The comment https://github.com/systemd/systemd/pull/21814#discussion_r771842132
suggests to introduce new helper, but it is used only one place.
Let's not add such, but simply replace the goto with a flag.
Nishal Kulkarni [Fri, 17 Dec 2021 16:36:31 +0000 (22:06 +0530)]
logind: Use new macros
Migrate logind to use the new macros to declare a D-Bus method or signal.
Replaced SD_BUS_METHOD_WITH_NAMES with SD_BUS_METHOD_WITH_ARGS.
Replaced SD_BUS_SIGNAL_WITH_NAMES with SD_BUS_SIGNAL_WITH_ARGS.
Daan De Meyer [Fri, 17 Dec 2021 19:01:31 +0000 (20:01 +0100)]
core: Add trigger limit for path units
When conditions fail on a service unit, a path unit can cause
PID 1 to busy loop as it keeps trying to activate the service unit.
To avoid this from happening, add a trigger limit to the path unit,
identical to the trigger limit we have for socket units.
Initially, let's start with a high limit and not make it configurable.
If needed, we can add properties to configure the rate limit similar
to the ones we have for socket units.
Frantisek Sumsal [Fri, 17 Dec 2021 19:59:10 +0000 (20:59 +0100)]
man: fix machine-id(5) man page reference
Spotted whilst debugging:
```
[763/2094] Generating man/machine-info.html with a custom command
Element cite in namespace '' encountered in para, but no template matches.
[765/2094] Generating man/machine-info.5 with a custom command
Element cite in namespace '' encountered in para, but no template matches.
```
Now that kernel-install creates the machine-id directory, we don't need to do
this is 'bootctl install', and in fact it's better not to do this since it
might never be necessary. So let's change the default behaviour to 'no'.
I kept support for 'auto' to maintain backwards compatibility, even though the
default was changed. Previous behaviour can be requested by specifying
--make-machine-id-directory=auto.
bootctl: write KERNEL_INSTALL_LAYOUT=bls and KERNEL_INSTALL_MACHINE_ID=…
This is a natural extension of d6bce6e224: if we are installing sd-boot, we
want to use the sd-boot layout, so let's write the appropriate
KERNEL_INSTALL_LAYOUT setting. Effectively, if we do 'booctl install',
kernel-install will not autodetect the layout anymore.
And 357376d0bb added support for KERNEL_INSTALL_MACHINE_ID. We need to support
it here too. We both read it, so that we create the right directories, and also
write it if it wasn't written yet and we created some directories using it, so
that kernel-install that is executed later knows the machine-id that matches
the directories we crated.
The code is changed in some places to fail if we can't figure out the current
status. When installing the boot loader it's probably better not to guess.
Anita Zhang [Fri, 17 Dec 2021 12:19:53 +0000 (04:19 -0800)]
test: adjust MemoryHigh= on oomd extended test units
On some runs `sleep infinity` run by the user manager uses over 3M of
memory, which is higher than the MemoryHigh= set on testbloat and
testmunch. If no pgscan is generated, then systemd-oomd sorts by memory
usage which leads to a situation where testchill (using 3M) could be
targeted over testbloat (1M-2M).
Fix this by setting reasonable MemoryHigh= values for all of these test
units. Even if somehow testchill throttles a bit at 3M, testbloat and
testmunch should still be trying to use over 100M at memory and will
throttle down to 5M and 6M with the new values. This should reflect
the desired state in pgscan and memory usage during the test run.
наб [Tue, 16 Mar 2021 15:47:34 +0000 (16:47 +0100)]
kernel-install: replace 00-entry-directory with K_I_LAYOUT in k-i
341890de866f2ee34919a47ce3fc6c8cd3c1924c made "bootctl install" create
ESP\MID, in preparation of cf73f650890b56a59bfb713c4c82b4e29daa7316 that
followed it and created 00-entry-directory.install to make ESP\MID\KVER
if ESP\MID existed ‒ this meant that "bootctl install" followed by
"kernel-install $(uname -r) /boot/vml*$(uname -r) /boot/ini*$(uname -r)"
actually installed the kernel correctly.
Later, 31e57550b552e113bd3d44355b237c41e42beb58 reverted the first
commit, meaning, that now running those two commands first installs
sd-boot, but then does nothing. Everything appears to work right,
nothing errors out, but no changes are actually done. To the untrained
eye (all of them), even running with -v appears to work:
all the hooks are run, as is depmod, but, again, nothing happens.
This is horrible. Nothing in either manpage suggests what to do
(nor should it, really), but the user is left with a bootloader that
appears fully funxional, since nothing suggests a failure in the output,
but with an unbootable machine, /no way to boot it/, even if they drop
to an EFI shell, since the boot bundle isn't present on the ESP,
and no real recourse even if they boot into a recovery system,
apart from installing like GRUB or whatever.
00- is purely instrumentation for 90-,
and separating one from the other has led to downstream dissatisfaxion
(indeed, the last mentioned commit cited cited exactly that as the
reversion reason), while creating $ENTRY_DIR_ABS is only required
for bootloaders using the BLS, and shouldn't itself toggle anything.
