shaba/systemd
1
0
Fork 0
mirror of https://github.com/systemd/systemd.git synced 2024-11-07 01:27:11 +03:00
Code
ca4d708dc4
systemd/src/test/test-mount-util.c

259 lines
9.2 KiB
C
Raw Normal View History

Add SPDX license identifiers to source files under the LGPL This follows what the kernel is doing, c.f. https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=5fd54ace4721fc5ce2bb5aef6318fcf17f421460.
2017-11-18 19:09:20 +03:00
/* SPDX-License-Identifier: LGPL-2.1+ */
core: hook up MountFlags= to the transient unit logic This makes "systemd-run -p MountFlags=shared -t /bin/sh" work, by making MountFlags= to the list of properties that may be accessed transiently.
2016-11-22 22:19:08 +03:00
/***
This file is part of systemd.
Copyright 2016 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 <sys/mount.h>
mount-util: add new path_get_mnt_id() call that queries the mnt ID of a path This is a simple wrapper around name_to_handle_at_loop() and fd_fdinfo_mnt_id() to query the mnt ID of a path. It uses name_to_handle_at() where it can, and falls back to to fd_fdinfo_mnt_id() where that doesn't work. This is a best-effort thing of course, since neither name_to_handle_at() nor the fdinfo logic work on all kernels.
2017-11-20 18:05:41 +03:00
#include "alloc-util.h"
#include "def.h"
#include "fd-util.h"
#include "fileio.h"
#include "hashmap.h"
core: hook up MountFlags= to the transient unit logic This makes "systemd-run -p MountFlags=shared -t /bin/sh" work, by making MountFlags= to the list of properties that may be accessed transiently.
2016-11-22 22:19:08 +03:00
#include "log.h"
test-mount-util: move test_path_is_mount_point here path-util.c and mount-util.c are intertwined, but path_is_mount_point() is defined in mount-util.c. No functional difference.
2017-10-31 12:52:30 +03:00
#include "log.h"
core: hook up MountFlags= to the transient unit logic This makes "systemd-run -p MountFlags=shared -t /bin/sh" work, by making MountFlags= to the list of properties that may be accessed transiently.
2016-11-22 22:19:08 +03:00
#include "mount-util.h"
test: fix test-mount-util when handling duplicate mounts on the same location The test was written so far under the assumption that if two mounts are placed onto the same location the "upper" mount is listed later in /proc/self/mountinfo. This appears not to be guaranteed however, as running the tests in a normal nspawn shows. This patch fixes that: it reverses the hashmap of mounts we build: instead of keying by path, we key by mnt_id, and if we notice that path_get_mnt_id() doesn't match what a line in /proc/self/mountinfo says, we use the returned ID to check if maybe another line agrees. Fixes: #7431
2017-11-23 14:48:42 +03:00
#include "path-util.h"
test-mount-util: move test_path_is_mount_point here path-util.c and mount-util.c are intertwined, but path_is_mount_point() is defined in mount-util.c. No functional difference.
2017-10-31 12:52:30 +03:00
#include "rm-rf.h"
core: hook up MountFlags= to the transient unit logic This makes "systemd-run -p MountFlags=shared -t /bin/sh" work, by making MountFlags= to the list of properties that may be accessed transiently.
2016-11-22 22:19:08 +03:00
#include "string-util.h"
Modify mount_propagation_flags_from_string to return a normal int code This means that callers can distiguish an error from flags==0, and don't have to special-case the empty string.
2016-12-03 21:57:42 +03:00
static void test_mount_propagation_flags(const char *name, int ret, unsigned long expected) {
long unsigned flags;
core: hook up MountFlags= to the transient unit logic This makes "systemd-run -p MountFlags=shared -t /bin/sh" work, by making MountFlags= to the list of properties that may be accessed transiently.
2016-11-22 22:19:08 +03:00
tree-wide: do not work in assert() (#7271) Fixes #7270.
2017-11-07 16:04:20 +03:00
assert_se(mount_propagation_flags_from_string(name, &flags) == ret);
Modify mount_propagation_flags_from_string to return a normal int code This means that callers can distiguish an error from flags==0, and don't have to special-case the empty string.
