We had errno_to_name() which works for "known" errnos, and returns NULL for
unknown ones, and then ERRNO_NAME which always returns an answer, possibly just
a number as a string, but requires a helper buffer.
It is possible for the kernel to add a new errno. We recently learned that some
architectures define custom errno names. Or for some function to unexpectedly
return a bogus errno value. In almost all cases it's better to print that value
rather than "n/a" or "(null)". So let's use ERRNO_NAME is most error handling
code. Noteably, our code wasn't very good in handling the potential NULL, so
in various places we could print "(null)". Since this is supposed to be used
most of the time, let's shorten the names to ERRNO_NAME/errno_name.
There are a few places where we don't want to use the fallback path, in
particular for D-Bus error names or when saving the error name. Let's rename
errno_to_name() to errno_name_no_fallback() to make the distinction clearer.
With this change we fully duplicate the functionality in glibc. This patch
was motivated by glibc missing some names, but those are in the process of
being added:
https://sourceware.org/pipermail/libc-alpha/2025-July/169007.html
ci: add test coverage for secure boot and addons, and shim integration (#38377)
Now that we can enable SB on GHA do it by default in mkosi images with
selfenroll, and add a smoke test.
Also add one CI job that tests the shim integration. This will catch
regressions like https://github.com/systemd/systemd/issues/38349
Debian provides a signed shim that trusts sdboot and can be
installed without pulling in grub automatically. Install it
in the debian mkosi CI job, and build a custom efivars with
the mkosi cert enrolled in MOK but not DB, to test those
code paths.
* cb1a3c9049 FirmwareVariables: allow generating during image build
* 6104923534 env: export $EFI_ARCHITECTURE in hook scripts on EFI arches
* fef33f96a2 mkosi-tools: ukify moved to systemd-ukify in openSUSE
* ec4475a846 ensure builds with cache over device boundaries
* 7be5159f24 Change UnifiedKernelImages to enum and accept signed/unsigned
* 071ac4a575 mkosi-vm: install systemd-boot-efi-signed where available
* 1865be628e opensuse: Install OpenSUSE-release if another release package is not installed
* 0381b17819 qemu: Disable hpet for x86 VMs
* 4f63700eb3 mkosi-tools: install systemd-boot-tools for bootctl
* 1230ed333b man: remove duplicate 'the' in FirmwareVariables description
conf-files: CHASE_MUST_BE_DIRECTORY can be set with CHASE_NONEXISTENT
With the previous commit, now CHASE_MUST_BE_DIRECTORY can be set with
CHASE_NONEXISTENT. Let's unconditionally set the flag to chase the
directory part of the conf file.
chase: check the result is a directory or regular file only when the resolved path exists
Otherwise, if it is called with CHASE_NONEXISTENT, when we call
stat_verify_directory()/_regular() the struct stat is for one of the
parent directory, rather than for the result path.
With this change, we can safely specify CHASE_MUST_BE_DIRECTORY/REGULAR
with CHASE_NONEXISTENT.
More importantly, chaseat() internally sets CHASE_MUST_BE_DIRECTORY when
the input path ends with "/", "/,", "/..". Hence, without this change,
we cannot specify CHASE_NONEXISTENT safely.
Fixes the following error when running with sanitizers:
```
TEST-87-AUX-UTILS-VM.sh[670]: + bootctl install --make-entry-directory=yes
TEST-87-AUX-UTILS-VM.sh[695]: Copied "/usr/lib/systemd/boot/efi/systemd-bootx64.efi.signed" to "/boot/EFI/systemd/systemd-bootx64.efi".
TEST-87-AUX-UTILS-VM.sh[695]: Copied "/usr/lib/systemd/boot/efi/systemd-bootx64.efi.signed" to "/boot/EFI/BOOT/BOOTX64.EFI".
TEST-87-AUX-UTILS-VM.sh[695]: Created "/boot/fedora".
TEST-87-AUX-UTILS-VM.sh[695]: Random seed file /boot/loader/random-seed successfully refreshed (32 bytes).
