450442cf93
This adds "systemd-mount" which is for transient mount and automount units what "systemd-run" is for transient service, scope and timer units. The tool allows establishing mounts and automounts during runtime. It is very similar to the usual /bin/mount commands, but can pull in additional dependenices on access (for example, it pulls in fsck automatically), an take benefit of the automount logic. This tool is particularly useful for mount removable file systems (such as USB sticks), as the automount logic (together with automatic unmount-on-idle), as well as automatic fsck on first access ensure that the removable file system has a high chance to remain in a fully clean state even when it is unplugged abruptly, and returns to a clean state on the next re-plug. This is a follow-up for #2471, as it adds a simple client-side for the transient automount logic added in that PR. In later work it might make sense to invoke this tool automatically from udev rules in order to implement a simpler and safer version of removable media management á la udisks.
559 lines
20 KiB
C
559 lines
20 KiB
C
/***
|
|
This file is part of systemd.
|
|
|
|
Copyright 2010 Lennart Poettering
|
|
|
|
systemd is free software; you can redistribute it and/or modify it
|
|
under the terms of the GNU Lesser General Public License as published by
|
|
the Free Software Foundation; either version 2.1 of the License, or
|
|
(at your option) any later version.
|
|
|
|
systemd is distributed in the hope that it will be useful, but
|
|
WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
|
Lesser General Public License for more details.
|
|
|
|
You should have received a copy of the GNU Lesser General Public License
|
|
along with systemd; If not, see <http://www.gnu.org/licenses/>.
|
|
***/
|
|
|
|
#include <errno.h>
|
|
#include <stdlib.h>
|
|
#include <string.h>
|
|
#include <sys/mount.h>
|
|
#include <sys/stat.h>
|
|
#include <sys/statvfs.h>
|
|
#include <unistd.h>
|
|
|
|
#include "alloc-util.h"
|
|
#include "escape.h"
|
|
#include "fd-util.h"
|
|
#include "fileio.h"
|
|
#include "hashmap.h"
|
|
#include "mount-util.h"
|
|
#include "parse-util.h"
|
|
#include "path-util.h"
|
|
#include "set.h"
|
|
#include "stdio-util.h"
|
|
#include "string-util.h"
|
|
|
|
static int fd_fdinfo_mnt_id(int fd, const char *filename, int flags, int *mnt_id) {
|
|
char path[strlen("/proc/self/fdinfo/") + DECIMAL_STR_MAX(int)];
|
|
_cleanup_free_ char *fdinfo = NULL;
|
|
_cleanup_close_ int subfd = -1;
|
|
char *p;
|
|
int r;
|
|
|
|
if ((flags & AT_EMPTY_PATH) && isempty(filename))
|
|
xsprintf(path, "/proc/self/fdinfo/%i", fd);
|
|
else {
|
|
subfd = openat(fd, filename, O_CLOEXEC|O_PATH);
|
|
if (subfd < 0)
|
|
return -errno;
|
|
|
|
xsprintf(path, "/proc/self/fdinfo/%i", subfd);
|
|
}
|
|
|
|
r = read_full_file(path, &fdinfo, NULL);
|
|
if (r == -ENOENT) /* The fdinfo directory is a relatively new addition */
|
|
return -EOPNOTSUPP;
|
|
if (r < 0)
|
|
return -errno;
|
|
|
|
p = startswith(fdinfo, "mnt_id:");
|
|
if (!p) {
|
|
p = strstr(fdinfo, "\nmnt_id:");
|
|
if (!p) /* The mnt_id field is a relatively new addition */
|
|
return -EOPNOTSUPP;
|
|
|
|
p += 8;
|
|
}
|
|
|
|
p += strspn(p, WHITESPACE);
|
|
p[strcspn(p, WHITESPACE)] = 0;
|
|
|
|
return safe_atoi(p, mnt_id);
|
|
}
|
|
|
|
int fd_is_mount_point(int fd, const char *filename, int flags) {
|
|
union file_handle_union h = FILE_HANDLE_INIT, h_parent = FILE_HANDLE_INIT;
|
|
int mount_id = -1, mount_id_parent = -1;
|
|
bool nosupp = false, check_st_dev = true;
|
|
struct stat a, b;
|
|
int r;
|
|
|
|
assert(fd >= 0);
|
|
assert(filename);
|
|
|
|
/* First we will try the name_to_handle_at() syscall, which
|
|
* tells us the mount id and an opaque file "handle". It is
|
|
* not supported everywhere though (kernel compile-time
|
|
* option, not all file systems are hooked up). If it works
|
|
* the mount id is usually good enough to tell us whether
|
|
* something is a mount point.
