int mount_cgroup_controllers(char ***join_controllers) {
_cleanup_set_free_free_ Set *controllers = NULL;
int r;
if (!cg_is_legacy_wanted())
return 0;
/* Mount all available cgroup controllers that are built into the kernel. */
controllers = set_new(&string_hash_ops);
if (!controllers)
return log_oom();
r = cg_kernel_controllers(controllers);
if (r < 0)
return log_error_errno(r, "Failed to enumerate cgroup controllers: %m");
for (;;) {
_cleanup_free_ char *options = NULL, *controller = NULL, *where = NULL;
MountPoint p = {
.what = "cgroup",
.type = "cgroup",
.flags = MS_NOSUID|MS_NOEXEC|MS_NODEV,
.mode = MNT_IN_CONTAINER,
};
char ***k = NULL;
controller = set_steal_first(controllers);
if (!controller)
break;
if (join_controllers)
for (k = join_controllers; *k; k++)
if (strv_find(*k, controller))
break;
if (k && *k) {
char **i, **j;
for (i = *k, j = *k; *i; i++) {
if (!streq(*i, controller)) {
_cleanup_free_ char *t;
t = set_remove(controllers, *i);
if (!t) {
free(*i);
continue;
}
}
*(j++) = *i;
}
*j = NULL;
options = strv_join(*k, ",");
if (!options)
return log_oom();
} else {
options = controller;
controller = NULL;
}
where = strappend("/sys/fs/cgroup/", options);
if (!where)
return log_oom();
p.where = where;
p.options = options;
r = mount_one(&p, true);
if (r < 0)
return r;
if (r > 0 && k && *k) {
char **i;
for (i = *k; *i; i++) {
_cleanup_free_ char *t = NULL;
t = strappend("/sys/fs/cgroup/", *i);
if (!t)
return log_oom();
r = symlink(options, t);
if (r < 0 && errno != EEXIST)
return log_error_errno(errno, "Failed to create symlink %s: %m", t);
#ifdef SMACK_RUN_LABEL
r = mac_smack_copy(t, options);
if (r < 0 && r != -EOPNOTSUPP)
return log_error_errno(r, "Failed to copy smack label from %s to %s: %m", options, t);
#endif
}
}
}
/* Now that we mounted everything, let's make the tmpfs the
* cgroup file systems are mounted into read-only. */
(void) mount("tmpfs", "/sys/fs/cgroup", "tmpfs", MS_REMOUNT|MS_NOSUID|MS_NOEXEC|MS_NODEV|MS_STRICTATIME|MS_RDONLY, "mode=755");
return 0;
}
#if defined(HAVE_SELINUX) || defined(HAVE_SMACK)
static int nftw_cb(
const char *fpath,
const struct stat *sb,
int tflag,
struct FTW *ftwbuf) {
/* No need to label /dev twice in a row... */
if (_unlikely_(ftwbuf->level == 0))
return FTW_CONTINUE;
label_fix(fpath, false, false);
/* /run/initramfs is static data and big, no need to
* dynamically relabel its contents at boot... */
if (_unlikely_(ftwbuf->level == 1 &&
tflag == FTW_D &&
streq(fpath, "/run/initramfs")))
return FTW_SKIP_SUBTREE;
return FTW_CONTINUE;
};
#endif
int mount_setup(bool loaded_policy) {
unsigned i;
int r = 0;
for (i = 0; i < ELEMENTSOF(mount_table); i ++) {
int j;
j = mount_one(mount_table + i, loaded_policy);
if (j != 0 && r >= 0)
r = j;
}
if (r < 0)
return r;
#if defined(HAVE_SELINUX) || defined(HAVE_SMACK)
/* Nodes in devtmpfs and /run need to be manually updated for
* the appropriate labels, after mounting. The other virtual
* API file systems like /sys and /proc do not need that, they
* use the same label for all their files. */
if (loaded_policy) {
usec_t before_relabel, after_relabel;
char timespan[FORMAT_TIMESPAN_MAX];
before_relabel = now(CLOCK_MONOTONIC);
nftw("/dev", nftw_cb, 64, FTW_MOUNT|FTW_PHYS|FTW_ACTIONRETVAL);
nftw("/run", nftw_cb, 64, FTW_MOUNT|FTW_PHYS|FTW_ACTIONRETVAL);
after_relabel = now(CLOCK_MONOTONIC);
log_info("Relabelled /dev and /run in %s.",
format_timespan(timespan, sizeof(timespan), after_relabel - before_relabel, 0));
}
#endif
/* Create a few default symlinks, which are normally created
* by udevd, but some scripts might need them before we start
* udevd. */
dev_setup(NULL, UID_INVALID, GID_INVALID);
