADF; int state = b; fd = open(i->path, O_CLOEXEC|O_RDONLY|O_NONBLOCK|O_NOCTTY); if (fd < 0) return -errno; r = fd_verify_block(fd); if (r < 0) return r; if (ioctl(fd, BLKROSET, &state) < 0) return -errno; break; } default: return -EOPNOTSUPP; } i->read_only = b; return 0; } static int make_lock_dir(RuntimeScope scope) { int r; _cleanup_free_ char *p = NULL; r = runtime_directory_generic(scope, "systemd", &p); if (r < 0) return r; _cleanup_close_ int pfd = open_mkdir_at(AT_FDCWD, p, O_CLOEXEC, 0755); if (pfd < 0) return pfd; _cleanup_close_ int nfd = open_mkdir_at(pfd, "nspawn", O_CLOEXEC, 0755); if (nfd < 0) return nfd; r = RET_NERRNO(mkdirat(nfd, "locks", 0700)); if (r == -EEXIST) return 0; if (r < 0) return r; return 1; } int image_path_lock( RuntimeScope scope, const char *path, int operation, LockFile *ret_global, LockFile *ret_local) { _cleanup_free_ char *p = NULL; LockFile t = LOCK_FILE_INIT; struct stat st; bool exclusive; int r; assert(path); assert(ret_local); /* Locks an image path. This actually creates two locks: one "local" one, next to the image path * itself, which might be shared via NFS. And another "global" one, in /run/, that uses the * device/inode number. This has the benefit that we can even lock a tree that is a mount point, * correctly. */ if (!path_is_absolute(path)) return -EINVAL; switch (operation & (LOCK_SH|LOCK_EX)) { case LOCK_SH: exclusive = false; break; case LOCK_EX: exclusive = true; break; default: return -EINVAL; } if (getenv_bool("SYSTEMD_NSPAWN_LOCK") == 0) { *ret_local = LOCK_FILE_INIT; if (ret_global) *ret_global = LOCK_FILE_INIT; return 0; } /* Prohibit taking exclusive locks on the host image. We can't allow this, since we ourselves are * running off it after all, and we don't want any images to manipulate the host image. We make an * exception for shared locks however: we allow those (and make them NOPs since there's no point in * taking them if there can't be exclusive locks). Strictly speaking these are questionable as well, * since it means changes made to the host might propagate to the container as they happen (and a * shared lock kinda suggests that no changes happen at all while it is in place), but it's too * useful not to allow read-only containers off the host root, hence let's support this, and trust * the user to do the right thing with this. */ if (path_equal(path, "/")) { if (exclusive) return -EBUSY; *ret_local = LOCK_FILE_INIT; if (ret_global) *ret_global = LOCK_FILE_INIT; return 0; } if (ret_global) { if (stat(path, &st) < 0) log_debug_errno(errno, "Failed to stat() image '%s', not locking image: %m", path); else { r = runtime_directory_generic(scope, "systemd/nspawn/locks/", &p); if (r < 0) return r; if (S_ISBLK(st.st_mode)) r = strextendf(&p, "block-" DEVNUM_FORMAT_STR, DEVNUM_FORMAT_VAL(st.st_rdev)); else if (S_ISDIR(st.st_mode) || S_ISREG(st.st_mode)) r = strextendf(&p, "inode-%" PRIu64 ":%" PRIu64, (uint64_t) st.st_dev, (uint64_t) st.st_ino); else return -ENOTTY; if (r < 0) return r; } } /* For block devices we don't need the "local" lock, as the major/minor lock above should be * sufficient, since block devices are host local anyway. */ if (!path_startswith(path, "/dev/")) { r = make_lock_file_for(path, operation, &t); if (r < 0) { if (exclusive || r != -EROFS) return r; log_debug_errno(r, "Failed to create shared lock for '%s', ignoring: %m", path); } } if (p) { (void) make_lock_dir(scope); r = make_lock_file(p, operation, ret_global); if (r < 0) { release_lock_file(&t); return r; } } else if (ret_global) *ret_global = LOCK_FILE_INIT; *ret_local = t; return 0; } int image_set_limit(Image *i, uint64_t referenced_max) { int r; assert(i); if (image_is_vendor(i) || image_is_host(i)) return -EROFS; if (i->type != IMAGE_SUBVOLUME) return -EOPNOTSUPP; /* We set the quota both for the subvolume as well as for the * subtree. The latter is mostly for historical reasons, since * we didn't use to have a concept of subtree quota, and hence * only modified the subvolume quota. */ (void) btrfs_qgroup_set_limit(i->path, 0, referenced_max); (void) btrfs_subvol_auto_qgroup(i->path, 0, true); r = btrfs_subvol_set_subtree_quota_limit(i->path, 0, referenced_max); if (r < 0) return r; (void) image_update_quota(i, -EBADF); return 0; } int image_set_pool_limit(RuntimeScope scope, ImageClass class, uint64_t referenced_max) { int r; assert(scope >= 0 && scope < _RUNTIME_SCOPE_MAX); assert(class >= 0 && class < _IMAGE_CLASS_MAX); _cleanup_free_ char *pool = NULL; r = get_pool_directory(scope, class, /* fname= */ NULL, /* suffix= */ NULL, &pool); if (r < 0) return r; r = btrfs_qgroup_set_limit(pool, /* qgroupid= */ 0, referenced_max); if (ERRNO_IS_NEG_NOT_SUPPORTED(r)) return r; if (r < 0) log_debug_errno(r, "Failed to set limit on btrfs quota group for '%s', ignoring: %m", pool); r = btrfs_subvol_set_subtree_quota_limit(pool, /* subvol_id= */ 0, referenced_max); if (ERRNO_IS_NEG_NOT_SUPPORTED(r)) return r; if (r < 0) return log_debug_errno(r, "Failed to set subtree quota limit for '%s': %m", pool); return 0; } int image_get_pool_path(RuntimeScope scope, ImageClass class, char **ret) { assert(scope >= 0 && scope < _RUNTIME_SCOPE_MAX); assert(class >= 0 && class < _IMAGE_CLASS_MAX); assert(ret); return get_pool_directory(scope, class, /* fname= */ NULL, /* suffix= */ NULL, ret); } int image_get_pool_usage(RuntimeScope scope, ImageClass class, uint64_t *ret) { int r; assert(scope >= 0 && scope < _RUNTIME_SCOPE_MAX); assert(class >= 0 && class < _IMAGE_CLASS_MAX); assert(ret); _cleanup_free_ char *pool = NULL; r = get_pool_directory(scope, class, /* fname= */ NULL, /* suffix= */ NULL, &pool); if (r < 0) return r; _cleanup_close_ int fd = open(pool, O_RDONLY|O_CLOEXEC|O_DIRECTORY); if (fd < 0) return -errno; BtrfsQuotaInfo q; r = btrfs_subvol_get_subtree_quota_fd(fd, /* subvol_id= */ 0, &q); if (r < 0) return r; *ret = q.referenced; return 0; } int image_get_pool_limit(RuntimeScope scope, ImageClass class, uint64_t *ret) { int r; assert(scope >= 0 && scope < _RUNTIME_SCOPE_MAX); assert(class >= 0 && class < _IMAGE_CLASS_MAX); assert(ret); _cleanup_free_ char *pool = NULL; r = get_pool_directory(scope, class, /* fname= */ NULL, /* suffix= */ NULL, &pool); if (r < 0) return r; _cleanup_close_ int fd = open(pool, O_RDONLY|O_CLOEXEC|O_DIRECTORY); if (fd < 0) return -errno; BtrfsQuotaInfo q; r = btrfs_subvol_get_subtree_quota_fd(fd, /* subvol_id= */ 0, &q); if (r < 0) return r; *ret = q.referenced_max; return 