|
|
*
|
|
* If that didn't work we will try to read the mount id from
|
|
* /proc/self/fdinfo/<fd>. This is almost as good as
|
|
* name_to_handle_at(), however, does not return the
|
|
* opaque file handle. The opaque file handle is pretty useful
|
|
* to detect the root directory, which we should always
|
|
* consider a mount point. Hence we use this only as
|
|
* fallback. Exporting the mnt_id in fdinfo is a pretty recent
|
|
* kernel addition.
|
|
*
|
|
* As last fallback we do traditional fstat() based st_dev
|
|
* comparisons. This is how things were traditionally done,
|
|
* but unionfs breaks this since it exposes file
|
|
* systems with a variety of st_dev reported. Also, btrfs
|
|
* subvolumes have different st_dev, even though they aren't
|
|
* real mounts of their own. */
|
|
|
|
r = name_to_handle_at(fd, filename, &h.handle, &mount_id, flags);
|
|
if (r < 0) {
|
|
if (errno == ENOSYS)
|
|
/* This kernel does not support name_to_handle_at()
|
|
* fall back to simpler logic. */
|
|
goto fallback_fdinfo;
|
|
else if (errno == EOPNOTSUPP)
|
|
/* This kernel or file system does not support
|
|
* name_to_handle_at(), hence let's see if the
|
|
* upper fs supports it (in which case it is a
|
|
* mount point), otherwise fallback to the
|
|
* traditional stat() logic */
|
|
nosupp = true;
|
|
else
|
|
return -errno;
|
|
}
|
|
|
|
r = name_to_handle_at(fd, "", &h_parent.handle, &mount_id_parent, AT_EMPTY_PATH);
|
|
if (r < 0) {
|
|
if (errno == EOPNOTSUPP) {
|
|
if (nosupp)
|
|
/* Neither parent nor child do name_to_handle_at()?
|
|
We have no choice but to fall back. */
|
|
goto fallback_fdinfo;
|
|
else
|
|
/* The parent can't do name_to_handle_at() but the
|
|
* directory we are interested in can?
|
|
* If so, it must be a mount point. */
|
|
return 1;
|
|
} else
|
|
return -errno;
|
|
}
|
|
|
|
/* The parent can do name_to_handle_at() but the
|
|
* directory we are interested in can't? If so, it
|
|
* must be a mount point. */
|
|
if (nosupp)
|
|
return 1;
|
|
|
|
/* If the file handle for the directory we are
|
|
* interested in and its parent are identical, we
|
|
* assume this is the root directory, which is a mount
|
|
* point. */
|
|
|
|
if (h.handle.handle_bytes == h_parent.handle.handle_bytes &&
|
|
h.handle.handle_type == h_parent.handle.handle_type &&
|
|
memcmp(h.handle.f_handle, h_parent.handle.f_handle, h.handle.handle_bytes) == 0)
|
|
return 1;
|
|
|
|
return mount_id != mount_id_parent;
|
|
|
|
fallback_fdinfo:
|
|
r = fd_fdinfo_mnt_id(fd, filename, flags, &mount_id);
|
|
if (r == -EOPNOTSUPP)
|
|
goto fallback_fstat;
|
|
if (r < 0)
|
|
return r;
|
|
|
|
r = fd_fdinfo_mnt_id(fd, "", AT_EMPTY_PATH, &mount_id_parent);
|
|
if (r < 0)
|
|
return r;
|
|
|
|
if (mount_id != mount_id_parent)
|
|
return 1;
|
|
|
|
/* Hmm, so, the mount ids are the same. This leaves one
|
|
* special case though for the root file system. For that,
|
|
* let's see if the parent directory has the same inode as we
|
|
* are interested in. Hence, let's also do fstat() checks now,
|
|
* too, but avoid the st_dev comparisons, since they aren't
|
|
* that useful on unionfs mounts. */
|
|
check_st_dev = false;
