Compare commits

...

13 Commits

Author SHA1 Message Date
Lennart Poettering 1bf9c5b2bf
Merge 8fd917a74d into f6793bbcf0 2024-11-20 10:47:25 -08:00
Lennart Poettering f6793bbcf0 killall: gracefully handle processes inserted into containers via nsenter -a
"nsenter -a" doesn't migrate the specified process into the target
cgroup (it really should). Thus the cgroup will remain in a cgroup
that is (due to cgroup ns) outside our visibility. The kernel will
report the cgroup path of such cgroups as starting with "/../". Detect
that and print a reasonably error message instead of trying to resolve
that.
2024-11-20 18:11:38 +00:00
Mike Yuan f87863a8ff process-util: refuse to operate on remote PidRef
Follow-up for 7e3e540b88
2024-11-20 18:10:26 +00:00
Antonio Alvarez Feijoo 58c3c2886d cryptenroll: fix typo 2024-11-20 18:03:44 +00:00
Daan De Meyer dbbe895807 test-audit-util: Migrate to new assertion macros 2024-11-20 16:48:55 +00:00
Yu Watanabe 52b0351a15
core/exec-invoke: suppress placeholder home only in build_environment() (#35219)
Alternative to https://github.com/systemd/systemd/pull/34789
Closes #34789
2024-11-20 17:34:25 +09:00
Luca Boccassi fe077a1a58 units: add initrd directory to list of conditions for systemd-confext
systemd-sysext has the same check, but it was forgotten for confexts.
Needed to activate confexts from the ESP in the initrd.
2024-11-20 09:12:24 +01:00
Mike Yuan b718b86e1b
core/exec-invoke: suppress placeholder home only in build_environment()
Currently, get_fixed_user() employs USER_CREDS_SUPPRESS_PLACEHOLDER,
meaning home path is set to NULL if it's empty or root. However,
the path is also used for applying WorkingDirectory=~, and we'd
spuriously use the invoking user's home as fallback even if
User= is changed in that case.

Let's instead delegate such suppression to build_environment(),
so that home is proper initialized for usage at other steps.
shell doesn't actually suffer from such problem, but it's changed
too for consistency.

Alternative to #34789
2024-11-19 00:38:18 +01:00
Mike Yuan d911778877
core/exec-invoke: minor cleanup for apply_working_directory() error handling
Assign exit_status at the same site where error log is emitted,
for readability.
2024-11-19 00:38:18 +01:00
Mike Yuan eea9d3eb10
basic/user-util: split out placeholder suppression from USER_CREDS_CLEAN into its own flag
No functional change, preparation for later commits.
2024-11-19 00:38:18 +01:00
Mike Yuan 579ce77ead
basic/user-util: introduce shell_is_placeholder() helper 2024-11-19 00:38:18 +01:00
Lennart Poettering 8fd917a74d tpm2-util: allocate pcrlock NV indexes from our newly assigned range 2024-07-04 16:32:03 +02:00
Lennart Poettering 6cda26f3d6 doc: add document stating how we assign NV indexes
The Trusted Computing Group has assigned the systemd project an NV index
range, let's document properly which one that is, and how we use it.

Thank you so much, TCG!
2024-07-04 16:32:03 +02:00
14 changed files with 195 additions and 60 deletions

