chiark / gitweb /
10 years agosystemctl: fix corrupted output of units
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.

Do not use INT_MAX for unlimited width.

10 years agolocaled: handle multiple X11 kbd layouts
Lennart Poettering [Thu, 29 Sep 2011 17:44:32 +0000 (19:44 +0200)]
localed: handle multiple X11 kbd layouts

10 years agosystemctl: fix error message when no connection to systemd is available
Lennart Poettering [Thu, 29 Sep 2011 14:00:29 +0000 (16:00 +0200)]
systemctl: fix error message when no connection to systemd is available

10 years agosystemctl: fix detection whether the dbus connection is available
Lennart Poettering [Thu, 29 Sep 2011 13:31:53 +0000 (15:31 +0200)]
systemctl: fix detection whether the dbus connection is available

10 years agobuild-sys: fix build
Lennart Poettering [Thu, 29 Sep 2011 13:09:18 +0000 (15:09 +0200)]
build-sys: fix build

10 years agovconsole: don't parse Fedora's KEYMAP= kernel parameters anymore
Lennart Poettering [Wed, 28 Sep 2011 02:25:20 +0000 (04:25 +0200)]
vconsole: don't parse Fedora's KEYMAP= kernel parameters anymore

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 Since the initrd might get out of date we need to make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.

10 years agolocaled: add SetX11Keyboard() and SetVConsoleKeyboard() bus calls
Lennart Poettering [Wed, 28 Sep 2011 02:25:13 +0000 (04:25 +0200)]
localed: add SetX11Keyboard() and SetVConsoleKeyboard() bus calls

10 years agoanalyze: always draw top kernel, initramfs, userspace bars
Kay Sievers [Tue, 27 Sep 2011 10:52:09 +0000 (12:52 +0200)]
analyze: always draw top kernel, initramfs, userspace bars

Based on Koen Kooi's patch. Renamed initrd to initramfs.

10 years agosystemctl: better width and alignment of fields
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.

Keep the columns aligned even with '--full'.

10 years agosystemctl: print the legend regardless of stdout type, add --no-legend
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.


10 years agopager: add a trivial internal pager
Michal Schmidt [Mon, 26 Sep 2011 21:02:56 +0000 (23:02 +0200)]
pager: add a trivial internal pager

In the very unlikely scenario where none of the external pagers is
available, use an internal implementation to pass stdin to stdout.

Don't bother with trying 'cat', because it's no more useful than the
internal pager.

10 years agobuild-sys: bump release v36
Lennart Poettering [Fri, 23 Sep 2011 16:30:02 +0000 (18:30 +0200)]
build-sys: bump release

10 years agoreadahead: opt out of bridge logging for readahead in order not to delay things unnec...
Lennart Poettering [Fri, 23 Sep 2011 15:42:54 +0000 (17:42 +0200)]
readahead: opt out of bridge logging for readahead in order not to delay things unnecessarily

10 years agoservice: fix up std output/error before we add dependencies to the bridge socket
Lennart Poettering [Fri, 23 Sep 2011 15:42:21 +0000 (17:42 +0200)]
service: fix up std output/error before we add dependencies to the bridge socket

10 years agovirt: add missing header inclusion
Lennart Poettering [Fri, 23 Sep 2011 15:20:45 +0000 (17:20 +0200)]
virt: add missing header inclusion

10 years agocondition: optionally test against type of virtualization (vm vs. container)
Lennart Poettering [Fri, 23 Sep 2011 15:09:49 +0000 (17:09 +0200)]
condition: optionally test against type of virtualization (vm vs. container)

10 years agoutil: move virtualization detection into its own files, and extend return codes
Lennart Poettering [Fri, 23 Sep 2011 15:00:33 +0000 (17:00 +0200)]
util: move virtualization detection into its own files, and extend return codes

10 years agoanalyze: report startup time in plot mode as well
Koen Kooi [Thu, 22 Sep 2011 13:24:18 +0000 (15:24 +0200)]
analyze: report startup time in plot mode as well

It now prints something like "Startup finished in 1507ms (kernel) + 850ms (userspace) = 2357ms" below the legend.