|
|
|
|
fallback_fstat:
|
|
/* yay for fstatat() taking a different set of flags than the other
|
|
* _at() above */
|
|
if (flags & AT_SYMLINK_FOLLOW)
|
|
flags &= ~AT_SYMLINK_FOLLOW;
|
|
else
|
|
flags |= AT_SYMLINK_NOFOLLOW;
|
|
if (fstatat(fd, filename, &a, flags) < 0)
|
|
return -errno;
|
|
|
|
if (fstatat(fd, "", &b, AT_EMPTY_PATH) < 0)
|
|
return -errno;
|
|
|
|
/* A directory with same device and inode as its parent? Must
|
|
* be the root directory */
|
|
if (a.st_dev == b.st_dev &&
|
|
a.st_ino == b.st_ino)
|
|
return 1;
|
|
|
|
return check_st_dev && (a.st_dev != b.st_dev);
|
|
}
|
|
|
|
/* flags can be AT_SYMLINK_FOLLOW or 0 */
|
|
int path_is_mount_point(const char *t, int flags) {
|
|
_cleanup_close_ int fd = -1;
|
|
_cleanup_free_ char *canonical = NULL, *parent = NULL;
|
|
|
|
assert(t);
|
|
|
|
if (path_equal(t, "/"))
|
|
return 1;
|
|
|
|
/* we need to resolve symlinks manually, we can't just rely on
|
|
* fd_is_mount_point() to do that for us; if we have a structure like
|
|
* /bin -> /usr/bin/ and /usr is a mount point, then the parent that we
|
|
* look at needs to be /usr, not /. */
|
|
if (flags & AT_SYMLINK_FOLLOW) {
|
|
canonical = canonicalize_file_name(t);
|
|
if (!canonical)
|
|
return -errno;
|
|
|
|
t = canonical;
|
|
}
|
|
|
|
parent = dirname_malloc(t);
|
|
if (!parent)
|
|
return -ENOMEM;
|
|
|
|
fd = openat(AT_FDCWD, parent, O_DIRECTORY|O_CLOEXEC|O_PATH);
|
|
if (fd < 0)
|
|
return -errno;
|
|
|
|
return fd_is_mount_point(fd, basename(t), flags);
|
|
}
|
|
|
|
int umount_recursive(const char *prefix, int flags) {
|
|
bool again;
|
|
int n = 0, r;
|
|
|
|
/* Try to umount everything recursively below a
|
|
* directory. Also, take care of stacked mounts, and keep
|
|
* unmounting them until they are gone. */
|
|
|
|
do {
|
|
_cleanup_fclose_ FILE *proc_self_mountinfo = NULL;
|
|
|
|
again = false;
|
|
r = 0;
|
|
|
|
proc_self_mountinfo = fopen("/proc/self/mountinfo", "re");
|
|
if (!proc_self_mountinfo)
|
|
return -errno;
|
|
|
|
for (;;) {
|
|
_cleanup_free_ char *path = NULL, *p = NULL;
|
|
int k;
|
|
|
|
k = fscanf(proc_self_mountinfo,
|
|
"%*s " /* (1) mount id */
|
|
"%*s " /* (2) parent id */
|
|
"%*s " /* (3) major:minor */
|
|
"%*s " /* (4) root */
|
|
"%ms " /* (5) mount point */
|
|
"%*s" /* (6) mount options */
|
|
"%*[^-]" /* (7) optional fields */
|
|
"- " /* (8) separator */
|
|
"%*s " /* (9) file system type */
|
|
"%*s" /* (10) mount source */
|
|
"%*s" /* (11) mount options 2 */
|
|
"%*[^\n]", /* some rubbish at the end */
|
|
&path);
|
|
if (k != 1) {
|
|
if (k == EOF)
|
|
break;
|
|
|
|
continue;
|
|
}
|
|
|
|
r = cunescape(path, UNESCAPE_RELAX, &p);
|
|
if (r < 0)
|
|
return r;
|
|
|
|
if (!path_startswith(p, prefix))
|
|
continue;
|
|
|
|
if (umount2(p, flags) < 0) {
|
|
r = -errno;
|
|
continue;
|
|
}
|
|
|
|
again = true;
|
|
n++;
|
|
|
|
break;
|
|
}
|
|
|
|
} while (again);
|
|
|
|
return r ? r : n;
|
|
}
|
|
|
|
static int get_mount_flags(const char *path, unsigned long *flags) {
|
|
struct statvfs buf;
|
|
|
|
if (statvfs(path, &buf) < 0)
|
|
return -errno;
|
|
*flags = buf.f_flag;
|
|
return 0;
|
|
}
|
|
|
|
int bind_remount_recursive(const char *prefix, bool ro) {