View File

@ -0,0 +1,70 @@
---
title: TPM2 NV Index Assignment by systemd
category: Booting
layout: default
SPDX-License-Identifier: LGPL-2.1-or-later
---
# TPM 2.0 NV Index Assignments
The Trusted Computing Group (TCG) maintains a [Registry of Reserved TPM 2.0
Handles and Localities](https://trustedcomputinggroup.org/resource/registry/)
which assigns TPM 2.0 NV index ranges (among ther things, see section 2.2) to
organizations (by convention only!). It has assigned the NV index range
**0x01800400-0x018005FF** to the systemd project. This NV index range is subdivided
and used by systemd for the following purposes:
## As Storage for a Disk Encryption PolicyAuthorizeNV Policy Hash
*Scope*: Dynamic allocation at OS installation time, one for each installed
Linux/systemd based OS that uses `systemd-pcrlock` based disk encryption policies.
*Subrange*: **0x01800400-0x0180041F**
*Number of NV Indexes*: **32**
*Size*: Stores one policy hash. Under the assumption SHA256 policy hashes are
used, this means **32 byte**.
## As Storage for Additional PCRs Implemented in NV Indexes
*Scope*: Static allocation by the systemd project, one for each additional NV
Indexed based PCR (systemd calls these "NvPCRs"). These can be shared between
multiple Linux/systemd based OSes installed on the same system.
*Subrange*: **0x01800420-0x01800423**
*Number of NV Indexes*: **4**
*Size*: Stores one PCR hash each (`TPMA_NT_EXTEND`). We'd expect that typically
SHA256 PCR hashes are used, hence this means **32 byte**.
*Detailed Assignments*:
| NVIndex | Purpose |
|------------|---------------------------------------------------------------|
| 0x01800420 | Used LUKS unlock mechanism (TPM2, PKCS11, FIDO2, …) |
| 0x01800421 | Product UUID |
| 0x01800422 | System Extension Images (sysexts) applied to the host |
| 0x01800423 | Configuration Extension Images (confexts) applied to the host |
## Currently Unused Range
The following range is currently not used by the systemd project, but might be
allocated later: **0x01800424-0x018005FF**
## Summary:
| NVIndex Range | Number | Purpose |
|-----------------------|--------|------------------------------------------------|
| 0x01800400-0x0180041F | 32 | Assigned to systemd, used for pcrlock policies |
| 0x01800420-0x01800423 | 4 | Assigned to systemd, used as additional PCRs |
| 0x01800424-0x018005FF | 476 | Assigned to systemd, currently unused |
# Relationship with TCG
This document is referenced by the aforementioned registry for details about
assignments of the NV Index range delegated to the systemd project. Hence,
particular care should be taken that this page is not moved, and its URL
remains stable as
[`https://systemd.io/TPM2_NVINDEX_ASSIGNMENTS`](https://systemd.io/TPM2_NVINDEX_ASSIGNMENTS).

View File

@ -803,6 +803,10 @@ int cg_pid_get_path(const char *controller, pid_t pid, char **ret_path) {
if (!path)
return -ENOMEM;
/* Refuse cgroup paths from outside our cgroup namespace */
if (startswith(path, "/../"))
return -EUNATCH;
/* Truncate suffix indicating the process is a zombie */
e = endswith(path, " (deleted)");
if (e)

View File

@ -102,8 +102,8 @@ int pid_get_comm(pid_t pid, char **ret) {
_cleanup_free_ char *escaped = NULL, *comm = NULL;
int r;
assert(ret);
assert(pid >= 0);
assert(ret);
if (pid == 0 || pid == getpid_cached()) {
comm = new0(char, TASK_COMM_LEN + 1); /* Must fit in 16 byte according to prctl(2) */
@ -143,6 +143,9 @@ int pidref_get_comm(const PidRef *pid, char **ret) {
if (!pidref_is_set(pid))
return -ESRCH;
if (pidref_is_remote(pid))
return -EREMOTE;
r = pid_get_comm(pid->pid, &comm);
if (r < 0)
return r;
@ -289,6 +292,9 @@ int pidref_get_cmdline(const PidRef *pid, size_t max_columns, ProcessCmdlineFlag
if (!pidref_is_set(pid))
return -ESRCH;
if (pidref_is_remote(pid))
return -EREMOTE;
r = pid_get_cmdline(pid->pid, max_columns, flags, &s);
if (r < 0)
return r;
@ -331,6 +337,9 @@ int pidref_get_cmdline_strv(const PidRef *pid, ProcessCmdlineFlags flags, char *
if (!pidref_is_set(pid))
return -ESRCH;
if (pidref_is_remote(pid))
return -EREMOTE;
r = pid_get_cmdline_strv(pid->pid, flags, &args);
if (r < 0)
return r;
@ -477,6 +486,9 @@ int pidref_is_kernel_thread(const PidRef *pid) {
if (!pidref_is_set(pid))
return -ESRCH;
if (pidref_is_remote(pid))
return -EREMOTE;
result = pid_is_kernel_thread(pid->pid);
if (result < 0)
return result;
@ -594,6 +606,9 @@ int pidref_get_uid(const PidRef *pid, uid_t *ret) {
if (!pidref_is_set(pid))
return -ESRCH;
if (pidref_is_remote(pid))
return -EREMOTE;
r = pid_get_uid(pid->pid, &uid);
if (r < 0)
return r;
@ -794,6 +809,9 @@ int pidref_get_start_time(const PidRef *pid, usec_t *ret) {
if (!pidref_is_set(pid))
return -ESRCH;
if (pidref_is_remote(pid))
return -EREMOTE;
r = pid_get_start_time(pid->pid, ret ? &t : NULL);
if (r < 0)
return r;
@ -1093,6 +1111,9 @@ int pidref_is_my_child(const PidRef *pid) {
if (!pidref_is_set(pid))
return -ESRCH;
if (pidref_is_remote(pid))
return -EREMOTE;
result = pid_is_my_child(pid->pid);
if (result < 0)
return result;
@ -1128,6 +1149,9 @@ int pidref_is_unwaited(const PidRef *pid) {
if (!pidref_is_set(pid))
return -ESRCH;
if (pidref_is_remote(pid))
return -EREMOTE;
if (pid->pid == 1 || pidref_is_self(pid))
return true;
@ -1169,6 +1193,9 @@ int pidref_is_alive(const PidRef *pidref) {
if (!pidref_is_set(pidref))
return -ESRCH;
if (pidref_is_remote(pidref))
return -EREMOTE;
result = pid_is_alive(pidref->pid);
if (result < 0) {
assert(result != -ESRCH);