10 years agoanalyze: always draw 1s marker for scale
Koen Kooi [Thu, 22 Sep 2011 09:30:04 +0000 (11:30 +0200)]
analyze: always draw 1s marker for scale

In situations like this:

root@omap4430-panda:~# systemd-analyze
Startup finished in 1499ms (kernel) + 916ms (userspace) = 2416ms

The svg plot will only have the 0s marker and no subsequent markers for scale. This patch forces the 1s marker to always be drawn.

10 years agounit: fix complementing of requirement deps with After deps for targets
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' 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.

It did not work because of a forgotten negation.

10 years agoutil: don't fail if no id was passed to detect_container()
Lennart Poettering [Fri, 23 Sep 2011 03:33:33 +0000 (05:33 +0200)]
util: don't fail if no id was passed to detect_container()

10 years agoutil: detect systemd-nspawn without relying on ns cgroup tree
Lennart Poettering [Fri, 23 Sep 2011 02:38:39 +0000 (04:38 +0200)]
util: detect systemd-nspawn without relying on ns cgroup tree

10 years agonspawn: set env var container=systemd-nspawn, following the scheme lxc introduced
Lennart Poettering [Fri, 23 Sep 2011 02:03:18 +0000 (04:03 +0200)]
nspawn: set env var container=systemd-nspawn, following the scheme lxc introduced

10 years agologinctl: fix typo
Lennart Poettering [Fri, 23 Sep 2011 00:53:55 +0000 (02:53 +0200)]
loginctl: fix typo

10 years agollvm-analyze: change a few things to make llvm-analyze show fewer false positives
Lennart Poettering [Fri, 23 Sep 2011 00:39:54 +0000 (02:39 +0200)]
llvm-analyze: change a few things to make llvm-analyze show fewer false positives

10 years agollvm-analyze: fix some bugs found by llvm-analyze
Lennart Poettering [Fri, 23 Sep 2011 00:39:28 +0000 (02:39 +0200)]
llvm-analyze: fix some bugs found by llvm-analyze

10 years agoutil: improve signal_to_string, signal_from_string
Michal Schmidt [Fri, 23 Sep 2011 07:39:40 +0000 (09:39 +0200)]
util: improve signal_to_string, signal_from_string

 Produce names for SIGRTMIN+n.
 Never give an "n/a" result. In the worst case give the number itself as
 a string.

 Parse "RTMIN+n".
 Parse any valid signal number.

10 years agomanager: fix job mode for SIGRTMIN+1, +2
Michal Schmidt [Fri, 23 Sep 2011 07:37:45 +0000 (09:37 +0200)]
manager: fix job mode for SIGRTMIN+1, +2

The test did not work as intended and always resulted in JOB_REPLACE.

10 years agocondition: fix reversed tests if path does not exist at all
Michal Schmidt [Fri, 23 Sep 2011 00:10:00 +0000 (02:10 +0200)]
condition: fix reversed tests if path does not exist at all

CONDITION_FILE_IS_EXECUTABLE gave reversed results when the path
did not exist at all.

10 years agocoverity: change a few things so that coverity doesn't show so many false positives
Lennart Poettering [Thu, 22 Sep 2011 23:44:36 +0000 (01:44 +0200)]
coverity: change a few things so that coverity doesn't show so many false positives

10 years agocoverity: fix a couple of bugs found by coverity
Lennart Poettering [Thu, 22 Sep 2011 23:43:28 +0000 (01:43 +0200)]
coverity: fix a couple of bugs found by coverity

10 years agounits: fix rescue.service race with plymouth
Michal Schmidt [Thu, 22 Sep 2011 21:52:38 +0000 (23:52 +0200)]
units: fix rescue.service race with plymouth

It was possible for the "ExecStartPre=-/bin/plymouth quit" to race
with plymouth-start.service which is pulled in indirectly by ->

The race left plymouth running on the terminal, making it unusable for
rescue purposes.

