Michal Schmidt [Tue, 29 Nov 2011 22:14:36 +0000 (23:14 +0100)]
shutdownd: use PassCred=yes in the socket unit
Since Linux 3.2 in order to receive SCM_CREDENTIALS it is not sufficient
to set SO_PASSCRED just before recvmsg(). The option has to be already
set when the sender sends the message.
With socket activation it is too late to set the option in the service.
It must be set on the socket right from the start.
See the kernel commit: 16e57262 af_unix: dont send SCM_CREDENTIALS by default
Michal Schmidt [Wed, 16 Nov 2011 23:21:16 +0000 (00:21 +0100)]
execute: log errors from "sd(EXEC)"
To give the administrator more hints about failures occuring in spawning
of commands than just the exit code, log the strerror.
All fds are closed, so reopen the log.
Michal Schmidt [Wed, 16 Nov 2011 22:45:01 +0000 (23:45 +0100)]
execute: avoid logging to closed fds
Several functions called from the "sd(EXEC)" process try to log messages
when all the file descriptors are already closed, including the logging
ones. The logging functions do not expect their fds to be closed and
they hit an assertion failure. The failure wants to be logged too,
so there is an infinite recursion, ended by a SIGSEGV.
When we close all fds, we must let log.c know about it.
Michal Schmidt [Thu, 10 Nov 2011 08:55:47 +0000 (09:55 +0100)]
service: don't warn if the pidfile still exists after SIGCHLD
A service that drops its privileges may not be able to remove it when it
exits. The stale pidfile is not a problem as long as the service
carefully recognizes it on its next start.
systemd would produce a warning after the service exits:
PID ... read from file ... does not exist. Your service or init
script might be broken.
Silence the warning in this case. Still warn if this error is detected
when loading the pidfile after service start.
Noticed by Miroslav Lichvar in
https://bugzilla.redhat.com/show_bug.cgi?id=752396
Michal Schmidt [Wed, 9 Nov 2011 07:42:03 +0000 (08:42 +0100)]
units: drop unnecessary 'StandardOutput=syslog'
DefaultStandardOutput is syslog anyway. There's no reason to assume that
the administrator would want these units to be excluded when he configures
a different DefaultStandardOutput.
cgroup: immediately remove all cgroups which run empty
Some controllers have scaling problems when many empty cgroups exist.
Hence, as soon as we get a notification that a cgroup is empty, delete
it. This is also nice to keep the systemd-cgls output short.
Ran Benita [Wed, 2 Nov 2011 08:48:49 +0000 (10:48 +0200)]
bash-completion: update with new verbs and arguments
Adds arguments --root= --runtime --no-legend.
Adds verbs link mask unmask reenable list-unit-files.
Also uses list-unit-files to make nicer enable and disable completions.
Dexter Morgan [Fri, 28 Oct 2011 17:16:33 +0000 (19:16 +0200)]
Add Mageia support
This patch adds support for the Mageia Linux distribution:
http://www.mageia.org/
Mageia is a fork of Mandriva although some divergence has already occured
and thus inclusion of these changes upstream allow us to (hopefully)
migrate more rapidly to the new standard approaches systemd offers.
Indeed, we already use the preferred mechanism of OS identification via
the /etc/os-release file rather than a distro specific variation.
This patch mostly mirrors the patch added previously for Mandriva
support. In addition to those original authors, this patch was mostly
written by Dexter Morgan with help from Colin Guthrie and Eugeni Dodonov.
Tom Gundersen [Mon, 17 Oct 2011 11:01:08 +0000 (13:01 +0200)]
cryptsetup-generator: avoid ordering cycle on swap
Devices with random keys (swap), should not be ordered before local-fs.target,
as this creates a cycle with systemd-load-random-seed.service (and also it
does not make sense, a swap device is not a local-fs).
mount: order remote mounts after both network.target and remote-fs-pre.target
Since remote-fs-pre.target is optional we cannot count on it to order
remote mounts after network.target, so let's add that order explicitly
in addition to remote-fs-pre.target.
cgroup: always recreate cgroup before we try to apply attributes
We might have trimmed the cgroup tree previously, hence don't trust our
own "realized" flag, always recreate cgroup tree before applying our
attributes to make sure this actually works out.
initctl: don't use dbus connection after PID 1 got respawned
After reexec PID 1 our bus connection is invalidated. Hence don't try to
reuse it, just terminate so that when we are spawned the next time we
just get a new one.
systemctl: make list-unit-files output more economical
The first column is given the width of the widest entry,
if possible, otherwise all entries are ellipsized to fit
in ($COLUMNS - (width of second column)).
