test-bus-introspect is also applied to the tables from test-bus-vtable.c.
test-bus-vtable.c is also used as C++ sources to produce test-bus-vtable-cc,
and our hashmap headers are not C++ compatible. So let's do the introspection
part only in the C version.
In cgroup v2 we have protection tunables -- currently MemoryLow and
MemoryMin (there will be more in future for other resources, too). The
design of these protection tunables requires not only intermediate
cgroups to propagate protections, but also the units at the leaf of that
resource's operation to accept it (by setting MemoryLow or MemoryMin).
This makes sense from an low-level API design perspective, but it's a
good idea to also have a higher-level abstraction that can, by default,
propagate these resources to children recursively. In this patch, this
happens by having descendants set memory.low to N if their ancestor has
DefaultMemoryLow=N -- assuming they don't set a separate MemoryLow
value.
Any affected unit can opt out of this propagation by manually setting
`MemoryLow` to some value in its unit configuration. A unit can also
stop further propagation by setting `DefaultMemoryLow=` with no
argument. This removes further propagation in the subtree, but has no
effect on the unit itself (for that, use `MemoryLow=0`).
Our use case in production is simplifying the configuration of machines
which heavily rely on memory protection tunables, but currently require
tweaking a huge number of unit files to make that a reality. This
directive makes that significantly less fragile, and decreases the risk
of misconfiguration.
After this patch is merged, I will implement DefaultMemoryMin= using the
same principles.
This avoids double compilation. Those files are tiny, so it doesn't save time,
but we avoid repeated warnings and errors, and it's generally cleaner to it
this way.
The number of commands in 'ninja -C build clean && ninja -C build' drops from
1462 to 1455 for me.
Let's move the shutdown binary into its own subdirectory in
src/shutdown, after all it is relatively isolated from the normal PID 1
sources, being a different binary and all.
Unfortunately it's not possible to move some of the code, since it is
shared with PID 1, that I wished we could move, but I still think it's
worth it.
This new setting allows configuration of CFS period on the CPU cgroup, instead
of using a hardcoded default of 100ms.
Tested:
- Legacy cgroup + Unified cgroup
- systemctl set-property
- systemctl show
- Confirmed that the cgroup settings (such as cpu.cfs_period_ns) were set
appropriately, including updating the CPU quota (cpu.cfs_quota_ns) when
CPUQuotaPeriodSec= is updated.
- Checked that clamping works properly when either period or (quota * period)
are below the resolution of 1ms, or if period is above the max of 1s.
The test queries some domain names. If the DNS servers are unreachable,
e.g. in a rawhide container I get the total runtime of 24.5s usually, but
sometimes slightly longer, enough to reach the default timeout of 30s.
Apparently, people do use nscd, hence play somewhat nice with it, and
let's explicitly flush nscd caches whenever we register a new
user/group.
This patch only adds the actual refresh request invocation. Later
commits then issue this call at appropriate moments.
Note that the nscd protocol is not officially documented though very
simple. This code is written very defensively so that incompatibilities
don't affect us much.
Given that glibc really has a duty to maintain compat between
differently compiled programs and their system nscd they can't break API
and thus it should be safe for us to implement an alternative,
minimalistic client.
Ideally this kind of explicit, global cache flushing would not be necessary.
However nscd currently has no cache coherency protocol, hence we can't
really implement this better. The only concept it knows is a TTL for
positive hosts lookups. Hoewver for negative lookups or any of the other
tables nothing is available.
Let's generalize this, so that we can use this in nspawn later on, which
is pretty useful as we need to be able to mask files from the inner
child of nspawn too, where the host's /run/systemd/inaccessible
directory is not visible anymore. Moreover, if nspawn can create these
nodes on its own before the payload this means the payload can run with
fewer privileges.
We would accept e.g. FailureAction=reboot-force in user units and then do an
exit in the user manager. Let's be stricter, and define "exit"/"exit-force" as
the only supported actions in user units.
v2:
- rename 'exit' to 'exit-force' and add new 'exit'
- add test for the parsing function
This makes it so that tests no longer need to know the absolute paths to the
source and build dirs, instead using the systemd-runtest.env file to get these
paths when running from the build tree.
Confirmed that test-catalog works on `ninja test`, when called standalone and
also when the environment file is not present, in which case it will use the
installed location under /usr/lib/systemd/catalog.
The location can now also be overridden for this test by setting the
$SYSTEMD_CATALOG_DIR environment variable.