10 years agoman: document that PathChanged= does not care for simple writes but only for close...
Lennart Poettering [Thu, 22 Sep 2011 19:37:01 +0000 (21:37 +0200)]
man: document that PathChanged= does not care for simple writes but only for close()-after-write()

10 years agounit: don't recheck conditions when a unit is already starting but unit_start() is...
Lennart Poettering [Thu, 22 Sep 2011 19:32:18 +0000 (21:32 +0200)]
unit: don't recheck conditions when a unit is already starting but unit_start() is invoked

10 years agoman: document pkg-config line for sd-daemon
Lennart Poettering [Thu, 22 Sep 2011 19:16:33 +0000 (21:16 +0200)]
man: document pkg-config line for sd-daemon

10 years agoman: document that sd-daemon.[ch] is now available as shared library
Lennart Poettering [Thu, 22 Sep 2011 19:13:41 +0000 (21:13 +0200)]
man: document that sd-daemon.[ch] is now available as shared library

10 years agoservice: warn if a service fails to write its PID file
Michal Schmidt [Thu, 22 Sep 2011 11:10:32 +0000 (13:10 +0200)]
service: warn if a service fails to write its PID file

Warn if a service promises to write a PID file (using 'PIDFile=' in the
unit file or '# pidfile:' in SysV header), but fails to keep the

This warning will likely trigger also for the forking services with a
racy daemonization, which exit the original process before the PID file
is written.

10 years agoupdate TODO
Lennart Poettering [Thu, 22 Sep 2011 02:18:18 +0000 (04:18 +0200)]
update TODO

10 years agoselinux: reopen log when loading the SELinux policy fails
Lennart Poettering [Thu, 22 Sep 2011 02:05:40 +0000 (04:05 +0200)]
selinux: reopen log when loading the SELinux policy fails

10 years agoservice: if StandardInput=socket and StandardOutput=inherit imply socket for output...
Lennart Poettering [Thu, 22 Sep 2011 01:29:51 +0000 (03:29 +0200)]
service: if StandardInput=socket and StandardOutput=inherit imply socket for output, don't imply default output

This is useful for inetd-style per-connection services, so that they
again can simply specify StandardOutput=socket to connect all three fds
to the socket.

10 years agoudev: asynchronously mount fuse/confifgs as soon as the kmods are loaded
Lennart Poettering [Thu, 22 Sep 2011 01:02:15 +0000 (03:02 +0200)]
udev: asynchronously mount fuse/confifgs as soon as the kmods are loaded

10 years agounits: introduce unit file for FUSE fs
Lennart Poettering [Thu, 22 Sep 2011 01:00:04 +0000 (03:00 +0200)]
units: introduce unit file for FUSE fs

FUSE and configfs is very very similar, so handle both the same way.

10 years agounits: In MeeGo there was more than one ExecStart lines present that caused systemd...
Marko Saukko [Wed, 21 Sep 2011 19:38:46 +0000 (21:38 +0200)]
units: In MeeGo there was more than one ExecStart lines present that caused systemd to refuse the rescue.service

10 years agomount: pull in quota tools from fstab lines with 'quota' option, too
Lennart Poettering [Wed, 21 Sep 2011 18:19:43 +0000 (20:19 +0200)]
mount: pull in quota tools from fstab lines with 'quota' option, too

10 years agojob: after converting a job from restart to start readd it to run queue
Lennart Poettering [Wed, 21 Sep 2011 17:56:15 +0000 (19:56 +0200)]
job: after converting a job from restart to start readd it to run queue

When we converted a restart job it's time to look again whether any of
its dependencies are now runnable.

10 years agoupdate TODO
Kay Sievers [Wed, 21 Sep 2011 17:33:52 +0000 (19:33 +0200)]
update TODO

10 years agounits: we don't need to check whether /sys/kernel/security is already mounted since...
Lennart Poettering [Wed, 21 Sep 2011 16:06:24 +0000 (18:06 +0200)]
units: we don't need to check whether /sys/kernel/security is already mounted since we'll enumerate that fact anyway

10 years agoupdate TODO
Lennart Poettering [Wed, 21 Sep 2011 01:56:07 +0000 (03:56 +0200)]
update TODO

10 years agounits: drop automount unit for /dev/hugepages
Lennart Poettering [Wed, 21 Sep 2011 01:35:19 +0000 (03:35 +0200)]
units: drop automount unit for /dev/hugepages

Since hugetlbfs cannot be compiled as kernel module there's little point
in doing on-demand mounting via autofs for it.