[ Added a few fixes, calculate state_cols too, respect '--no-legend',
better handling of '--full' -- michich ]
Dave Reisner [Wed, 12 Oct 2011 00:56:53 +0000 (20:56 -0400)]
systemctl-completion: always invoke with --no-legend
In the case of completion for the 'restart' verb, passing the invalid
unit name (the colums header) causes completion to cease functioning
entirely, with the error:
Failed to issue method call: Unit name UNIT is not valid.
This adds a small wrapper function for systemctl which can have common
options added to it.
Jonathan Nieder [Mon, 17 Oct 2011 19:01:40 +0000 (21:01 +0200)]
audit: do not complain if kernel lacks audit
When running on a kernel without audit support, systemd currently
writes a mysterious-sounding error to its log:
systemd[1]: Failed to connect to audit log: Protocol not supported
Better to suppress the audit_open() failure message when (and only
when) it is due to running on a kernel without audit support, since in
this case the admin probably does not mind systemd not writing to the
audit log. This way, more serious errors like ENOMEM and EACCES will
stand out more.
Tollef Fog Heen [Mon, 17 Oct 2011 19:00:42 +0000 (21:00 +0200)]
service: Drop rcN.d runlevels from SysV services that also exist in rcS.d
Services which claim to start in both rcN.d and rcS.d generate
loops which for some reason seems to usually end up with dbus not
starting and the whole machine being quite unhappy. We now rather
assume that if a service can be started in rcS, it should not also
start in rcN.d.
Michal Schmidt [Mon, 17 Oct 2011 09:12:12 +0000 (11:12 +0200)]
manager: fix a crash in isolating
HASHMAP_FOREACH is safe against the removal of the current entry, but
not against the removal of other entries. job_finish_and_invalidate()
can recursively remove other entries.
It triggered an assertion failure:
Assertion 'j->installed' failed at src/manager.c:1218, function
transaction_apply(). Aborting.
Fix the crash by iterating from the beginning when there is a
possibility that the iterator could be invalid.
It is O(n^2) in the worst case, but that's better than a crash.
service: don't try to guess PID for SysV services anymore
As it turns out there are quite a number of SysV services too broken to
make the guessing work: instead of returning in the parent only after
the child is fully initialized they return immediately. The effect is
that the guessing in systemd might happen too early, at a time where the
final main process doesn't exist yet.
By turning this off we won't try to detect the main pid anymore, with
the effect that all processes of the service in question are considered
equally likely to be the main process.
src/pager.c: In function ‘pager_fallback’:
src/pager.c:35:13: warning: function might be possible candidate for attribute ‘noreturn’ [-Wmissing-noreturn]
units: remount root and API FS before all mount units are applied
In order to ensure that bind mounts copy the final mount settings to the
new bind mount make the root and API FS mount options are applied before
the other file systems are mounted.
Michal Schmidt [Thu, 29 Sep 2011 19:18:17 +0000 (21:18 +0200)]
systemctl: fix corrupted output of units
On some systems the list of units in systemctl output came out wrong,
all on one line and with missing descriptions.
It turns out printf() really attempts to mmap 2G memory when INT_MAX is
passed as the field width. On machines with small virtual memory
without overcommit the mmap inside printf() failed.
KEYMAP are use to pass keymap configuration to initrd, but not to the
system itself. Since the initrd might get out of date we need to make
sure that changes made in userspace override the settings from the
cmdline, hence drpo any use of it all for these variables.
Michal Schmidt [Tue, 27 Sep 2011 09:20:20 +0000 (11:20 +0200)]
systemctl: better width and alignment of fields
Provide nicer output by taking advantage of the available columns.
Expand UNIT and DESCRIPTION if possible. Economize the space:
- never give UNIT more width than it can use.
- if there's a shortage of space, give 25 columns to UNIT, even
if it would result in overflowing the total columns. The field would not
be useful below width 25.
- then if possible, assign up to 25 columns to DESCRIPTION.
- then split any remaining space evenly between UNIT and DESCRIPTION.
Michal Schmidt [Mon, 26 Sep 2011 22:48:40 +0000 (00:48 +0200)]
systemctl: print the legend regardless of stdout type, add --no-legend
Print the legend (the column headers and the footer with hints) by
default even to non-tty output. People seem to get confused by the
difference when they redirect the output.
Add a parameter to suppress the printing of the legend.
Michal Schmidt [Fri, 23 Sep 2011 11:58:00 +0000 (13:58 +0200)]
unit: fix complementing of requirement deps with After deps for targets
'man systemd.target' says:
Unless DefaultDependencies= is set to false, target units will
implicitly complement all configured dependencies of type
Wants=, Requires=, RequiresOverridable= with dependencies of type
After= if the units in question also have DefaultDependencies=true.
projects / systemd / log