2016-12-03 21:57:42 +03:00
if (ret >= 0) {
const char *c;
assert_se(flags == expected);
c = mount_propagation_flags_to_string(flags);
if (isempty(name))
assert_se(isempty(c));
else
assert_se(streq(c, name));
}
core: hook up MountFlags= to the transient unit logic This makes "systemd-run -p MountFlags=shared -t /bin/sh" work, by making MountFlags= to the list of properties that may be accessed transiently.
2016-11-22 22:19:08 +03:00
}
mount-util: add new path_get_mnt_id() call that queries the mnt ID of a path This is a simple wrapper around name_to_handle_at_loop() and fd_fdinfo_mnt_id() to query the mnt ID of a path. It uses name_to_handle_at() where it can, and falls back to to fd_fdinfo_mnt_id() where that doesn't work. This is a best-effort thing of course, since neither name_to_handle_at() nor the fdinfo logic work on all kernels.
2017-11-20 18:05:41 +03:00
static void test_mnt_id(void) {
_cleanup_fclose_ FILE *f = NULL;
Hashmap *h;
Iterator i;
test: fix test-mount-util when handling duplicate mounts on the same location The test was written so far under the assumption that if two mounts are placed onto the same location the "upper" mount is listed later in /proc/self/mountinfo. This appears not to be guaranteed however, as running the tests in a normal nspawn shows. This patch fixes that: it reverses the hashmap of mounts we build: instead of keying by path, we key by mnt_id, and if we notice that path_get_mnt_id() doesn't match what a line in /proc/self/mountinfo says, we use the returned ID to check if maybe another line agrees. Fixes: #7431
2017-11-23 14:48:42 +03:00
char *p;
void *k;
mount-util: add new path_get_mnt_id() call that queries the mnt ID of a path This is a simple wrapper around name_to_handle_at_loop() and fd_fdinfo_mnt_id() to query the mnt ID of a path. It uses name_to_handle_at() where it can, and falls back to to fd_fdinfo_mnt_id() where that doesn't work. This is a best-effort thing of course, since neither name_to_handle_at() nor the fdinfo logic work on all kernels.
2017-11-20 18:05:41 +03:00
int r;
assert_se(f = fopen("/proc/self/mountinfo", "re"));
test: fix test-mount-util when handling duplicate mounts on the same location The test was written so far under the assumption that if two mounts are placed onto the same location the "upper" mount is listed later in /proc/self/mountinfo. This appears not to be guaranteed however, as running the tests in a normal nspawn shows. This patch fixes that: it reverses the hashmap of mounts we build: instead of keying by path, we key by mnt_id, and if we notice that path_get_mnt_id() doesn't match what a line in /proc/self/mountinfo says, we use the returned ID to check if maybe another line agrees. Fixes: #7431
2017-11-23 14:48:42 +03:00
assert_se(h = hashmap_new(&trivial_hash_ops));
mount-util: add new path_get_mnt_id() call that queries the mnt ID of a path This is a simple wrapper around name_to_handle_at_loop() and fd_fdinfo_mnt_id() to query the mnt ID of a path. It uses name_to_handle_at() where it can, and falls back to to fd_fdinfo_mnt_id() where that doesn't work. This is a best-effort thing of course, since neither name_to_handle_at() nor the fdinfo logic work on all kernels.
2017-11-20 18:05:41 +03:00
for (;;) {
_cleanup_free_ char *line = NULL, *path = NULL;
int mnt_id;
r = read_line(f, LONG_LINE_MAX, &line);
if (r == 0)
break;
assert_se(r > 0);
assert_se(sscanf(line, "%i %*s %*s %*s %ms", &mnt_id, &path) == 2);
test: fix test-mount-util when handling duplicate mounts on the same location The test was written so far under the assumption that if two mounts are placed onto the same location the "upper" mount is listed later in /proc/self/mountinfo. This appears not to be guaranteed however, as running the tests in a normal nspawn shows. This patch fixes that: it reverses the hashmap of mounts we build: instead of keying by path, we key by mnt_id, and if we notice that path_get_mnt_id() doesn't match what a line in /proc/self/mountinfo says, we use the returned ID to check if maybe another line agrees. Fixes: #7431
2017-11-23 14:48:42 +03:00
assert_se(hashmap_put(h, INT_TO_PTR(mnt_id), path) >= 0);
mount-util: add new path_get_mnt_id() call that queries the mnt ID of a path This is a simple wrapper around name_to_handle_at_loop() and fd_fdinfo_mnt_id() to query the mnt ID of a path. It uses name_to_handle_at() where it can, and falls back to to fd_fdinfo_mnt_id() where that doesn't work. This is a best-effort thing of course, since neither name_to_handle_at() nor the fdinfo logic work on all kernels.