10 years agounits: drop automount unit for /dev/mqueue
Lennart Poettering [Wed, 21 Sep 2011 01:29:47 +0000 (03:29 +0200)]
units: drop automount unit for /dev/mqueue

Since the mqueue support cannot be built as a module there's little
benefit in having an autofs mount point set up for this.

10 years agounits: drop automount unit for /sys/kernel/security
Lennart Poettering [Wed, 21 Sep 2011 01:26:37 +0000 (03:26 +0200)]
units: drop automount unit for /sys/kernel/security

Since securityfs cannot be build as module there's little value in
having an on-demand autofs mount point for it.

10 years agounits: drop automount unit for /sys/kernel/debug
Lennart Poettering [Wed, 21 Sep 2011 01:20:28 +0000 (03:20 +0200)]
units: drop automount unit for /sys/kernel/debug

Since debugfs cannot be compiled as module there's little benefit in
having it as autofs mount point.

10 years agounits: drop /sys/kernel/config automount unit
Lennart Poettering [Wed, 21 Sep 2011 01:16:00 +0000 (03:16 +0200)]
units: drop /sys/kernel/config automount unit

The mount point directory /sys/kernel/config is only created after the
module is loaded, hence there's little value in having this an automount
unit: the runtime penalty for mounting an autofs here should be the same
as for a real mount.

10 years agomodules-load: filter out double modules
Lennart Poettering [Wed, 21 Sep 2011 00:57:02 +0000 (02:57 +0200)]
modules-load: filter out double modules

10 years agoservice: minor change in service_load_pid_file return value
Michal Schmidt [Wed, 21 Sep 2011 00:25:17 +0000 (02:25 +0200)]
service: minor change in service_load_pid_file return value

Return 0 only if the PID was really loaded. If no PIDFile= is defined,
return -ENOENT.

Only one caller cares about the return value of this function and this
change makes the usage nicer.

10 years agocondition: add ConditionPathIsSymbolicLink
Michal Schmidt [Tue, 20 Sep 2011 23:29:38 +0000 (01:29 +0200)]
condition: add ConditionPathIsSymbolicLink

10 years agocondition: always follow symlinks for condition checks, to mimic test
Lennart Poettering [Tue, 20 Sep 2011 23:07:25 +0000 (01:07 +0200)]
condition: always follow symlinks for condition checks, to mimic test

10 years agoservice: handle forking services that move to a new PID
Michal Schmidt [Tue, 20 Sep 2011 19:43:30 +0000 (21:43 +0200)]
service: handle forking services that move to a new PID

When some forking daemons receive a SIGHUP, they re-execute themselves
and consequently change to a new main PID. As long as they update the
PID file in the right order (before exiting the old PID), we can detect
that and avoid killing them.

10 years agounits: don't initialize automount point for securityfs if the initrd already mounted...
Lennart Poettering [Tue, 20 Sep 2011 22:45:22 +0000 (00:45 +0200)]
units: don't initialize automount point for securityfs if the initrd already mounted the directory

10 years agounit: introduce ConditionPathIsMountPoint=
Lennart Poettering [Tue, 20 Sep 2011 22:44:51 +0000 (00:44 +0200)]
unit: introduce ConditionPathIsMountPoint=

10 years agoman: document which condition tests follow symlinks
Michal Schmidt [Tue, 20 Sep 2011 10:28:22 +0000 (12:28 +0200)]
man: document which condition tests follow symlinks

ConditionPathExists and ConditionFileIsExecutable follow symlinks.
ConditionPathIsDirectory does not follow symlinks.
(XXX: is this actually the desired behavior?)

10 years agounits: direct stdout/stderr of rescue shells to tty
Michal Schmidt [Tue, 20 Sep 2011 10:11:04 +0000 (12:11 +0200)]
units: direct stdout/stderr of rescue shells to tty

With output of services going to syslog by default now, the rescue shell
units need to direct their output to tty explicitly.
Specify stderr too, just in case.

10 years agounits: Add support for automounting configfs, ala debugfs, etc.
Bill Nottingham [Tue, 13 Sep 2011 15:01:56 +0000 (11:01 -0400)]
units: Add support for automounting configfs, ala debugfs, etc.