0; } static int check_btrfs(const char *path) { struct statfs sfs; int r; if (statfs(path, &sfs) < 0) { if (errno != ENOENT) return -errno; _cleanup_free_ char *parent = NULL; r = path_extract_directory(path, &parent); if (r < 0) return r; if (statfs(parent, &sfs) < 0) return -errno; } return F_TYPE_EQUAL(sfs.f_type, BTRFS_SUPER_MAGIC); } int image_setup_pool(RuntimeScope scope, ImageClass class, bool use_btrfs_subvol, bool use_btrfs_quota) { int r; assert(class >= 0 && class < _IMAGE_CLASS_MAX); _cleanup_free_ char *pool = NULL; r = image_get_pool_path(scope, class, &pool); if (r < 0) return r; r = check_btrfs(pool); if (r <= 0) return r; if (!use_btrfs_subvol) return 0; (void) btrfs_subvol_make_label(pool); if (!use_btrfs_quota) return 0; r = btrfs_quota_enable(pool, /* b= */ true); if (r < 0) log_warning_errno(r, "Failed to enable quota for %s, ignoring: %m", pool); r = btrfs_subvol_auto_qgroup(pool, /* subvol_id= */ 0, /* create_intermediary_qgroup= */ true); if (r < 0) log_warning_errno(r, "Failed to set up default quota hierarchy for %s, ignoring: %m", pool); return 0; } int image_read_metadata(Image *i, const char *root, const ImagePolicy *image_policy, RuntimeScope scope) { _cleanup_(release_lock_file) LockFile global_lock = LOCK_FILE_INIT, local_lock = LOCK_FILE_INIT; int r; assert(i); r = image_path_lock(scope, i->path, LOCK_SH|LOCK_NB, &global_lock, &local_lock); if (r < 0) return r; switch (i->type) { case IMAGE_SUBVOLUME: case IMAGE_DIRECTORY: { _cleanup_strv_free_ char **machine_info = NULL, **os_release = NULL, **sysext_release = NULL, **confext_release = NULL; _cleanup_free_ char *hostname = NULL, *path = NULL; sd_id128_t machine_id = SD_ID128_NULL; if (i->class == IMAGE_SYSEXT) { r = extension_has_forbidden_content(i->path); if (r < 0) return r; if (r > 0) return log_debug_errno(SYNTHETIC_ERRNO(ENOMEDIUM), "Conflicting content found in image %s, refusing.", i->name); } r = chase("/etc/hostname", i->path, CHASE_PREFIX_ROOT|CHASE_TRAIL_SLASH, &path, NULL); if (r < 0 && r != -ENOENT) log_debug_errno(r, "Failed to chase /etc/hostname in image %s: %m", i->name); else if (r >= 0) { r = read_etc_hostname(path, /* substitute_wildcards= */ false, &hostname); if (r < 0) log_debug_errno(r, "Failed to read /etc/hostname of image %s: %m", i->name); } path = mfree(path); r = id128_get_machine(i->path, &machine_id); if (r < 0) log_debug_errno(r, "Failed to read machine ID in image %s, ignoring: %m", i->name); r = chase("/etc/machine-info", i->path, CHASE_PREFIX_ROOT|CHASE_TRAIL_SLASH, &path, NULL); if (r < 0 && r != -ENOENT) log_debug_errno(r, "Failed to chase /etc/machine-info in image %s: %m", i->name); else if (r >= 0) { r = load_env_file_pairs(NULL, path, &machine_info); if (r < 0) log_debug_errno(r, "Failed to parse machine-info data of %s: %m", i->name); } r = load_os_release_pairs(i->path, &os_release); if (r < 0) log_debug_errno(r, "Failed to read os-release in image, ignoring: %m"); r = load_extension_release_pairs(i->path, IMAGE_SYSEXT, i->name, /* relax_extension_release_check= */ false, &sysext_release); if (r < 0) log_debug_errno(r, "Failed to read sysext-release in image, ignoring: %m"); r = load_extension_release_pairs(i->path, IMAGE_CONFEXT, i->name, /* relax_extension_release_check= */ false, &confext_release); if (r < 0) log_debug_errno(r, "Failed to read confext-release in image, ignoring: %m"); free_and_replace(i->hostname, hostname); i->machine_id = machine_id; strv_free_and_replace(i->machine_info, machine_info); strv_free_and_replace(i->os_release, os_release); strv_free_and_replace(i->sysext_release, sysext_release); strv_free_and_replace(i->confext_release, confext_release); break; } case IMAGE_RAW: case IMAGE_BLOCK: { _cleanup_(verity_settings_done) VeritySettings verity = VERITY_SETTINGS_DEFAULT; _cleanup_(loop_device_unrefp) LoopDevice *d = NULL; _cleanup_(dissected_image_unrefp) DissectedImage *m = NULL; DissectImageFlags flags = DISSECT_IMAGE_GENERIC_ROOT | DISSECT_IMAGE_REQUIRE_ROOT | DISSECT_IMAGE_RELAX_VAR_CHECK | DISSECT_IMAGE_READ_ONLY | DISSECT_IMAGE_USR_NO_ROOT | DISSECT_IMAGE_ADD_PARTITION_DEVICES | DISSECT_IMAGE_PIN_PARTITION_DEVICES | DISSECT_IMAGE_VALIDATE_OS | DISSECT_IMAGE_VALIDATE_OS_EXT | DISSECT_IMAGE_ALLOW_USERSPACE_VERITY; r = loop_device_make_by_path( i->path, O_RDONLY, /* sector_size= */ UINT32_MAX, LO_FLAGS_PARTSCAN, LOCK_SH, &d); if (r < 0) return log_debug_errno(r, "Failed to create loopback device of '%s': %m", i->path); r = dissect_loop_device( d, &verity, /* mount_options= */ NULL, image_policy, /* image_filter= */ NULL, flags, &m); if (r < 0) return log_debug_errno(r, "Failed to dissect image '%s': %m", i->path); r = dissected_image_load_verity_sig_partition( m, d->fd, &verity); if (r < 0) return log_debug_errno(r, "Failed to load Verity signature partition of '%s': %m", i->path); r = dissected_image_guess_verity_roothash( m, &verity); if (r < 0) return log_debug_errno(r, "Failed to guess Verity root hash of '%s': %m", i->path); r = dissected_image_decrypt( m, root, /* passphrase= */ NULL, &verity, image_policy, flags); if (r < 0) return log_debug_errno(r, "Failed to decrypt image '%s': %m", i->path); /* Do not use the image name derived from the backing file of the loop device */ r = free_and_strdup(&m->image_name, i->name); if (r < 0) return r; r = dissected_image_acquire_metadata( m, /* userns_fd= */ -EBADF, flags); if (r < 0) return log_debug_errno(r, "Failed to acquire metadata from image '%s': %m", i->path); free_and_replace(i->hostname, m->hostname); i->machine_id = m->machine_id; strv_free_and_replace(i->machine_info, m->machine_info); strv_free_and_replace(i->os_release, m->os_release); strv_free_and_replace(i->sysext_release, m->sysext_release); strv_free_and_replace(i->confext_release, m->confext_release); break; } default: return -EOPNOTSUPP; } i->metadata_valid = true; return 0; } int image_name_lock( RuntimeScope scope, const char *name, int operation, LockFile *ret) { int r; assert(name); assert(ret); /* Locks an image name, regardless of the precise path used. */ if (streq(name, ".host")) return -EBUSY; if (!image_name_is_valid(name)) return -EINVAL; if (getenv_bool("SYSTEMD_NSPAWN_LOCK") == 0) { *ret = (LockFile) LOCK_FILE_INIT; return 0; } (void) make_lock_dir(scope); _cleanup_free_ char *p = NULL; r = runtime_directory_generic(scope, "systemd/nspawn/locks/name-", &p); if (r < 0) return r; if (!strextend(&p, name)) return -ENOMEM; return make_lock_file(p, operation, ret); } bool