TEST-87-AUX-UTILS-VM.sh[695]: ../src/shared/efi-api.c:618:38: runtime error: left shift of 243 by 24 places cannot be represented in type 'int'
```
TEST-23-UNIT-FILE: do not wait indefinitely but set a reasonable timeout
Otherwise, the test does not finish until the global timeout is reached.
This is for making the test fail earlier when something spurious happens:
```
[FAILED] Failed to start TEST-23-UNIT-FILE-short-lived.service - Shortlived Unit.
TEST-23-UNIT-FILE.sh[776]: + '[' 0 -eq 0 ']'
TEST-23-UNIT-FILE.sh[776]: + sleep .5
(snip)
58/98 systemd:integration-tests / TEST-23-UNIT-FILE TIMEOUT 1800.52s killed by signal 9 SIGKILL
```
- move scripts from test/units/ to the test specific units directory,
- drop meaningless true from silent-success.service,
- call journalctl from the same bash invocation of echo.
No functional change, just refactoring and preparation for the next
commit.
We want to check if the magic we got from statfs() is one of the magics listed
for one of the file systems in the given group. To do this, we'd iteratate over
the file system names, convert each name to an array of magics, and compare
those to the one we got. We were using gperf-generated lookup table for this,
so the string lookups were quick, but still this seems unnecessarily complex.
Let's just generate a simple lookup function, because we can:
meson: rework generation of file system lookup helpers
Previously, the gperf table was the main "source of truth", and additional
information (e.g. which file system names are obsolete) was scattered in
various files. We would then parse the gperf file, using python, awk, grep, and
bash, and use the results in various ways. This is hard to understand and
maintain. Let's replace all of this with a single python script that generates
the requested outputs as appropriate.
$ diff -u <(git show @{u}:src/basic/filesystems-gperf.gperf) \
<(src/basic/filesystem-sets.py gperf)
shows that the outputs are the same except for comments.
Similarly, 'src/basic/filesystem-sets.py fs-type-to-string' and
'src/basic/filesystem-sets.py filesystem-sets' can be used to view
the generated code.
The check that the kernel doesn't define any new file system magics is
converted into a normal test. It doesn't seem necessary to fail the build
when that happens.
I was looking at strace for systemd-getty-generator and noticed the call to
faccessat2(3</sys>, "", W_OK, AT_EMPTY_PATH), even though we already did
fstatfs(3</sys>), which should give us all the necessary information. Let's
only do this additional check when it's likely to yield something useful, i.e.
for network fses and otherwise skip the syscall.
The call to statvfs is replaced by statfs because that gives us the .f_type
field and allows is_network_fs() to be called.
I'm a bit worried that the is_network_fs() is somewhat costly. This will be
improved in later commits.
sd-bus: escape invalid characters in error message
At many places, we pass arguments of dbus method calls to
sd_bus_error_setf(), and produces unprintable log messages.
Let's always escape the generated error message.
NEWS: announce legacy iptables/libiptc support will be dropped in v259
nftables is available since kernel 3.13 (released on 19 January 2014).
Major distributions have already provided nftables, and marked/called
iptables as deprecated or legacy.
Moreover, currently, iptables/libiptc backend does not support IPv6.
Hence, it is not necessary to keep iptables/libiptc backend anymore.
Let's drop it in the next release.
Note, fedora/centos have already disabled iptables/libiptc support since v249.
TEST-64-UDEV-STORAGE: wait for partition devices being created before calling udevadm trigger
For some reasons, kernel or sfdisk once remove the created partitions
and recreated them. And if 'udevadm trigger' triggers devices currently
being removed, the udevd does not receive the triggered events, and the
command stuck.
```
[ 33.150452] TEST-64-UDEV-STORAGE.sh[546]: + sfdisk --wipe=always /dev/md/mdmirpar
[ 33.478336] systemd-udevd[442]: md127: Device is queued (SEQNUM=2163, ACTION=change)
[ 33.480153] kernel: md127: p1 p2 p3
[ 33.483772] systemd-udevd[442]: md127p1: Device is queued (SEQNUM=2164, ACTION=add)
[ 33.483914] systemd-udevd[442]: md127p2: Device is queued (SEQNUM=2165, ACTION=add)
[ 33.484999] systemd-udevd[442]: md127p3: Device is queued (SEQNUM=2166, ACTION=add)
[ 33.485564] systemd-udevd[442]: md127: Received inotify event of watch handle 164.
