If the systemd version on the host is too old and there's no local
build directory, use the default tools tree which will build an
image containing all the tooling required to build systemd and use
that to build the other presets.
Let's not build all presets by default, but only the system preset.
Also, let's only make the system preset depend on the initrd preset
if we're building a bootable disk image. This means that if we build
the system image as a cpio, uki or directory, the initrd preset will
be skipped as it's not necessary, making the build a little faster.
The kernel has this horrible build system feature where distros
can ship /sbin/installkernel and it'll automatically be used by
make install. Let's make sure that doesn't happen as on Debian this
script puts the kernel under the wrong name causing mkosi build
failures.
Let's stop manually destroying the timers when we fail to spawn a child.
We don't do this in any of the similar codepaths in any of the unit
types, only in two specific ones in socket/swap. Destroying the timer is
unnecessary, since this is done anyway in the _set_state() call of each
unit type if not appropriate, and every failure path here runs through
that anyway.
This brings all these similar codepaths into sync.
The first conversion to PidRef. It's mostly an excercise of
search/replace, but with some special care taken for life-cycle (i.e. we
need to destroy the structure properly once done, to release the pidfd).
It also uses pidfd based killing for some of the killing but leaves most
as it is to make the conversion minimal.
Let's start with the conversion of PID 1 to pidfds. Let's add a simple
structure with just two fields that can be used to maintain a reference
to arbitrary processes via both pid_t and pidfd.
This is an embeddable struct, to keep it in line with where we
previously used a pid_t directly to track a process.
Of course, since this might contain an fd on systems where we have pidfd
this structure has a proper lifecycle.
(Note that this is quite different from sd_event_add_child() event
source objects as that one is only for child processes and collects
process results, while this infra is much simpler and more generic and
can be used to reference any process, anywhere in the tree.)
I'm not quite sure what the original intent of this line was, but it
doesn't work in the one call-site the "required" argument is actually
used. The "writable" flag was indexed as a scalar leaving only the
"e" to compare against.
Instead, let's just sort the parsed flags and compare the whole thing.
Also substitute "required" as a pattern, so that pattern based
comparisons may be supported.
Follow-up for 53c0397b1d
* Split do_wall into wall_utmp and wall_logind
* Don't pass unused arguments
* Add missing asserts
* Modernize error handling
* Don't do anything if neither utmp nor logind is enabled
Without memory accounting explicitly disabled, we may not obtain current
consumption from all units on the ancestry path.
Use a descendant value as lower bound estimate for ancestors if
ancestor's consumption cannot be directly queried.
This makes MemoryAvailable= an upper bound of available values.
unit_get_memory_available() duplicates similar logic contained in
unit_get_memory_current(). Instead, it can call it for each unit it
needs data for.
Additionally, simplify the flow by treating all units from leaf to root
uniformly in one loop.
Functional change when a queried unit does not have MemoryAccounting=yes
(or cgroup_path), we will try getting an estimate from ancestors.
Currently, querying a unit's available memory would result in infinity
if there are no limits set on the unit or ancestors.
That undermines semantics implied by the name, so look at the physical
memory if the search propagates up to the -.slice.
This makes sense even in systemd user instances, limits of -.slice are
still looked at too.
Also change printed representation of infinite MemoryAvailable which
means we could not figure out a good estimate.
Otherwise we'll skip parsing main config directly.
Replaces #27152
Replaces #28956:
chase() is an overkill for determining whether
two files are the same. We only need to check
if the inodes are the same.
If the user does not specify a config file to use, ukify will try looking for one at {/run,/etc,/usr/local/lib,/usr/lib}/systemd/ukify.conf in order and then use the first one found. Also made sure the --config input is a pathlib.Path by specifying its type in its CONFIG_ITEMS entry.
Big cheers to Zbigniew Jędrzejewski-Szmek <zbyszek@in.waw.pl> for helping!
This is unused these days, we instead calculate policy hashes ourselves
and do not need trial sessions for that anymore. Given how unreliable
they are in actual TPMs, we don't want to use them anyway, so I don't
think we ever want to readd this.
hence, let's drop this unused feature.
Luca Boccassi
Daan De Meyer
Mike Yuan
Lennart Poettering
Michal Koutný
Ronan Pigott
Alvin Alvarado
Yu Watanabe
Frantisek Sumsal
Jan Janssen
Zbigniew Jędrzejewski-Szmek