2017-11-20 18:05:41 +03:00
path = NULL;
}
HASHMAP_FOREACH_KEY(p, k, h, i) {
test: fix test-mount-util when handling duplicate mounts on the same location The test was written so far under the assumption that if two mounts are placed onto the same location the "upper" mount is listed later in /proc/self/mountinfo. This appears not to be guaranteed however, as running the tests in a normal nspawn shows. This patch fixes that: it reverses the hashmap of mounts we build: instead of keying by path, we key by mnt_id, and if we notice that path_get_mnt_id() doesn't match what a line in /proc/self/mountinfo says, we use the returned ID to check if maybe another line agrees. Fixes: #7431
2017-11-23 14:48:42 +03:00
int mnt_id = PTR_TO_INT(k), mnt_id2;
mount-util: add new path_get_mnt_id() call that queries the mnt ID of a path This is a simple wrapper around name_to_handle_at_loop() and fd_fdinfo_mnt_id() to query the mnt ID of a path. It uses name_to_handle_at() where it can, and falls back to to fd_fdinfo_mnt_id() where that doesn't work. This is a best-effort thing of course, since neither name_to_handle_at() nor the fdinfo logic work on all kernels.
2017-11-20 18:05:41 +03:00
test: fix test-mount-util when handling duplicate mounts on the same location The test was written so far under the assumption that if two mounts are placed onto the same location the "upper" mount is listed later in /proc/self/mountinfo. This appears not to be guaranteed however, as running the tests in a normal nspawn shows. This patch fixes that: it reverses the hashmap of mounts we build: instead of keying by path, we key by mnt_id, and if we notice that path_get_mnt_id() doesn't match what a line in /proc/self/mountinfo says, we use the returned ID to check if maybe another line agrees. Fixes: #7431
2017-11-23 14:48:42 +03:00
r = path_get_mnt_id(p, &mnt_id2);
mount-util: add new path_get_mnt_id() call that queries the mnt ID of a path This is a simple wrapper around name_to_handle_at_loop() and fd_fdinfo_mnt_id() to query the mnt ID of a path. It uses name_to_handle_at() where it can, and falls back to to fd_fdinfo_mnt_id() where that doesn't work. This is a best-effort thing of course, since neither name_to_handle_at() nor the fdinfo logic work on all kernels.
2017-11-20 18:05:41 +03:00
if (r == -EOPNOTSUPP) { /* kernel or file system too old? */
test: fix test-mount-util when handling duplicate mounts on the same location The test was written so far under the assumption that if two mounts are placed onto the same location the "upper" mount is listed later in /proc/self/mountinfo. This appears not to be guaranteed however, as running the tests in a normal nspawn shows. This patch fixes that: it reverses the hashmap of mounts we build: instead of keying by path, we key by mnt_id, and if we notice that path_get_mnt_id() doesn't match what a line in /proc/self/mountinfo says, we use the returned ID to check if maybe another line agrees. Fixes: #7431
2017-11-23 14:48:42 +03:00
log_debug("%s doesn't support mount IDs\n", p);
mount-util: add new path_get_mnt_id() call that queries the mnt ID of a path This is a simple wrapper around name_to_handle_at_loop() and fd_fdinfo_mnt_id() to query the mnt ID of a path. It uses name_to_handle_at() where it can, and falls back to to fd_fdinfo_mnt_id() where that doesn't work. This is a best-effort thing of course, since neither name_to_handle_at() nor the fdinfo logic work on all kernels.
2017-11-20 18:05:41 +03:00
continue;
}
if (IN_SET(r, -EACCES, -EPERM)) {
test: fix test-mount-util when handling duplicate mounts on the same location The test was written so far under the assumption that if two mounts are placed onto the same location the "upper" mount is listed later in /proc/self/mountinfo. This appears not to be guaranteed however, as running the tests in a normal nspawn shows. This patch fixes that: it reverses the hashmap of mounts we build: instead of keying by path, we key by mnt_id, and if we notice that path_get_mnt_id() doesn't match what a line in /proc/self/mountinfo says, we use the returned ID to check if maybe another line agrees. Fixes: #7431
2017-11-23 14:48:42 +03:00
log_debug("Can't access %s\n", p);
mount-util: add new path_get_mnt_id() call that queries the mnt ID of a path This is a simple wrapper around name_to_handle_at_loop() and fd_fdinfo_mnt_id() to query the mnt ID of a path. It uses name_to_handle_at() where it can, and falls back to to fd_fdinfo_mnt_id() where that doesn't work. This is a best-effort thing of course, since neither name_to_handle_at() nor the fdinfo logic work on all kernels.