[ 33.503016] TEST-64-UDEV-STORAGE.sh[546]: + SYSTEMD_LOG_LEVEL=debug
[ 33.503016] TEST-64-UDEV-STORAGE.sh[546]: + timeout 30 udevadm trigger --settle --parent-match /dev/md/mdmirpar
[ 33.485905] systemd-udevd[442]: Successfully forked off '(udev-synth)' as PID 3208.
[ 33.486067] systemd-udevd[442]: md127: Removing watch handle 164.
[ 33.489035] systemd-udevd[442]: md127p1: Device is queued (SEQNUM=2167, ACTION=remove)
[ 33.489048] systemd-udevd[442]: Received inotify event about removal of watch handle 164.
[ 33.489507] systemd-udevd[442]: md127p2: Device is queued (SEQNUM=2168, ACTION=remove)
[ 33.496298] systemd-udevd[442]: md127p3: Device is queued (SEQNUM=2169, ACTION=remove)
[ 33.500628] systemd-udevd[442]: md127: Device is queued (SEQNUM=2170, ACTION=change)
[ 33.502355] systemd-udevd[442]: md127p1: Device is queued (SEQNUM=2171, ACTION=add)
[ 33.509371] TEST-64-UDEV-STORAGE.sh[3211]: md127: Triggered device with action 'change'.
[ 33.509371] TEST-64-UDEV-STORAGE.sh[3211]: md127p1: Triggered device with action 'change'.
[ 33.509371] TEST-64-UDEV-STORAGE.sh[3211]: md127p2: Triggered device with action 'change'.
[ 33.512532] systemd-udevd[442]: md127: Device is queued (SEQNUM=2172, ACTION=change, UUID=a0b75692-08ad-428a-859b-9ef8772874d7)
[ 33.512666] systemd-udevd[442]: md127p1: Device is queued (SEQNUM=2173, ACTION=change, UUID=4cd75a91-aa5b-4678-878c-0420b6c2e1e9)
[ 33.512796] systemd-udevd[442]: md127p2: Device is queued (SEQNUM=2174, ACTION=add)
[ 33.512910] systemd-udevd[442]: md127p3: Device is queued (SEQNUM=2175, ACTION=add)
[ 33.531834] TEST-64-UDEV-STORAGE.sh[3211]: md127: Got uevent without UUID, ignoring: No such file or directory
[ 33.553563] TEST-64-UDEV-STORAGE.sh[3211]: md127p1: Got uevent without UUID, ignoring: No such file or directory
[ 33.561262] TEST-64-UDEV-STORAGE.sh[3211]: md127p2: Got uevent without UUID, ignoring: No such file or directory
[ 33.562468] TEST-64-UDEV-STORAGE.sh[3211]: md127p2: Got uevent without UUID, ignoring: No such file or directory
[ 33.563143] TEST-64-UDEV-STORAGE.sh[3211]: md127p3: Got uevent without UUID, ignoring: No such file or directory
[ 33.564174] TEST-64-UDEV-STORAGE.sh[3211]: md127p1: Got uevent without UUID, ignoring: No such file or directory
[ 33.567614] TEST-64-UDEV-STORAGE.sh[3211]: md127p3: Got uevent without UUID, ignoring: No such file or directory
[ 33.597750] TEST-64-UDEV-STORAGE.sh[3211]: md127: Got uevent without UUID, ignoring: No such file or directory
[ 33.623522] TEST-64-UDEV-STORAGE.sh[3211]: md127p1: Got uevent without UUID, ignoring: No such file or directory
[ 33.676268] TEST-64-UDEV-STORAGE.sh[3211]: md127p3: Got uevent without UUID, ignoring: No such file or directory
[ 33.686088] TEST-64-UDEV-STORAGE.sh[3211]: md127p2: Got uevent without UUID, ignoring: No such file or directory
```
Let's wait for partition devices being actually created, and wait for
all queued events being processed. Then, call 'udevadm trigger'.