10 years agoupdate TODO
Lennart Poettering [Tue, 20 Sep 2011 01:33:49 +0000 (03:33 +0200)]
update TODO

10 years agostdout-syslog-bridge: properly handle overly long log lines
Lennart Poettering [Tue, 20 Sep 2011 00:46:04 +0000 (02:46 +0200)]
stdout-syslog-bridge: properly handle overly long log lines

10 years agoservice: drop broken support for X-Interactive fields in LSB headers
Lennart Poettering [Mon, 19 Sep 2011 23:33:48 +0000 (01:33 +0200)]
service: drop broken support for X-Interactive fields in LSB headers

The logic never worked since reading from the boot console is useless
when a service is started after boot. Hence drop this half-baked code,
since we now have a place document incompatibilities like this.

10 years agocondition: in ConditionPathIsExecutable follow symlinks
Lennart Poettering [Mon, 19 Sep 2011 23:28:00 +0000 (01:28 +0200)]
condition: in ConditionPathIsExecutable follow symlinks

10 years agocondition: opt out of /proc/cmdline parsing only when running in a container, not...
Lennart Poettering [Mon, 19 Sep 2011 21:51:15 +0000 (23:51 +0200)]
condition: opt out of /proc/cmdline parsing only when running in a container, not in a VM

10 years agohostnamed: Fix "SetStaticHostname" debug output
Bastien Nocera [Sun, 18 Sep 2011 13:10:07 +0000 (14:10 +0100)]
hostnamed: Fix "SetStaticHostname" debug output

Was using the wrong property (eg. not the one we just set)

10 years agoman: list specifiers in a table
Mathieu Bridon [Mon, 5 Sep 2011 08:35:57 +0000 (16:35 +0800)]
man: list specifiers in a table

It was noted on IRC that the current layout is not necessarily the
easiest to parse by humans. I personnally struggled understanding it the
first time I read it.

Hopefully, this change makes it easier to understand.

Note: I only reformatted the information that was already present in the
documentation, I didn't add anything.

10 years agoman: fix typo in systemd manual page
Albert Strasheim [Mon, 5 Sep 2011 08:45:12 +0000 (08:45 +0000)]
man: fix typo in systemd manual page

10 years agotimedate: fix interfaces list
Lennart Poettering [Sun, 4 Sep 2011 13:00:13 +0000 (15:00 +0200)]
timedate: fix interfaces list

10 years agodbus: terminate mechanism daemons after a while
Lennart Poettering [Sat, 3 Sep 2011 01:26:58 +0000 (03:26 +0200)]
dbus: terminate mechanism daemons after a while

10 years agostdout-bridge: set facility of messages with no facility to configured facility inste...
Lennart Poettering [Thu, 1 Sep 2011 19:15:00 +0000 (21:15 +0200)]
stdout-bridge: set facility of messages with no facility to configured facility instead of LOG_USER

10 years agofsck: show progress while fscking at boot
Lennart Poettering [Thu, 1 Sep 2011 19:05:06 +0000 (21:05 +0200)]
fsck: show progress while fscking at boot

10 years agobuild-sys: bump release v35
Lennart Poettering [Thu, 1 Sep 2011 00:10:25 +0000 (02:10 +0200)]
build-sys: bump release

10 years agolabel: fix database check
Lennart Poettering [Wed, 31 Aug 2011 23:16:55 +0000 (01:16 +0200)]
label: fix database check

10 years agopo: Initial Polish translation
Piotr Drąg [Mon, 22 Aug 2011 13:45:38 +0000 (15:45 +0200)]
po: Initial Polish translation

10 years agolocale: support $LANGUAGE too
Lennart Poettering [Wed, 31 Aug 2011 12:09:14 +0000 (14:09 +0200)]
locale: support $LANGUAGE too

$LANGUAGE is a GNU extension that is probably worth supporting, since it
allows specifiying an order of languages.