View File

@ -220,9 +220,9 @@ static int synthesize_user_creds(
if (ret_gid)
*ret_gid = GID_NOBODY;
if (ret_home)
*ret_home = FLAGS_SET(flags, USER_CREDS_CLEAN) ? NULL : "/";
*ret_home = FLAGS_SET(flags, USER_CREDS_SUPPRESS_PLACEHOLDER) ? NULL : "/";
if (ret_shell)
*ret_shell = FLAGS_SET(flags, USER_CREDS_CLEAN) ? NULL : NOLOGIN;
*ret_shell = FLAGS_SET(flags, USER_CREDS_SUPPRESS_PLACEHOLDER) ? NULL : NOLOGIN;
return 0;
}
@ -244,6 +244,7 @@ int get_user_creds(
assert(username);
assert(*username);
assert((ret_home || ret_shell) || !(flags & (USER_CREDS_SUPPRESS_PLACEHOLDER|USER_CREDS_CLEAN)));
if (!FLAGS_SET(flags, USER_CREDS_PREFER_NSS) ||
(!ret_home && !ret_shell)) {
@ -315,17 +316,14 @@ int get_user_creds(
if (ret_home)
/* Note: we don't insist on normalized paths, since there are setups that have /./ in the path */
*ret_home = (FLAGS_SET(flags, USER_CREDS_CLEAN) &&
(empty_or_root(p->pw_dir) ||
!path_is_valid(p->pw_dir) ||
!path_is_absolute(p->pw_dir))) ? NULL : p->pw_dir;
*ret_home = (FLAGS_SET(flags, USER_CREDS_SUPPRESS_PLACEHOLDER) && empty_or_root(p->pw_dir)) ||
(FLAGS_SET(flags, USER_CREDS_CLEAN) && (!path_is_valid(p->pw_dir) || !path_is_absolute(p->pw_dir)))
? NULL : p->pw_dir;
if (ret_shell)
*ret_shell = (FLAGS_SET(flags, USER_CREDS_CLEAN) &&
(isempty(p->pw_shell) ||
!path_is_valid(p->pw_shell) ||
!path_is_absolute(p->pw_shell) ||
is_nologin_shell(p->pw_shell))) ? NULL : p->pw_shell;
*ret_shell = (FLAGS_SET(flags, USER_CREDS_SUPPRESS_PLACEHOLDER) && shell_is_placeholder(p->pw_shell)) ||
(FLAGS_SET(flags, USER_CREDS_CLEAN) && (!path_is_valid(p->pw_shell) || !path_is_absolute(p->pw_shell)))
? NULL : p->pw_shell;
if (patch_username)
*username = p->pw_name;