2017-11-20 18:05:41 +03:00
continue;
}
test: fix test-mount-util when handling duplicate mounts on the same location The test was written so far under the assumption that if two mounts are placed onto the same location the "upper" mount is listed later in /proc/self/mountinfo. This appears not to be guaranteed however, as running the tests in a normal nspawn shows. This patch fixes that: it reverses the hashmap of mounts we build: instead of keying by path, we key by mnt_id, and if we notice that path_get_mnt_id() doesn't match what a line in /proc/self/mountinfo says, we use the returned ID to check if maybe another line agrees. Fixes: #7431
2017-11-23 14:48:42 +03:00
log_debug("mnt id of %s is %i\n", p, mnt_id2);
if (mnt_id == mnt_id2)
continue;
mount-util: add new path_get_mnt_id() call that queries the mnt ID of a path This is a simple wrapper around name_to_handle_at_loop() and fd_fdinfo_mnt_id() to query the mnt ID of a path. It uses name_to_handle_at() where it can, and falls back to to fd_fdinfo_mnt_id() where that doesn't work. This is a best-effort thing of course, since neither name_to_handle_at() nor the fdinfo logic work on all kernels.
2017-11-20 18:05:41 +03:00
test: fix test-mount-util when handling duplicate mounts on the same location The test was written so far under the assumption that if two mounts are placed onto the same location the "upper" mount is listed later in /proc/self/mountinfo. This appears not to be guaranteed however, as running the tests in a normal nspawn shows. This patch fixes that: it reverses the hashmap of mounts we build: instead of keying by path, we key by mnt_id, and if we notice that path_get_mnt_id() doesn't match what a line in /proc/self/mountinfo says, we use the returned ID to check if maybe another line agrees. Fixes: #7431
2017-11-23 14:48:42 +03:00
/* The ids don't match? If so, then there are two mounts on the same path, let's check if that's really
* the case */
assert_se(path_equal_ptr(hashmap_get(h, INT_TO_PTR(mnt_id2)), p));
mount-util: add new path_get_mnt_id() call that queries the mnt ID of a path This is a simple wrapper around name_to_handle_at_loop() and fd_fdinfo_mnt_id() to query the mnt ID of a path. It uses name_to_handle_at() where it can, and falls back to to fd_fdinfo_mnt_id() where that doesn't work. This is a best-effort thing of course, since neither name_to_handle_at() nor the fdinfo logic work on all kernels.
2017-11-20 18:05:41 +03:00
}
Use hashmap_free_free where appropriate
2017-11-28 14:31:27 +03:00
hashmap_free_free(h);
mount-util: add new path_get_mnt_id() call that queries the mnt ID of a path This is a simple wrapper around name_to_handle_at_loop() and fd_fdinfo_mnt_id() to query the mnt ID of a path. It uses name_to_handle_at() where it can, and falls back to to fd_fdinfo_mnt_id() where that doesn't work. This is a best-effort thing of course, since neither name_to_handle_at() nor the fdinfo logic work on all kernels.
2017-11-20 18:05:41 +03:00
}
test-mount-util: move test_path_is_mount_point here path-util.c and mount-util.c are intertwined, but path_is_mount_point() is defined in mount-util.c. No functional difference.