This adds -Dnss-resolve= and -Dnss-mymachines= meson options.
By using this option, e.g., resolved can be built without nss-resolve.
When no nss modules are built, then test-nss is neither built.
Also, This changes the option name -Dmyhostname= to -Dnss-myhostname=
for consistency to other nss related options.
Closes#9596.
This shows a minor memleak:
==1883== 24 bytes in 1 blocks are definitely lost in loss record 1 of 1
==1883== at 0x4C2DBAB: malloc (vg_replace_malloc.c:299)
==1883== by 0x4E9D385: malloc_multiply (alloc-util.h:69)
==1883== by 0x4EA2959: bus_request_name_async_may_reload_dbus (bus-util.c:1841)
==1883== by ...
The exchange of messages is truncated at two different points: once right
after the first callback is requested, and the second time after the full
sequence has run (usually resulting in an error because of policy).
We have plenty of code in our codebase that outputs tables to the
console, and all is homegrown and awful. Let's replace it with a generic
implementation that can do automatically what the old implementations
did manually.
Features:
1. Ellipsation (for fields overly long) and alignment (for
fields overly short)
2. Sorting of rows
3. automatically copies formatting from the same cell in the row above
4. Heavy use of varargs to make putting together tables easy
5. can expand and compress tables, with weights
6. Has a minimal understanding of unicode wide characters in order to
match unicode strings to character cell terminals.
7. Columns can be reordered and individually turned off.
8. pretty printing for various data types
And more.
We go through the whole file system, so this test can take arbitrary time. But
this test is still quite useful, so let's at least try to make it more efficent
by not descending at all into the directories we would filter out later on
anyway.
Also increase the timeout, in case the previous step doesn't help enough.
Files which are installed as-is (any .service and other unit files, .conf
files, .policy files, etc), are left as is. My assumption is that SPDX
identifiers are not yet that well known, so it's better to retain the
extended header to avoid any doubt.
I also kept any copyright lines. We can probably remove them, but it'd nice to
obtain explicit acks from all involved authors before doing that.
Case sensitive or case insensitive matching can be requested using
--case-sensitive[=yes|no].
Unless specified, matching is case sensitive if the pattern contains any
uppercase letters, and case insensitive otherwise. This matches what
forward-search does in emacs, and recently also --ignore-case in less. This
works surprisingly well, because usually when one is wants to do case-sensitive
matching, the pattern is usually camel-cased. In the less frequent case when
case-sensitive matching is required with an all-lowercase pattern,
--case-sensitive can be used to override the automatic logic.
Previously, we'd maintain two hashmaps keyed by PIDs, pointing to Unit
interested in SIGCHLD events for them. This scheme allowed a specific
PID to be watched by exactly 0, 1 or 2 units.
With this rework this is replaced by a single hashmap which is primarily
keyed by the PID and points to a Unit interested in it. However, it
optionally also keyed by the negated PID, in which case it points to a
NULL terminated array of additional Unit objects also interested. This
scheme means arbitrary numbers of Units may now watch the same PID.
Runtime and memory behaviour should not be impact by this change, as for
the common case (i.e. each PID only watched by a single unit) behaviour
stays the same, but for the uncommon case (a PID watched by more than
one unit) we only pay with a single additional memory allocation for the
array.
Why this all? Primarily, because allowing exactly two units to watch a
specific PID is not sufficient for some niche cases, as processes can
belong to more than one unit these days:
1. sd_notify() with MAINPID= can be used to attach a process from a
different cgroup to multiple units.
2. Similar, the PIDFile= setting in unit files can be used for similar
setups,
3. By creating a scope unit a main process of a service may join a
different unit, too.
4. On cgroupsv1 we frequently end up watching all processes remaining in
a scope, and if a process opens lots of scopes one after the other it
might thus end up being watch by many of them.
This patch hence removes the 2-unit-per-PID limit. It also makes a
couple of other changes, some of them quite relevant:
- manager_get_unit_by_pid() (and the bus call wrapping it) when there's
ambiguity will prefer returning the Unit the process belongs to based on
cgroup membership, and only check the watch-pids hashmap if that
fails. This change in logic is probably more in line with what people
expect and makes things more stable as each process can belong to
exactly one cgroup only.