View File

@ -12,6 +12,8 @@
#include <sys/types.h>
#include <unistd.h>
#include "string-util.h"
/* Users managed by systemd-homed. See https://systemd.io/UIDS-GIDS for details how this range fits into the rest of the world */
#define HOME_UID_MIN ((uid_t) 60001)
#define HOME_UID_MAX ((uid_t) 60513)
@ -36,10 +38,20 @@ static inline int parse_gid(const char *s, gid_t *ret_gid) {
char* getlogname_malloc(void);
char* getusername_malloc(void);
const char* default_root_shell_at(int rfd);
const char* default_root_shell(const char *root);
bool is_nologin_shell(const char *shell);
static inline bool shell_is_placeholder(const char *shell) {
return isempty(shell) || is_nologin_shell(shell);
}
typedef enum UserCredsFlags {
USER_CREDS_PREFER_NSS = 1 << 0, /* if set, only synthesize user records if database lacks them. Normally we bypass the userdb entirely for the records we can synthesize */
USER_CREDS_ALLOW_MISSING = 1 << 1, /* if a numeric UID string is resolved, be OK if there's no record for it */
USER_CREDS_CLEAN = 1 << 2, /* try to clean up shell and home fields with invalid data */
USER_CREDS_SUPPRESS_PLACEHOLDER = 1 << 3, /* suppress home and/or shell fields if value is placeholder (root/empty/nologin) */
} UserCredsFlags;
int get_user_creds(const char **username, uid_t *ret_uid, gid_t *ret_gid, const char **ret_home, const char **ret_shell, UserCredsFlags flags);
@ -125,10 +137,6 @@ int fgetsgent_sane(FILE *stream, struct sgrp **sg);
int putsgent_sane(const struct sgrp *sg, FILE *stream);
#endif
bool is_nologin_shell(const char *shell);
const char* default_root_shell_at(int rfd);
const char* default_root_shell(const char *root);
int is_this_me(const char *username);
const char* get_home_root(void);

View File

@ -855,9 +855,6 @@ static int get_fixed_user(
assert(user_or_uid);
assert(ret_username);
/* Note that we don't set $HOME or $SHELL if they are not particularly enlightening anyway
* (i.e. are "/" or "/bin/nologin"). */
r = get_user_creds(&user_or_uid, ret_uid, ret_gid, ret_home, ret_shell, USER_CREDS_CLEAN);
if (r < 0)
return r;
@ -1883,7 +1880,10 @@ static int build_environment(
}
}
if (home && set_user_login_env) {
/* Note that we don't set $HOME or $SHELL if they are not particularly enlightening anyway
* (i.e. are "/" or "/bin/nologin"). */
if (home && set_user_login_env && !empty_or_root(home)) {
x = strjoin("HOME=", home);
if (!x)
return -ENOMEM;
@ -1892,7 +1892,7 @@ static int build_environment(
our_env[n_env++] = x;
}
if (shell && set_user_login_env) {
if (shell && set_user_login_env && !shell_is_placeholder(shell)) {
x = strjoin("SHELL=", shell);
if (!x)
return -ENOMEM;
@ -3471,20 +3471,16 @@ static int apply_working_directory(
const ExecContext *context,
const ExecParameters *params,
ExecRuntime *runtime,
const char *home,
int *exit_status) {
const char *home) {
const char *wd;
int r;
assert(context);
assert(exit_status);
if (context->working_directory_home) {
if (!home) {
*exit_status = EXIT_CHDIR;
if (!home)
return -ENXIO;
}
wd = home;
} else
@ -3503,13 +3499,7 @@ static int apply_working_directory(
if (r >= 0)
r = RET_NERRNO(fchdir(dfd));
}
if (r < 0 && !context->working_directory_missing_ok) {
*exit_status = EXIT_CHDIR;
return r;
}
return 0;
return context->working_directory_missing_ok ? 0 : r;
}
static int apply_root_directory(
@ -3785,7 +3775,7 @@ static int acquire_home(const ExecContext *c, const char **home, char **ret_buf)
if (!c->working_directory_home)
return 0;
if (c->dynamic_user)
if (c->dynamic_user || (c->user && is_this_me(c->user) <= 0))
return -EADDRNOTAVAIL;
r = get_home_dir(ret_buf);
@ -4543,7 +4533,7 @@ int exec_invoke(
r = acquire_home(context, &home, &home_buffer);
if (r < 0) {
*exit_status = EXIT_CHDIR;
return log_exec_error_errno(context, params, r, "Failed to determine $HOME for user: %m");
return log_exec_error_errno(context, params, r, "Failed to determine $HOME for the invoking user: %m");
}
/* If a socket is connected to STDIN/STDOUT/STDERR, we must drop O_NONBLOCK */
@ -5382,9 +5372,11 @@ int exec_invoke(
* running this service might have the correct privilege to change to the working directory. Also, it
* is absolutely 💣 crucial 💣 we applied all mount namespacing rearrangements before this, so that
* the cwd cannot be used to pin directories outside of the sandbox. */
r = apply_working_directory(context, params, runtime, home, exit_status);
if (r < 0)
r = apply_working_directory(context, params, runtime, home);
if (r < 0) {
*exit_status = EXIT_CHDIR;
return log_exec_error_errno(context, params, r, "Changing to the requested working directory failed: %m");
}
if (needs_sandboxing) {
/* Apply other MAC contexts late, but before seccomp syscall filtering, as those should really be last to