2017-10-31 12:52:30 +03:00
static void test_path_is_mount_point(void) {
int fd;
char tmp_dir[] = "/tmp/test-path-is-mount-point-XXXXXX";
_cleanup_free_ char *file1 = NULL, *file2 = NULL, *link1 = NULL, *link2 = NULL;
_cleanup_free_ char *dir1 = NULL, *dir1file = NULL, *dirlink1 = NULL, *dirlink1file = NULL;
_cleanup_free_ char *dir2 = NULL, *dir2file = NULL;
assert_se(path_is_mount_point("/", NULL, AT_SYMLINK_FOLLOW) > 0);
assert_se(path_is_mount_point("/", NULL, 0) > 0);
assert_se(path_is_mount_point("/proc", NULL, AT_SYMLINK_FOLLOW) > 0);
assert_se(path_is_mount_point("/proc", NULL, 0) > 0);
assert_se(path_is_mount_point("/proc/1", NULL, AT_SYMLINK_FOLLOW) == 0);
assert_se(path_is_mount_point("/proc/1", NULL, 0) == 0);
assert_se(path_is_mount_point("/sys", NULL, AT_SYMLINK_FOLLOW) > 0);
assert_se(path_is_mount_point("/sys", NULL, 0) > 0);
/* we'll create a hierarchy of different kinds of dir/file/link
* layouts:
*
* <tmp>/file1, <tmp>/file2
* <tmp>/link1 -> file1, <tmp>/link2 -> file2
* <tmp>/dir1/
* <tmp>/dir1/file
* <tmp>/dirlink1 -> dir1
* <tmp>/dirlink1file -> dirlink1/file
* <tmp>/dir2/
* <tmp>/dir2/file
*/
/* file mountpoints */
assert_se(mkdtemp(tmp_dir) != NULL);
file1 = path_join(NULL, tmp_dir, "file1");
assert_se(file1);
file2 = path_join(NULL, tmp_dir, "file2");
assert_se(file2);
fd = open(file1, O_WRONLY|O_CREAT|O_EXCL|O_CLOEXEC, 0664);
assert_se(fd > 0);
close(fd);
fd = open(file2, O_WRONLY|O_CREAT|O_EXCL|O_CLOEXEC, 0664);
assert_se(fd > 0);
close(fd);
link1 = path_join(NULL, tmp_dir, "link1");
assert_se(link1);
assert_se(symlink("file1", link1) == 0);
link2 = path_join(NULL, tmp_dir, "link2");
assert_se(link1);
assert_se(symlink("file2", link2) == 0);
assert_se(path_is_mount_point(file1, NULL, AT_SYMLINK_FOLLOW) == 0);
assert_se(path_is_mount_point(file1, NULL, 0) == 0);
assert_se(path_is_mount_point(link1, NULL, AT_SYMLINK_FOLLOW) == 0);
assert_se(path_is_mount_point(link1, NULL, 0) == 0);
/* directory mountpoints */
dir1 = path_join(NULL, tmp_dir, "dir1");
assert_se(dir1);
assert_se(mkdir(dir1, 0755) == 0);
dirlink1 = path_join(NULL, tmp_dir, "dirlink1");
assert_se(dirlink1);
assert_se(symlink("dir1", dirlink1) == 0);
dirlink1file = path_join(NULL, tmp_dir, "dirlink1file");
assert_se(dirlink1file);
assert_se(symlink("dirlink1/file", dirlink1file) == 0);
dir2 = path_join(NULL, tmp_dir, "dir2");
assert_se(dir2);
assert_se(mkdir(dir2, 0755) == 0);
assert_se(path_is_mount_point(dir1, NULL, AT_SYMLINK_FOLLOW) == 0);
assert_se(path_is_mount_point(dir1, NULL, 0) == 0);
assert_se(path_is_mount_point(dirlink1, NULL, AT_SYMLINK_FOLLOW) == 0);
assert_se(path_is_mount_point(dirlink1, NULL, 0) == 0);
/* file in subdirectory mountpoints */
dir1file = path_join(NULL, dir1, "file");
assert_se(dir1file);
fd = open(dir1file, O_WRONLY|O_CREAT|O_EXCL|O_CLOEXEC, 0664);
assert_se(fd > 0);
close(fd);
assert_se(path_is_mount_point(dir1file, NULL, AT_SYMLINK_FOLLOW) == 0);
assert_se(path_is_mount_point(dir1file, NULL, 0) == 0);
assert_se(path_is_mount_point(dirlink1file, NULL, AT_SYMLINK_FOLLOW) == 0);
assert_se(path_is_mount_point(dirlink1file, NULL, 0) == 0);
/* these tests will only work as root */
if (mount(file1, file2, NULL, MS_BIND, NULL) >= 0) {
int rt, rf, rlt, rlf, rl1t, rl1f;
/* files */
/* capture results in vars, to avoid dangling mounts on failure */
rf = path_is_mount_point(file2, NULL, 0);
rt = path_is_mount_point(file2, NULL, AT_SYMLINK_FOLLOW);
rlf = path_is_mount_point(link2, NULL, 0);