Previously, ordered_set_put_strdupv() and friends returns the number of
pushed entries, but that is potentially larger than INT_MAX (of course,
realistically, OOM is triggered in that case).
No caller uses the number of the new entries. Let's return 1 when at
least one element is added.
bootctl: automatically set --graceful when running in chroot
Installing stuff in a chroot should not fail because efivars are
not available. When running in a container touching efivars is
completely disabled, but there are some cases (recovery) where
it is needed to touch them in a chroot, so don't disable them but
avoid failing the run instead.
meson: use files() not find_program() for helper scripts
We went back and forth between 'prog.sh', files('prog.sh'), and
find_program('prog.sh'). We want to use files() or find_program() so that we
get a good error message if the file is missing. Behaviour of meson changed
over time, and in the past not all forms could be used in all places. For
example 0f4c4f3824891aea6a76b2e861f38a6045b2a246 added find_program() in many
places to avoid repeated messages. But it seems that all recent meson versions
work fine with files().
find_program prints silly messages:
Program tools/make-man-index.py found: YES
(/home/zbyszek/src/systemd/tools/make-man-index.py)
Program tools/meson-render-jinja2.py found: YES
(/home/zbyszek/src/systemd/tools/meson-render-jinja2.py)
...
We know that those files will be found, they are part of the git checkout.
With files() this is gone and the meson output is easier to read.
All those files are standalone programs that can be executed directly.
Some .py files were marked executable, others weren't, probably accidentally.
Mark them all as executable in preparation for subsequent changes.
Some files were using that, others weren't. Since those are user-facing
docs, we should use the nice clean form, not the workaround for strange
systems that we use in other places.
Before 7d247d3cb875e1778e6cc81dbf255b13092b29e0, we needed the file name
twice. But now we only need it once, so no need to overcomplicate things
by defining a variable with a single use.
When a variable is used, it's also easier to make a mistake and e.g.
accidentally reuse the variable later.
- use DEFINE_TEST_MAIN() macro,
- use ASSERT_OK() and friends,
- drop restriction that the test must run as root,
- replace sleep(2) with loop of shorter sleep and cg_is_empty(),
- applied several coding style cleanups.
TEST-10-MOUNT: wait for userspace mount options being loaded
When a device is mounted with userspace options such as _netdev, even when the mount event source is
triggered, only /proc/self/mountinfo may be updated, and /run/mount/utab may not be updated yet.
Hence, the mount unit may be created/updated without the userspace options. In that case, the mount
event source will be retriggered when /run/mount/utab is updated, and the mount unit will be updated
again with the userspace options. Typically, the window between the two calls is very short, but when
the mount event source is ratelimited after the first event, processing the second event may be delayed
about 1 secound. Hence, here we need to wait for a while.
By adding a debugging logs in mount_setup_unit(), the userspace mount is
not obtained in the first event, and the second event is delayed by the ratelimit.
```
[ 20.023086] H TEST-10-MOUNT.sh[446]: + mount -t ext4 -o _netdev /dev/loop1p1 /tmp/deptest
[ 20.026255] H kernel: EXT4-fs (loop1p1): mounted filesystem c1fa00ea-2ba8-46b2-9002-2ac997f4cda9 r/w with ordered data mode. Quota mode: none.
[ 20.026537] H TEST-10-MOUNT.sh[446]: + timeout 10 bash -c 'until systemctl -q is-active tmp-deptest.mount; do sleep .1; done'
[ 20.032293] H systemd[1]: tmp-deptest.mount: mount_setup_unit: proc: yes, netdev: no
[ 20.035978] H systemd[1]: Unit blockdev@dev-loop1p1.target has alias blockdev@.target.
[ 20.039765] H systemd[1]: tmp-deptest.mount: Changed dead -> mounted
[ 20.046598] H systemd[1]: Event source 0x7c73093e05e0 (mount-monitor-dispatch) entered rate limit state.
```
test: move testcase_dependencies() to TEST-10-MOUNT
TEST-60-MOUNT_RATELIMIT is run on nspawn by default, and currently run
on vm only on arch mkosi. Let's move the test case to new TEST-10-MOUNT,
which always run on vm.
projects / thirdparty / systemd.git / log