To that end, introduce an /{e,l}/k/install.conf file that allows
overriding the detected layout, and detect it as "bls" if
$BOOT_ROOT/$MACHINE_ID ($ENTRY_DIR_ABS/..) exists, otherwise "other" ‒
if a user wishes to select a different bootloader,
like GRUB, they (or, indeed, the postinst script) can specify
layout=grub. This disables 90- and $ENTRY_DIR_ABS manipulation.
The code was correct, but looked suspicious: we were comparing
strlen(x) with sizeof(y), with looks like an off-by-one. But we actually
want x to be one longer than y, so that's fine. Let's use STRLEN() to
make this more obvious.
Michael Biebl [Thu, 16 Dec 2021 18:18:28 +0000 (19:18 +0100)]
test: record missing openssl as a failure
The openssl binary is an optional dependency.
If systemd has been built with OpenSSL support, we want to test its
OpenSSL functionality.
So record a failure message in /failed if the binary is missing.
See https://github.com/systemd/systemd/pull/21724#issuecomment-992707614
Daan De Meyer [Tue, 14 Dec 2021 16:09:08 +0000 (17:09 +0100)]
kernel-install: Introduce KERNEL_INSTALL_MACHINE_ID in /etc/machine-info
If KERNEL_INSTALL_MACHINE_ID is defined in /etc/machine-info, prefer it
over the machine ID from /etc/machine-id. If a machine ID is defined in
neither /etc/machine-info nor in /etc/machine-id, generate a new UUID
and try to write it to /etc/machine-info as KERNEL_INSTALL_MACHINE_ID
and use it as the machine ID if writing it to /etc/machine-info succeeds.
In practice, this means we have a more robust fallback if there's no
machine ID in /etc/machine-id than just using "Default" and allows
image builders to force kernel-install to use KERNEL_INSTALL_MACHINE_ID
by simply writing it to /etc/machine-info themselves.
Frantisek Sumsal [Thu, 16 Dec 2021 10:59:09 +0000 (11:59 +0100)]
test: settle before checking logs
Otherwise we might miss the "Device path too long" message:
```
[ 21.083274] testsuite-64.sh[374]: swapoff /dev/vda1
[ 21.089841] testsuite-64.sh[376]: ++ mktemp
[ 21.095115] testsuite-64.sh[271]: + logfile=/tmp/tmp.a1MULA35wL
[ 21.095115] testsuite-64.sh[271]: + journalctl -b -q --no-pager -o short-monotonic -p info --grep 'Device path.*vda.?'\'' too long to fit into unit name'
...
[ 21.277360] systemd[1]: testsuite-64.service: Main process exited, code=exited, status=1/FAILURE
[ 21.277508] systemd[1]: testsuite-64.service: Failed with result 'exit-code'.
...
[ 21.323500] systemd[1]: Device path '/sys/devices/pci0000:00/0000:00:03.0/0000:01:00.0/0000:02:00.0/0000:03:00.0/0000:04:00.0/0000:05:00.0/0000:06:00.0/0000:07:00.0/0000:08:00.0/0000:09:00.0/0000:0a:00.0/0000:0b:00.0/0000:0c:00.0/0000:0d:00.0/0000:0e:00.0/0000:0f:00.0/0000:10:00.0/0000:11:00.0/0000:12:00.0/0000:13:00.0/0000:14:00.0/0000:15:00.0/0000:16:00.0/0000:17:00.0/0000:18:00.0/0000:19:00.0/0000:1a:00.0/virtio0/block/vda/vda1' too long to fit into unit name, ignoring device.
meson: drop three more single-use convenience libraries
The way that the cryptsetup plugins were built was unnecessarilly complicated.
We would build three static libraries that would then be linked into dynamic
libraries. No need to do this.
While at it, let's use a convenience library to avoid compiling the shared code
more than once.
We want the output .so files to be located in the main build directory,
like with all consumable build artifacts, so we need to maintain the split
between src/cryptsetup/cryptsetup-token/meson.build and the main meson.build
file.
AFAICT, the build artifacts are the same: exported and undefined symbols are
identical. There is a tiny difference in size, but I think it might be caused
by a different build directory name.
meson: drop convenience library that was only used in one place
It doesn't make much sense to do this, the result is very similar to including
to objects directly in the output binary without going through an intermediate
target.
The linkage of test-libudev was rather strange too: udev_link_with is used to
allow udev to be linked to a static version of libshared, so that udev is not
linked to libshared.so. But here we were using both, defeating the purpose of
udev_link_with. I don't think it matters what the test is linked to, so let's
use the non-static linkage to save space.
meson: stop building out convenience libraries by default
The meson default for static_library() are:
build_by_default=true, install=false. We never interact with the
static libraries, and we only care about them as a stepping-stone towards
the installable executables or libraries. Thus let's only build them if
they are a dependency of something else we are building.
While at it, let's drop install:false, since this appears to be the default.
This change would have fixed the issue with lib_import_common failing
to build too: we wouldn't attempt to build it.