|
|
_cleanup_set_free_free_ Set *done = NULL;
|
|
_cleanup_free_ char *cleaned = NULL;
|
|
int r;
|
|
|
|
/* Recursively remount a directory (and all its submounts)
|
|
* read-only or read-write. If the directory is already
|
|
* mounted, we reuse the mount and simply mark it
|
|
* MS_BIND|MS_RDONLY (or remove the MS_RDONLY for read-write
|
|
* operation). If it isn't we first make it one. Afterwards we
|
|
* apply MS_BIND|MS_RDONLY (or remove MS_RDONLY) to all
|
|
* submounts we can access, too. When mounts are stacked on
|
|
* the same mount point we only care for each individual
|
|
* "top-level" mount on each point, as we cannot
|
|
* influence/access the underlying mounts anyway. We do not
|
|
* have any effect on future submounts that might get
|
|
* propagated, they migt be writable. This includes future
|
|
* submounts that have been triggered via autofs. */
|
|
|
|
cleaned = strdup(prefix);
|
|
if (!cleaned)
|
|
return -ENOMEM;
|
|
|
|
path_kill_slashes(cleaned);
|
|
|
|
done = set_new(&string_hash_ops);
|
|
if (!done)
|
|
return -ENOMEM;
|
|
|
|
for (;;) {
|
|
_cleanup_fclose_ FILE *proc_self_mountinfo = NULL;
|
|
_cleanup_set_free_free_ Set *todo = NULL;
|
|
bool top_autofs = false;
|
|
char *x;
|
|
unsigned long orig_flags;
|
|
|
|
todo = set_new(&string_hash_ops);
|
|
if (!todo)
|
|
return -ENOMEM;
|
|
|
|
proc_self_mountinfo = fopen("/proc/self/mountinfo", "re");
|
|
if (!proc_self_mountinfo)
|
|
return -errno;
|
|
|
|
for (;;) {
|
|
_cleanup_free_ char *path = NULL, *p = NULL, *type = NULL;
|
|
int k;
|
|
|
|
k = fscanf(proc_self_mountinfo,
|
|
"%*s " /* (1) mount id */
|
|
"%*s " /* (2) parent id */
|
|
"%*s " /* (3) major:minor */
|
|
"%*s " /* (4) root */
|
|
"%ms " /* (5) mount point */
|
|
"%*s" /* (6) mount options (superblock) */
|
|
"%*[^-]" /* (7) optional fields */
|
|
"- " /* (8) separator */
|
|
"%ms " /* (9) file system type */
|
|
"%*s" /* (10) mount source */
|
|
"%*s" /* (11) mount options (bind mount) */
|
|
"%*[^\n]", /* some rubbish at the end */
|
|
&path,
|
|
&type);
|
|
if (k != 2) {
|
|
if (k == EOF)
|
|
break;
|
|
|
|
continue;
|
|
}
|
|
|
|
r = cunescape(path, UNESCAPE_RELAX, &p);
|
|
if (r < 0)
|
|
return r;
|
|
|
|
/* Let's ignore autofs mounts. If they aren't
|
|
* triggered yet, we want to avoid triggering
|
|
* them, as we don't make any guarantees for
|
|
* future submounts anyway. If they are
|
|
* already triggered, then we will find
|
|
* another entry for this. */
|
|
if (streq(type, "autofs")) {
|
|
top_autofs = top_autofs || path_equal(cleaned, p);
|
|
continue;
|
|
}
|
|
|
|
if (path_startswith(p, cleaned) &&
|
|
!set_contains(done, p)) {
|
|
|
|
r = set_consume(todo, p);
|
|
p = NULL;
|
|
|
|
if (r == -EEXIST)
|
|
continue;
|
|
if (r < 0)
|
|
return r;
|
|
}
|
|
}
|
|
|
|
/* If we have no submounts to process anymore and if
|
|
* the root is either already done, or an autofs, we
|
|
* are done */
|
|
if (set_isempty(todo) &&
|
|
(top_autofs || set_contains(done, cleaned)))
|
|
return 0;
|
|
|
|
if (!set_contains(done, cleaned) &&
|
|
!set_contains(todo, cleaned)) {
|
|
/* The prefix directory itself is not yet a
|
|