- Every SIGCHLD event is now dispatched to all units interested in its
PID. Previously, there was some magic conditionalization: the SIGCHLD
would only be dispatched to the unit if it was only interested in a
single PID only, or the PID belonged to the control or main PID or we
didn't dispatch a signle SIGCHLD to the unit in the current event loop
iteration yet. These rules were quite arbitrary and also redundant as
the the per-unit handlers would filter the PIDs anyway a second time.
With this change we'll hence relax the rules: all we do now is
dispatch every SIGCHLD event exactly once to each unit interested in
it, and it's up to the unit to then use or ignore this. We use a
generation counter in the unit to ensure that we only invoke the unit
handler once for each event, protecting us from confusion if a unit is
both associated with a specific PID through cgroup membership and
through the "watch_pids" logic. It also protects us from being
confused if the "watch_pids" hashmap is altered while we are
dispatching to it (which is a very likely case).
- sd_notify() message dispatching has been reworked to be very similar
to SIGCHLD handling now. A generation counter is used for dispatching
as well.
This also adds a new test that validates that "watch_pid" registration
and unregstration works correctly.
This adds a "watch-bind" feature to sd-bus connections. If set and the
AF_UNIX socket we are connecting to doesn't exist yet, we'll establish
an inotify watch instead, and wait for the socket to appear. In other
words, a missing AF_UNIX just makes connecting slower.
This is useful for daemons such as networkd or resolved that shall be
able to run during early-boot, before dbus-daemon is up, and want to
connect to dbus-daemon as soon as it becomes ready.
This reduces the meson man=false target count to 1281.
v2:
- link test-engine with libshared instead of libsystemd_static
Previous version built fine on F27, but fails on F26 with the following error:
/usr/bin/ld: /tmp/ccr8HRGw.ltrans6.ltrans.o: undefined reference to symbol '__start_BUS_ERROR_MAP@@SD_SHARED'
/home/zbyszek/fedora/systemd/systemd-9d5aae75c64f5583a110f03b94816aacc03bbf4d/x86_64-redhat-linux-gnu/src/shared/libsystemd-shared-236.so: error adding symbols: DSO missing from command line
v3:
- add libudev_basic
gcrypt_util_sources had to be moved because otherwise they appeared twice
in libshared.so halfproducts, causing an error.
-fvisibility=default is added to libbasic, libshared_static so that the symbols
appear properly in the exported symbol list in libshared.
The advantage is that files are not compiled twice. When configured with -Dman=false,
the ninja target list is reduced from 1588 to 1347 targets. The difference in compilation
time is small (<10%). I think this is because of -O0 and ccache and multiple cores, and
in different settings the compilation time could be reduced. The main advantage is that
errors and warnings are not reported twice.
We already use the "_static" suffix for libshared_static ("shared" is the name
of the library, "static" is the format) and other libs, so let's rename for
consistency.
Also change libsystemd_static_sources to libsystemd_sources, since the same
list is used for both and shorter is better.
In this way, individual errors in files can be treated differently than a
failure of the whole service.
A test is added to check that the expected value is returned.
Some parts are commented out, because it is not. This will be fixed in
a subsequent commit.
So far I avoided adding license headers to meson files, but they are pretty
big and important and should carry license headers like everything else.
I added my own copyright, even though other people modified those files too.
But this is mostly symbolic, so I hope that's OK.
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 kdbus_flag and memfd related parts are left behind, because they
are entangled with the "legacy" dbus support.
test-bus-benchmark is switched to "manual". It was already broken before
(in the non-kdbus mode) but apparently nobody noticed. Hopefully it can
be fixed later.
test-login.c is largely rewritten to use _cleanup_ and give more meaningful
messages (function names are used instead of creative terms like "active
session", so that when something unexpected is returned, it's much easier to
see what function is responsible).
The monitoring part is only activated if '-m' is passed on the command line.
It runs against the information from /run/systemd/ in the live system, but that
should be OK: logind/sd-login interface is supposed to be stable and both
backwards and forwards compatible.
If not running in a login session, some tests are skipped.
Those two changes together mean that it's possible to run test-login in the
test suite.
Tests for sd_pid_get_{unit,user_unit,slice} are added.
This test is mostly a compilation test that checks that various defines in
sd-bus-vtable.h are valid C++. The code is executed, but the results are not
checked (apart from sd-bus functions not returning an error). test-bus-objects
contains pretty extensive tests for this functionality.
The C++ version is only added to meson, since it's simpler there.
Because of the .cc extension, meson will compile the executable with c++.