View File

@ -193,7 +193,7 @@ int enroll_fido2(
fflush(stdout);
fprintf(stderr,
"\nPlease save this FIDO2 credential ID. It is required when unloocking the volume\n"
"\nPlease save this FIDO2 credential ID. It is required when unlocking the volume\n"
"using the associated FIDO2 keyslot which we just created. To configure automatic\n"
"unlocking using this FIDO2 token, add an appropriate entry to your /etc/crypttab\n"
"file, see %s for details.\n", link);

View File

@ -2297,7 +2297,8 @@ static int start_transient_scope(sd_bus *bus) {
uid_t uid;
gid_t gid;
r = get_user_creds(&arg_exec_user, &uid, &gid, &home, &shell, USER_CREDS_CLEAN|USER_CREDS_PREFER_NSS);
r = get_user_creds(&arg_exec_user, &uid, &gid, &home, &shell,
USER_CREDS_CLEAN|USER_CREDS_SUPPRESS_PLACEHOLDER|USER_CREDS_PREFER_NSS);
if (r < 0)
return log_error_errno(r, "Failed to resolve user %s: %m", arg_exec_user);

View File

@ -46,13 +46,17 @@ static bool argv_has_at(pid_t pid) {
return c == '@';
}
static bool is_survivor_cgroup(const PidRef *pid) {
static bool is_in_survivor_cgroup(const PidRef *pid) {
_cleanup_free_ char *cgroup_path = NULL;
int r;
assert(pidref_is_set(pid));
r = cg_pidref_get_path(/* root= */ NULL, pid, &cgroup_path);
if (r == -EUNATCH) {
log_warning_errno(r, "Process " PID_FMT " appears to originate in foreign namespace, ignoring.", pid->pid);
return true;
}
if (r < 0) {
log_warning_errno(r, "Failed to get cgroup path of process " PID_FMT ", ignoring: %m", pid->pid);
return false;
@ -86,7 +90,7 @@ static bool ignore_proc(const PidRef *pid, bool warn_rootfs) {
return true; /* also ignore processes where we can't determine this */
/* Ignore processes that are part of a cgroup marked with the user.survive_final_kill_signal xattr */
if (is_survivor_cgroup(pid))
if (is_in_survivor_cgroup(pid))
return true;
r = pidref_get_uid(pid, &uid);

View File

@ -5871,9 +5871,9 @@ int tpm2_unseal(Tpm2Context *c,
return 0;
}
static TPM2_HANDLE generate_random_nv_index(void) {
return TPM2_NV_INDEX_UNASSIGNED_FIRST +
(TPM2_HANDLE) random_u64_range(TPM2_NV_INDEX_UNASSIGNED_LAST - TPM2_NV_INDEX_UNASSIGNED_FIRST + 1);
static TPM2_HANDLE generate_random_pcrlock_nv_index(void) {
return TPM2_NV_INDEX_PCRLOCK_FIRST +
(TPM2_HANDLE) random_u64_range(TPM2_NV_INDEX_PCRLOCK_LAST - TPM2_NV_INDEX_PCRLOCK_FIRST + 1);
}
int tpm2_define_policy_nv_index(
@ -5907,7 +5907,7 @@ int tpm2_define_policy_nv_index(
if (requested_nv_index != 0)
nv_index = requested_nv_index;
else
nv_index = generate_random_nv_index();
nv_index = generate_random_pcrlock_nv_index();
TPM2B_NV_PUBLIC public_info = {
.size = sizeof_field(TPM2B_NV_PUBLIC, nvPublic),

