Apparently, PAM documents that the PAM_TTY should come with a /dev
prefix, but we don't expect it so far, except that Wayland ends up
setting it after all, the way the docs suggest. Hence, let's simply drop
the /dev prefix if it is there.
Fixes: #6516
This new helper removes a leading /dev if there is one. We have code
doing this all over the place, let's unify this, and correct it while
we are at it, by using path_startswith() rather than startswith() to
drop the prefix.
Currently if tmpfiles is run with force on symlink creation but there already
exists a directory at that location, the creation will fail. This change
updates the behavior to remove the directory with rm_fr and then attempts to
create the symlink again.
* Containers don't use serial-getty@console.service,
they use console-getty.service instead, and suppress
scanning for kernel or virtualizer consoles.
* Nowadays gettys are started on *all* configured kernel consoles.
* except for the line printer console, because that's not a tty.
(Seriously. Search CONFIG_LP_CONSOLE).
Without this "meson test" will end up running all tests in the same
cgroup root, and they all will try to manage it. Which usually isn't too
bad, except when they end up clearing up each other's cgroups. This race
is hard to trigger but has caused various CI runs to fail spuriously.
With this change we simply move every test that runs a manager object
into their own private cgroup. Note that we don't clean up the cgroup at
the end, we leave that to the cgroup manager around it.
This fixes races that become visible by test runs throwing out errors
like this:
```
exec-systemcallfilter-failing.service: Passing 0 fds to service
exec-systemcallfilter-failing.service: About to execute: /bin/echo 'This should not be seen'
exec-systemcallfilter-failing.service: Forked /bin/echo as 5693
exec-systemcallfilter-failing.service: Changed dead -> start
exec-systemcallfilter-failing.service: Failed to attach to cgroup /exec-systemcallfilter-failing.service: No such file or directory
Received SIGCHLD from PID 5693 ((echo)).
Child 5693 ((echo)) died (code=exited, status=219/CGROUP)
exec-systemcallfilter-failing.service: Child 5693 belongs to exec-systemcallfilter-failing.service
exec-systemcallfilter-failing.service: Main process exited, code=exited, status=219/CGROUP
exec-systemcallfilter-failing.service: Changed start -> failed
exec-systemcallfilter-failing.service: Unit entered failed state.
exec-systemcallfilter-failing.service: Failed with result 'exit-code'.
exec-systemcallfilter-failing.service: cgroup is empty
Assertion 'service->main_exec_status.status == status_expected' failed at ../src/src/test/test-execute.c:71, function check(). Aborting.
```
BTW, I tracked this race down by using perf:
```
# perf record -e cgroup:cgroup_mkdir,cgroup_rmdir
…
# perf script
```
Thanks a lot @iaguis, @alban for helping me how to use perf for this.
Fixes#5895.
some run_target() calls were using params from custom_target()
example message:
WARNING: Passed invalid keyword argument "input". This will become a hard error in the future.
New way to call targets:
ninja man/man
ninja man/html
ninja man/update-man-rules
An infiniband hardware address is 20 bytes, but sockaddr_ll.sll_addr is only 8
bytes. Explicitly ensure that sockaddr_union has enough space for infiniband
addresses, even if they run over sockaddr_ll and add a macro to compute the
proper size to pass to kernel.
Observed when running from the console of a systemd nspawn container
(see failure below).
The value of r was tested, when r was last set by
sd_session_can_graphical(). This did not correspond to the value expected.
Fix the code, so we compare relevant values now. Hopefully :).
Test failure
------------
/* Information printed is from the live system */
sd_pid_get_unit(0, …) → "session-13.scope"
sd_pid_get_user_unit(0, …) → "n/a"
sd_pid_get_slice(0, …) → "user-1000.slice"
sd_pid_get_session(0, …) → "13"
sd_pid_get_owner_uid(0, …) → 1000
sd_pid_get_cgroup(0, …) → "/user.slice/user-1000.slice/session-13.scope"
sd_uid_get_display(1000, …) → "13"
sd_uid_get_sessions(1000, …) → [2] "15 13"
sd_uid_get_seats(1000, …) → [1] "seat0"
sd_session_is_active("13") → yes
sd_session_is_remote("13") → no
sd_session_get_state("13") → "active"
sd_session_get_uid("13") → 1000
sd_session_get_type("13") → "tty"
sd_session_get_class("13") → "user"
sd_session_get_display("13") → "n/a"
sd_session_get_remote_user("13") → "n/a"
sd_session_get_remote_host("13") → "n/a"
sd_session_get_seat("13") → "seat0"
sd_session_can_multi_seat("seat0") → no
sd_session_can_tty("seat0") → no
sd_session_can_graphical("seat0") → no
sd_uid_get_state(1000, …) → active
Assertion '!!k == !!r' failed at ../src/libsystemd/sd-login/test-login.c:191, function test_login(). Aborting.
Fixes:
```
$ env -i valgrind --leak-check=full ./build/test-bus-chat
...
==7763== 1,888 (1,824 direct, 64 indirect) bytes in 1 blocks are
definitely lost in loss record 2 of 2
==7763== at 0x4C2FA50: calloc (vg_replace_malloc.c:711)
==7763== by 0x4F8FF9A: sd_bus_new (sd-bus.c:175)
==7763== by 0x4F938BF: sd_bus_open_user (sd-bus.c:1138)
==7763== by 0x109ACD: server_init (test-bus-chat.c:70)
==7763== by 0x10BCF8: main (test-bus-chat.c:526)
==7763==
```
Closes#6481
The directories are only used by the specific services, and
created before the services are started. So, it is not necessary
to create them by systemd-tmpfiles.
This commit moves the first-boot system preset-settings evaluation out
of main and into the manager startup logic itself. Notably, it reverses
the order between generators and presets evaluation, so that any changes
performed by first-boot generators are taken into the account by presets
logic.
After this change, units created by a generator can be enabled as part
of a preset.
I hit a test failure with the `max_gid+1` test. Problem is that we loop
over 0..r, but set `r` again within the loop (to 1). So max_gid is only
set based on the first supplementary GID.
ConditionGroup=1000 → 1
ConditionGroup=4 → 1
ConditionGroup=adm → 1
ConditionGroup=1001 → 1
Assertion 'r == 0' failed at ../src/test/test-condition.c:462, function
test_condition_test_group(). Aborting.
$ id
uid=1000(alan-sysop) gid=1000(alan-sysop) groups=1000(alan-sysop),4(adm),
10(wheel),1001(sshlogin)
Without this requirement, if proxy-to-nginx.socket was down, and the sysadmin
were to do:
systemctl start proxy-to-nginx.service
then the service would come up without a configured socket, which doesn't make
sense. Normally this isn't how we expect a socket-activated service to start,
but it's possible for an admin to do this (if the .socket were already running,
the systemd-socket-proxyd process will start effectively idle). But the
.service shouldn't end up in a broken state if the .socket isn't already
listening.
Adding the explicit Requires: should ensure that an admin with this
configuration state can't accidentally break their system.
This reverts commit 4f5e972279.
Building with gperf 3.0 works just fine; we had an autoconf check to
determine the correct data types, and this check was ported to meson.
All test/TEST* but TEST-02-CRYPTSETUP share the same check_result_qemu()
and test_cleanup(), so move them into test_functions and only override
them in TEST-02-CRYPTSETUP.
Also provide a common test_run() which by default assumes that both QEMU
and nspawn tests are run. Particular tests which don't support either
need to explicitly opt out by setting $TEST_NO_{QEMU,NSPAWN}. Do it this
way around to avoid accidentally forgetting to opt in, and to encourage
test authors to at least always support nspawn.