rlt = path_is_mount_point(link2, NULL, AT_SYMLINK_FOLLOW);
assert_se(umount(file2) == 0);
assert_se(rf == 1);
assert_se(rt == 1);
assert_se(rlf == 0);
assert_se(rlt == 1);
/* dirs */
dir2file = path_join(NULL, dir2, "file");
assert_se(dir2file);
fd = open(dir2file, O_WRONLY|O_CREAT|O_EXCL|O_CLOEXEC, 0664);
assert_se(fd > 0);
close(fd);
assert_se(mount(dir2, dir1, NULL, MS_BIND, NULL) >= 0);
rf = path_is_mount_point(dir1, NULL, 0);
rt = path_is_mount_point(dir1, NULL, AT_SYMLINK_FOLLOW);
rlf = path_is_mount_point(dirlink1, NULL, 0);
rlt = path_is_mount_point(dirlink1, NULL, AT_SYMLINK_FOLLOW);
/* its parent is a mount point, but not /file itself */
rl1f = path_is_mount_point(dirlink1file, NULL, 0);
rl1t = path_is_mount_point(dirlink1file, NULL, AT_SYMLINK_FOLLOW);
assert_se(umount(dir1) == 0);
assert_se(rf == 1);
assert_se(rt == 1);
assert_se(rlf == 0);
assert_se(rlt == 1);
assert_se(rl1f == 0);
assert_se(rl1t == 0);
} else
printf("Skipping bind mount file test: %m\n");
assert_se(rm_rf(tmp_dir, REMOVE_ROOT|REMOVE_PHYSICAL) == 0);
}
core: hook up MountFlags= to the transient unit logic This makes "systemd-run -p MountFlags=shared -t /bin/sh" work, by making MountFlags= to the list of properties that may be accessed transiently.
2016-11-22 22:19:08 +03:00
int main(int argc, char *argv[]) {
log_set_max_level(LOG_DEBUG);
Modify mount_propagation_flags_from_string to return a normal int code This means that callers can distiguish an error from flags==0, and don't have to special-case the empty string.
2016-12-03 21:57:42 +03:00
test_mount_propagation_flags("shared", 0, MS_SHARED);
test_mount_propagation_flags("slave", 0, MS_SLAVE);
test_mount_propagation_flags("private", 0, MS_PRIVATE);
test_mount_propagation_flags(NULL, 0, 0);
test_mount_propagation_flags("", 0, 0);
test_mount_propagation_flags("xxxx", -EINVAL, 0);
test_mount_propagation_flags(" ", -EINVAL, 0);
core: hook up MountFlags= to the transient unit logic This makes "systemd-run -p MountFlags=shared -t /bin/sh" work, by making MountFlags= to the list of properties that may be accessed transiently.
2016-11-22 22:19:08 +03:00
mount-util: add new path_get_mnt_id() call that queries the mnt ID of a path This is a simple wrapper around name_to_handle_at_loop() and fd_fdinfo_mnt_id() to query the mnt ID of a path. It uses name_to_handle_at() where it can, and falls back to to fd_fdinfo_mnt_id() where that doesn't work. This is a best-effort thing of course, since neither name_to_handle_at() nor the fdinfo logic work on all kernels.
2017-11-20 18:05:41 +03:00
test_mnt_id();
test-mount-util: move test_path_is_mount_point here path-util.c and mount-util.c are intertwined, but path_is_mount_point() is defined in mount-util.c. No functional difference.
2017-10-31 12:52:30 +03:00
test_path_is_mount_point();
mount-util: add new path_get_mnt_id() call that queries the mnt ID of a path This is a simple wrapper around name_to_handle_at_loop() and fd_fdinfo_mnt_id() to query the mnt ID of a path. It uses name_to_handle_at() where it can, and falls back to to fd_fdinfo_mnt_id() where that doesn't work. This is a best-effort thing of course, since neither name_to_handle_at() nor the fdinfo logic work on all kernels.
2017-11-20 18:05:41 +03:00
core: hook up MountFlags= to the transient unit logic This makes "systemd-run -p MountFlags=shared -t /bin/sh" work, by making MountFlags= to the list of properties that may be accessed transiently.
2016-11-22 22:19:08 +03:00
return 0;
}
Copy Permalink