/* Mark the root directory as shared in regards to mount
* propagation. The kernel defaults to "private", but we think
* it makes more sense to have a default of "shared" so that
* nspawn and the container tools work out of the box. If
* specific setups need other settings they can reset the
* propagation mode to private if needed. */
if (detect_container() <= 0)
if (mount(NULL, "/", NULL, MS_REC|MS_SHARED, NULL) < 0)
log_warning_errno(errno, "Failed to set up the root directory for shared mount propagation: %m");
/* Create a few directories we always want around, Note that
* sd_booted() checks for /run/systemd/system, so this mkdir
* really needs to stay for good, otherwise software that
* copied sd-daemon.c into their sources will misdetect
* systemd. */
mkdir_label("/run/systemd", 0755);
mkdir_label("/run/systemd/system", 0755);
mkdir_label("/run/systemd/inaccessible", 0000);
return 0;
}
/* Mount legacy cgroup hierarchy when cgroup namespaces are unsupported. */
static int mount_legacy_cgns_unsupported(
const char *dest,
CGroupUnified unified_requested,
bool userns,
uid_t uid_shift,
uid_t uid_range,
const char *selinux_apifs_context) {
_cleanup_set_free_free_ Set *controllers = NULL;
const char *cgroup_root;
int r;
cgroup_root = prefix_roota(dest, "/sys/fs/cgroup");
(void) mkdir_p(cgroup_root, 0755);
/* Mount a tmpfs to /sys/fs/cgroup if it's not mounted there yet. */
r = path_is_mount_point(cgroup_root, dest, AT_SYMLINK_FOLLOW);
if (r < 0)
return log_error_errno(r, "Failed to determine if /sys/fs/cgroup is already mounted: %m");
if (r == 0) {
_cleanup_free_ char *options = NULL;
r = tmpfs_patch_options("mode=755", uid_shift == 0 ? UID_INVALID : uid_shift, selinux_apifs_context, &options);
if (r < 0)
return log_oom();
r = mount_verbose(LOG_ERR, "tmpfs", cgroup_root, "tmpfs",
MS_NOSUID|MS_NOEXEC|MS_NODEV|MS_STRICTATIME, options);
if (r < 0)
return r;
}
r = cg_all_unified();
if (r < 0)
return r;
if (r > 0)
goto skip_controllers;
r = cg_kernel_controllers(&controllers);
if (r < 0)
return log_error_errno(r, "Failed to determine cgroup controllers: %m");
for (;;) {
_cleanup_free_ char *controller = NULL, *origin = NULL, *combined = NULL;
controller = set_steal_first(controllers);
if (!controller)
break;
origin = prefix_root("/sys/fs/cgroup/", controller);
if (!origin)
return log_oom();
r = readlink_malloc(origin, &combined);
if (r == -EINVAL) {
/* Not a symbolic link, but directly a single cgroup hierarchy */
r = mount_legacy_cgroup_hierarchy(dest, controller, controller, true);
if (r < 0)
return r;
} else if (r < 0)
return log_error_errno(r, "Failed to read link %s: %m", origin);
else {
_cleanup_free_ char *target = NULL;
target = prefix_root(dest, origin);
if (!target)
return log_oom();
/* A symbolic link, a combination of controllers in one hierarchy */
if (!filename_is_valid(combined)) {
log_warning("Ignoring invalid combined hierarchy %s.", combined);
continue;
}
r = mount_legacy_cgroup_hierarchy(dest, combined, combined, true);
if (r < 0)
return r;
r = symlink_idempotent(combined, target, false);
if (r == -EINVAL)
return log_error_errno(r, "Invalid existing symlink for combined hierarchy: %m");
if (r < 0)
return log_error_errno(r, "Failed to create symlink for combined hierarchy: %m");
}
}
skip_controllers:
if (unified_requested >= CGROUP_UNIFIED_SYSTEMD) {
r = mount_legacy_cgroup_hierarchy(dest, SYSTEMD_CGROUP_CONTROLLER_HYBRID, "unified", false);
if (r < 0)
return r;
}
r = mount_legacy_cgroup_hierarchy(dest, SYSTEMD_CGROUP_CONTROLLER_LEGACY, "systemd", false);
if (r < 0)
return r;
return mount_verbose(LOG_ERR, NULL, cgroup_root, NULL,
MS_REMOUNT|MS_NOSUID|MS_NOEXEC|MS_NODEV|MS_STRICTATIME|MS_RDONLY, "mode=755");
}