image_in_search_path( RuntimeScope scope, ImageClass class, const char *root, const char *image) { int r; assert(scope < _RUNTIME_SCOPE_MAX && scope != RUNTIME_SCOPE_GLOBAL); assert(class >= 0); assert(class < _IMAGE_CLASS_MAX); assert(image); _cleanup_strv_free_ char **search = NULL; r = pick_image_search_path(scope, class, root, &search); if (r < 0) return r; STRV_FOREACH(path, search) { const char *p, *q; size_t k; if (!empty_or_root(root)) { q = path_startswith(*path, root); if (!q) continue; } else q = *path; p = path_startswith(q, *path); if (!p) continue; /* Make sure there's a filename following */ k = strcspn(p, "/"); if (k == 0) continue; p += k; /* Accept trailing slashes */ if (p[strspn(p, "/")] == 0) return true; } return false; } bool image_is_vendor(const struct Image *i) { assert(i); return i->path && path_startswith(i->path, "/usr"); } bool image_is_host(const struct Image *i) { assert(i); if (i->name && streq(i->name, ".host")) return true; if (i->path && path_equal(i->path, "/")) return true; return false; } int image_to_json(const struct Image *img, sd_json_variant **ret) { assert(img); return sd_json_buildo( ret, SD_JSON_BUILD_PAIR_STRING("Type", image_type_to_string(img->type)), SD_JSON_BUILD_PAIR_STRING("Class", image_class_to_string(img->class)), SD_JSON_BUILD_PAIR_STRING("Name", img->name), SD_JSON_BUILD_PAIR_CONDITION(!!img->path, "Path", SD_JSON_BUILD_STRING(img->path)), SD_JSON_BUILD_PAIR_BOOLEAN("ReadOnly", image_is_read_only(img)), SD_JSON_BUILD_PAIR_CONDITION(img->crtime != 0, "CreationTimestamp", SD_JSON_BUILD_UNSIGNED(img->crtime)), SD_JSON_BUILD_PAIR_CONDITION(img->mtime != 0, "ModificationTimestamp", SD_JSON_BUILD_UNSIGNED(img->mtime)), SD_JSON_BUILD_PAIR_CONDITION(img->usage != UINT64_MAX, "Usage", SD_JSON_BUILD_UNSIGNED(img->usage)), SD_JSON_BUILD_PAIR_CONDITION(img->usage_exclusive != UINT64_MAX, "UsageExclusive", SD_JSON_BUILD_UNSIGNED(img->usage_exclusive)), SD_JSON_BUILD_PAIR_CONDITION(img->limit != UINT64_MAX, "Limit", SD_JSON_BUILD_UNSIGNED(img->limit)), SD_JSON_BUILD_PAIR_CONDITION(img->limit_exclusive != UINT64_MAX, "LimitExclusive", SD_JSON_BUILD_UNSIGNED(img->limit_exclusive))); } static const char* const image_type_table[_IMAGE_TYPE_MAX] = { [IMAGE_DIRECTORY] = "directory", [IMAGE_SUBVOLUME] = "subvolume", [IMAGE_RAW] = "raw", [IMAGE_BLOCK] = "block", [IMAGE_MSTACK] = "mstack", }; DEFINE_STRING_TABLE_LOOKUP(image_type, ImageType); int image_root_pick( RuntimeScope scope, ImageClass c, bool runtime, char **ret) { int r; assert(scope >= 0); assert(scope < _RUNTIME_SCOPE_MAX); assert(c >= 0); assert(c < _IMAGE_CLASS_MAX); assert(ret); /* Picks the primary target directory for downloads, depending on invocation contexts */ _cleanup_free_ char *s = NULL; switch (scope) { case RUNTIME_SCOPE_SYSTEM: { s = strdup(runtime ? image_root_runtime_to_string(c) : image_root_to_string(c)); if (!s) return -ENOMEM; break; } case RUNTIME_SCOPE_USER: r = sd_path_lookup(runtime ? SD_PATH_USER_RUNTIME : SD_PATH_USER_STATE_PRIVATE, "machines", &s); if (r < 0) return r; break; default: return -EOPNOTSUPP; } *ret = TAKE_PTR(s); return 0; } BUS_DEFINE_PROPERTY_GET(bus_property_get_image_is_read_only, "b", Image, (int) image_is_read_only);