In practice this changes very little, because we generally only declare static
libraries where there's something in the default target that will make use of
them. But it seems to be a better pattern to set build_by_default to false.
Use a 'convenience library' to do the compilation once and then link the
objects into all the files that need it. Those files are small, so this probably
doesn't matter too much for speed, but has the advantage that we don't get the
same error four times if something goes wrong.
The library is conditionalized in the same way importd itself, because we
cannot build it without the deps.
man: split out a paragraph and shorten the text about sd-network-generator
The ordering of the service wrt. to udevd is enforced by unit configuration,
so no need to tell the user about this. From users' POV, the only thing that
counts is that the unit is enabled and then the right thing happens.
units: enable systemd-network-generator by default
It is used by udevd and networkd. Since udevd is enabled statically, let's also
change the preset to "on". networkd is opt-in, so let's pull in the generator
when enabling networkd too.
Daan De Meyer [Wed, 15 Dec 2021 17:23:15 +0000 (18:23 +0100)]
journal: Stop reading in increments of block size during hole punching
Let's not try to be overly clever here. This code path is not overly
performance sensitive and we should avoid trying to outsmart the kernel
without proper benchmarking.
Daan De Meyer [Wed, 15 Dec 2021 17:18:25 +0000 (18:18 +0100)]
journal: Correctly advance offset when iterating hash table entries
pread() is not guaranteed to completely fill up the given buffer with
data which we assumed until now. Instead, only increment the offset by
the number of bytes that were actually read.
tree-wide: make FOREACH_DIRENT_ALL define the iterator variable
The variable is not useful outside of the loop (it'll always be null
after the loop is finished), so we can declare it inline in the loop.
This saves one variable declaration and reduces the chances that somebody
tries to use the variable outside of the loop.
For consistency, 'de' is used everywhere for the var name.
Daan De Meyer [Wed, 15 Dec 2021 12:58:24 +0000 (13:58 +0100)]
kernel-install: Remove "Default" from list of suffixes checked
This was an undocumented change in behavior introduced by 9e82a74cb0f08a288f9db228a0b5bec8a7188cdb. Previously, we only
checked for "Default" if we didn't find a machine ID. Let's make
sure we keep the previous behavior intact.
resolved: filter out our own stub resolvers when parsing servers
We get "upstream" dns server config from ~three places: /etc/resolv.conf,
config files, and runtime config via dbus. With this commit, we'll filter out
our own stub listeners if they are configured in either of the first two
sources. For /etc/resolv.conf this is done quitely, and for our own config
files, a LOG_INFO message is emitted, since this is a small inconsistency in
the config.
Setting loops like this over dbus is still allowed. The reason is that in the
past we didn't treat this as an error, and if we were to start responding with
an error, we could break a scenario that worked previously. E.g. NM sends us a
list of servers, and one happens to be the our own. We would just not use that
stub server before, but it'd still be shown in the dbus properties and such.
We would have to return error for the whole message, also rejecting the other
valid servers. I think it's easier to just keep that part unchanged.
Test case:
$ ls -l /etc/resolv.conf
-rw-r--r-- 1 root root 57 Dec 15 10:26 /etc/resolv.conf
$ cat /etc/resolv.conf
nameserver 192.168.150.1
options edns0 trust-ad
search .
$ cat /etc/systemd/resolved.conf.d/stub.conf
[Resolve]
DNSStubListenerExtra=192.168.150.1
$ resolvectl
...
Global
resolv.conf mode: foreign
DNS Servers: 192.168.150.1
Fallback DNS Servers: ...
(with the patch):
Global
resolv.conf mode: foreign
Fallback DNS Servers: ...
journactl: show info about journal range only at debug level (#21775)
The message that the "journal begins … ends …" has been always confusing to
users. (Before b91ae210e62 it was "logs begin … end …" which was arguably even
more confusing, but really the change in b91ae210e62 didn't substantially change
this.)
When the range shown is limited (by -e, -f, --since, or other options), it
doesn't really matter to the user what the oldest entries are, since they are
purposefully limiting the range. In fact, if we are showing the last few
entries with -e or -f, knowing that many months the oldest entries have is
completely useless.
And when such options are *not* used, the first entry generally corresponds to
the beginning of the range shown, and the last entry corresponds to the end of
that range. So again, it's not particularly useful, except when debugging
journalctl or such. Let's just treat it as a debug message.
hostnamectl: add a chassis symbol in status output
The idea is to be able to distinguish whether we're in a VM/container or something
more substantial at a glance.
Chassis: laptop 💻
Chassis: tablet 具
Chassis: vm 🖴
Chassis: server 🖳
Chassis: handset 🕻
Chassis: watch ⌚
Chassis: desktop 🖥
Chassis: container ☐
Nishal Kulkarni [Tue, 14 Dec 2021 08:43:13 +0000 (14:13 +0530)]
shell-completion: Add completion for systemd-analyze critical-chain
systemd-analyze critical-chain accepts an optional unit argument,
however currently there's no shell-completion for it
This change provides unit name completion for both bash and zsh.
projects / thirdparty / systemd.git / log