* mount, make it one. */
|
|
if (mount(cleaned, cleaned, NULL, MS_BIND|MS_REC, NULL) < 0)
|
|
return -errno;
|
|
|
|
orig_flags = 0;
|
|
(void) get_mount_flags(cleaned, &orig_flags);
|
|
orig_flags &= ~MS_RDONLY;
|
|
|
|
if (mount(NULL, prefix, NULL, orig_flags|MS_BIND|MS_REMOUNT|(ro ? MS_RDONLY : 0), NULL) < 0)
|
|
return -errno;
|
|
|
|
x = strdup(cleaned);
|
|
if (!x)
|
|
return -ENOMEM;
|
|
|
|
r = set_consume(done, x);
|
|
if (r < 0)
|
|
return r;
|
|
}
|
|
|
|
while ((x = set_steal_first(todo))) {
|
|
|
|
r = set_consume(done, x);
|
|
if (r == -EEXIST || r == 0)
|
|
continue;
|
|
if (r < 0)
|
|
return r;
|
|
|
|
/* Deal with mount points that are obstructed by a
|
|
* later mount */
|
|
r = path_is_mount_point(x, 0);
|
|
if (r == -ENOENT || r == 0)
|
|
continue;
|
|
if (r < 0)
|
|
return r;
|
|
|
|
/* Try to reuse the original flag set */
|
|
orig_flags = 0;
|
|
(void) get_mount_flags(x, &orig_flags);
|
|
orig_flags &= ~MS_RDONLY;
|
|
|
|
if (mount(NULL, x, NULL, orig_flags|MS_BIND|MS_REMOUNT|(ro ? MS_RDONLY : 0), NULL) < 0)
|
|
return -errno;
|
|
|
|
}
|
|
}
|
|
}
|
|
|
|
int mount_move_root(const char *path) {
|
|
assert(path);
|
|
|
|
if (chdir(path) < 0)
|
|
return -errno;
|
|
|
|
if (mount(path, "/", NULL, MS_MOVE, NULL) < 0)
|
|
return -errno;
|
|
|
|
if (chroot(".") < 0)
|
|
return -errno;
|
|
|
|
if (chdir("/") < 0)
|
|
return -errno;
|
|
|
|
return 0;
|
|
}
|
|
|
|
bool fstype_is_network(const char *fstype) {
|
|
static const char table[] =
|
|
"afs\0"
|
|
"cifs\0"
|
|
"smbfs\0"
|
|
"sshfs\0"
|
|
"ncpfs\0"
|
|
"ncp\0"
|
|
"nfs\0"
|
|
"nfs4\0"
|
|
"gfs\0"
|
|
"gfs2\0"
|
|
"glusterfs\0"
|
|
"pvfs2\0" /* OrangeFS */
|
|
"ocfs2\0"
|
|
;
|
|
|
|
const char *x;
|
|
|
|
x = startswith(fstype, "fuse.");
|
|
if (x)
|
|
fstype = x;
|
|
|
|
return nulstr_contains(table, fstype);
|
|
}
|
|
|
|
int repeat_unmount(const char *path, int flags) {
|
|
bool done = false;
|
|
|
|
assert(path);
|
|
|
|
/* If there are multiple mounts on a mount point, this
|
|
* removes them all */
|
|
|
|
for (;;) {
|
|
if (umount2(path, flags) < 0) {
|
|
|
|
if (errno == EINVAL)
|
|
return done;
|
|
|
|
return -errno;
|
|
}
|
|
|
|
done = true;
|
|
}
|
|
}
|
|
|
|
const char* mode_to_inaccessible_node(mode_t mode) {
|
|
/* This function maps a node type to the correspondent inaccessible node type.
|
|
* Character and block inaccessible devices may not be created (because major=0 and minor=0),
|
|
* in such case we map character and block devices to the inaccessible node type socket. */
|
|
switch(mode & S_IFMT) {
|
|
case S_IFREG:
|
|
return "/run/systemd/inaccessible/reg";
|
|
case S_IFDIR:
|
|
return "/run/systemd/inaccessible/dir";
|
|
case S_IFCHR:
|
|
if (access("/run/systemd/inaccessible/chr", F_OK) == 0)
|
|
return "/run/systemd/inaccessible/chr";
|
|
return "/run/systemd/inaccessible/sock";
|
|
case S_IFBLK:
|
|
if (access("/run/systemd/inaccessible/blk", F_OK) == 0)
|
|
return "/run/systemd/inaccessible/blk";
|
|
return "/run/systemd/inaccessible/sock";
|
|
case S_IFIFO:
|
|
return "/run/systemd/inaccessible/fifo";
|
|
case S_IFSOCK:
|
|
return "/run/systemd/inaccessible/sock";
|
|
}
|
|
return NULL;
|
|
}
|