View File

@ -500,13 +500,21 @@ enum {
int tpm2_pcr_index_from_string(const char *s) _pure_;
const char* tpm2_pcr_index_to_string(int pcr) _const_;
/* The first and last NV index handle that is not registered to any company, as per TCG's "Registry of
/* The first and last NV index handle that is assigned to the systemd project as per TCG's "Registry of
* Reserved TPM 2.0 Handles and Localities", section 2.2.2. */
#define TPM2_NV_INDEX_UNASSIGNED_FIRST UINT32_C(0x01800000)
#define TPM2_NV_INDEX_UNASSIGNED_LAST UINT32_C(0x01BFFFFF)
#define TPM2_NV_INDEX_SYSTEMD_FIRST UINT32_C(0x01800400)
#define TPM2_NV_INDEX_SYSTEMD_LAST UINT32_C(0x018005FF)
#if HAVE_TPM2
/* Verify that the above is indeed a subset of the general NV Index range */
assert_cc(TPM2_NV_INDEX_UNASSIGNED_FIRST >= TPM2_NV_INDEX_FIRST);
assert_cc(TPM2_NV_INDEX_UNASSIGNED_LAST <= TPM2_NV_INDEX_LAST);
assert_cc(TPM2_NV_INDEX_SYSTEMD_FIRST >= TPM2_NV_INDEX_FIRST);
assert_cc(TPM2_NV_INDEX_SYSTEMD_LAST <= TPM2_NV_INDEX_LAST);
#endif
/* A subrange we use to store pcrlock policies in */
#define TPM2_NV_INDEX_PCRLOCK_FIRST UINT32_C(0x01800400)
#define TPM2_NV_INDEX_PCRLOCK_LAST UINT32_C(0x0180041F)
assert_cc(TPM2_NV_INDEX_PCRLOCK_FIRST >= TPM2_NV_INDEX_SYSTEMD_FIRST);
assert_cc(TPM2_NV_INDEX_PCRLOCK_LAST <= TPM2_NV_INDEX_SYSTEMD_LAST);

View File

@ -7,24 +7,26 @@ TEST(audit_loginuid_from_pid) {
_cleanup_(pidref_done) PidRef self = PIDREF_NULL, pid1 = PIDREF_NULL;
int r;
assert_se(pidref_set_self(&self) >= 0);
assert_se(pidref_set_pid(&pid1, 1) >= 0);
ASSERT_OK(pidref_set_self(&self));
ASSERT_OK(pidref_set_pid(&pid1, 1));
uid_t uid;
r = audit_loginuid_from_pid(&self, &uid);
assert_se(r >= 0 || r == -ENODATA);
if (r != -ENODATA)
ASSERT_OK(r);
if (r >= 0)
log_info("self audit login uid: " UID_FMT, uid);
assert_se(audit_loginuid_from_pid(&pid1, &uid) == -ENODATA);
ASSERT_ERROR(audit_loginuid_from_pid(&pid1, &uid), ENODATA);
uint32_t sessionid;
r = audit_session_from_pid(&self, &sessionid);
assert_se(r >= 0 || r == -ENODATA);
if (r != -ENODATA)
ASSERT_OK(r);
if (r >= 0)
log_info("self audit session id: %" PRIu32, sessionid);
assert_se(audit_session_from_pid(&pid1, &sessionid) == -ENODATA);
ASSERT_ERROR(audit_session_from_pid(&pid1, &sessionid), ENODATA);
}
static int intro(void) {

View File

@ -0,0 +1,20 @@
#!/usr/bin/env bash
# SPDX-License-Identifier: LGPL-2.1-or-later
set -eux
set -o pipefail
# shellcheck source=test/units/util.sh
. "$(dirname "$0")"/util.sh
(! systemd-run --wait -p DynamicUser=yes \
-p EnvironmentFile=-/usr/lib/systemd/systemd-asan-env \
-p WorkingDirectory='~' true)
assert_eq "$(systemd-run --pipe --uid=root -p WorkingDirectory='~' pwd)" "/root"
assert_eq "$(systemd-run --pipe --uid=nobody -p WorkingDirectory='~' pwd)" "/"
assert_eq "$(systemd-run --pipe --uid=testuser -p WorkingDirectory='~' pwd)" "/home/testuser"
(! systemd-run --wait -p DynamicUser=yes -p User=testuser \
-p EnvironmentFile=-/usr/lib/systemd/systemd-asan-env \
-p WorkingDirectory='~' true)

View File

@ -16,6 +16,7 @@ ConditionDirectoryNotEmpty=|/run/confexts
ConditionDirectoryNotEmpty=|/var/lib/confexts
ConditionDirectoryNotEmpty=|/usr/local/lib/confexts
ConditionDirectoryNotEmpty=|/usr/lib/confexts
ConditionDirectoryNotEmpty=|/.extra/confext
DefaultDependencies=no
After=local-fs.target