This test is necessary to properly check the macros in sd-bus-vtable.h. Just
running the headers through g++ is not enough, because the macros are not
exercised.
Follow-up for #5941.
This adds a modified version of dhcp6_option_parse_domainname() that is
able to parse compressed domain names, borrowing the idea from
dns_packet_read_name(). It also adds pieces in networkd-link and
networkd-manager to properly save/load the added option field.
Resolves#2710.
The indentation for emacs'es meson-mode is added .dir-locals.
All files are reindented automatically, using the lasest meson-mode from git.
Indentation should now be fairly consistent.
This makes the helper binaries significantly bigger (in some cases, the final
size depends on link options and optimization level), and is only useful for
distributions which want to provide the option to install udev without systemd.
As the linking is improved, the difference between the columns might shrink,
but it's unlikely that linking libshared statically could ever be more
efficient.
E.g. with -O0, no -flto:
(static) (shared)
src/udev/ata_id 999176 85696
src/udev/cdrom_id 1024344 111656
src/udev/collect 990344 81280
src/udev/scsi_id 1023592 115656
src/udev/v4l_id 811736 17744
When linked dynamically, install_rpath must be specified, so add that.
test-dlopen is a very simple binary that is only linked with libc and
libdl. From it we do dlopen() on the nss and pam modules to check that they are
linked to all necessary libs.
(meson-compiled nss modules are linked to less libraries, for whatever reason.
I suspected that some deps are missing, but it turns out that my suspicions
weren't justified, and the modules load just fine. Let's keep the test though,
it is very quick, and might detect missing linkage in the future.)
This simplifies things and leads to a smaller installation footprint.
libsystemd_internal and libsystemd_journal_internal are linked into
libystemd-shared and available to all programs linked to libsystemd-shared.
libsystemd_journal_internal is not needed anymore, and libsystemd-shared
is used everwhere. The few exceptions are: libsystemd.so, test-engine,
test-bus-error, and various loadable modules.
The tests are included under the conditional too, instead of specifying
'ENABLE_NETWORKD' in the test definition array, because libnetworkd_core
dependency is undefined if networkd is disabled.
With mesonbuid/meson#1545, meson does not propagate deps of a library
when linking with that library. That's of course the right thing to do,
but it exposes a bunch of missing deps.
This compiles with both meson-0.39.1 and meson-git + pr/1545.
This is slightly complicated by the fact that files('libudev.h') cannot be used
as an argument in custom_target command (string is required). This restriction
should be lifted in future versions of mesons, so this could be simplified.
This is quite messy. I think libtool might have been using something
like -Wl,--whole-archive, but I don't think meson has support for that.
For now, just recompile all the sources for linking into libsystemd
directly. This should not matter much for efficiency, since it's a
few small files.
This is what autoconf-based build does, and it makes test-bus-error and
test-engine able to access the bus error mapping table. OTOH, this is a heavy
price to pay: it would be excellent to link libcore.a to libsystemd-shared-NNN.so.
Otherwise we duplicate the same code in 'systemd' and 'libsystemd-shared-NNN.so'.
-rwxrwxr-x. 1 4075544 Apr 6 20:30 systemd* <-- libcore linked against libsystemd-shared.so
-rwxrwxr-x. 1 5596504 Apr 9 14:07 systemd* <-- libcore linked against libsystemd-shared.a
v2:
- update for 6b5cf3ea62
Tests can be run with 'ninja-build test' or using 'mesontest'.
'-Dtests=unsafe' can be used to include the "unsafe" tests in the
test suite, same as with autotools.
v2:
- use more conf.get guards are optional components
- declare deps on generated headers for test-{af,arphrd,cap}-list
v3:
- define environment for tests
Most test don't need this, but to be consistent with autotools-based build, and
to avoid questions which tests need it and which don't, set the same environment
for all tests.
v4:
- rework test generation
Use a list of lists to define each test. This way we can reduce the
boilerplate somewhat, although the test listings are still pretty verbose. We
can also move the definitions of the tests to the subdirs. Unfortunately some
subdirs are included earlier than some of the libraries that test binaries
are linked to. So just dump all definitions of all tests that cannot be
defined earlier into src/test. The `executable` definitions are still at the
top level, so the binaries are compiled into the build root.
v5:
- tag test-dnssec-complex as manual
v6:
- fix HAVE_LIBZ typo
- add missing libgobject/libgio defs
- mark test-qcow2 as manual