10 years agogetty: simplify things a bit
Lennart Poettering [Wed, 31 Aug 2011 02:22:41 +0000 (04:22 +0200)]
getty: simplify things a bit

10 years agogetty: automatically add getty on hvsi0 virtualizer console
Lennart Poettering [Wed, 31 Aug 2011 02:15:42 +0000 (04:15 +0200)]
getty: automatically add getty on hvsi0 virtualizer console

10 years agodropin: don't fail if random files are stored in .wants directories
Lennart Poettering [Wed, 31 Aug 2011 01:55:38 +0000 (03:55 +0200)]
dropin: don't fail if random files are stored in .wants directories

10 years agosystemctl: if we managed to reexec the init system via the bus don't retry via signal
Lennart Poettering [Wed, 31 Aug 2011 01:31:27 +0000 (03:31 +0200)]
systemctl: if we managed to reexec the init system via the bus don't retry via signal

If we managed to reeexec the init system via the bus this allows us to
provide synchronous behaviour to callers. This is all lost if we then
repeat the reexecution via SIGTERM.

10 years agologind: use syspath instead of sysname to identify fb devices since udlfb got fixed now
Lennart Poettering [Wed, 31 Aug 2011 01:16:01 +0000 (03:16 +0200)]
logind: use syspath instead of sysname to identify fb devices since udlfb got fixed now

10 years agoman: document /etc/timezone
Lennart Poettering [Wed, 31 Aug 2011 01:09:22 +0000 (03:09 +0200)]
man: document /etc/timezone

10 years agomount: fix parsing of prio value
Lennart Poettering [Tue, 30 Aug 2011 23:46:38 +0000 (01:46 +0200)]
mount: fix parsing of prio value

10 years agoman: don't recommend anymore since we don't support non-socket...
Lennart Poettering [Tue, 30 Aug 2011 23:35:43 +0000 (01:35 +0200)]
man: don't recommend anymore since we don't support non-socket-activatable syslogs anymore where that was ncessary

10 years agogetty-generator: ignore if symlinks already exist
Lennart Poettering [Tue, 30 Aug 2011 23:30:57 +0000 (01:30 +0200)]
getty-generator: ignore if symlinks already exist

10 years agolabel: minor fixes
Lennart Poettering [Tue, 30 Aug 2011 23:21:02 +0000 (01:21 +0200)]
label: minor fixes

10 years agolabel: introduce label_bind() and make use of it where necessary
Lennart Poettering [Tue, 30 Aug 2011 23:02:00 +0000 (01:02 +0200)]
label: introduce label_bind() and make use of it where necessary

10 years agoservice: change default stdout/stderr to syslog
Lennart Poettering [Tue, 30 Aug 2011 20:57:55 +0000 (22:57 +0200)]
service: change default stdout/stderr to syslog

10 years agostdout-bridge: rename logger to stdout-syslog-bridge to make it more descriptive
Lennart Poettering [Tue, 30 Aug 2011 20:42:49 +0000 (22:42 +0200)]
stdout-bridge: rename logger to stdout-syslog-bridge to make it more descriptive

10 years agoselinux: explicitly relabel /run/initramfs/root-fsck and friends
Lennart Poettering [Tue, 30 Aug 2011 16:49:17 +0000 (18:49 +0200)]
selinux: explicitly relabel /run/initramfs/root-fsck and friends

10 years agospecifier: drop misplaced assert()
Lennart Poettering [Tue, 30 Aug 2011 14:49:17 +0000 (16:49 +0200)]
specifier: drop misplaced assert()

10 years agoselinux: don't relabel /run/initramfs
Lennart Poettering [Mon, 29 Aug 2011 22:16:00 +0000 (00:16 +0200)]
selinux: don't relabel /run/initramfs

/run/initramfs usually contains the initrd so that we can jump back into
it on shutdown. It's usually relatively large and static data, hence we
should avoid relabelling of it. On my netbook this saves 6s. (6.6s
needed for relabelling /dev and /run goes down to 600ms -- still way too
much, but much better).

10 years agoselinux: retest selinux after we loaded the policy
Lennart Poettering [Mon, 29 Aug 2011 21:36:10 +0000 (23:36 +0200)]
selinux: retest selinux after we loaded the policy

10 years agologind: downgrade session create/remove log messages to LOG_DEBUG for
Lennart Poettering [Mon, 29 Aug 2011 20:08:19 +0000 (22:08 +0200)]
logind: downgrade session create/remove log messages to LOG_DEBUG for
cronjobs and similar which have no tty/X11 server assigned

10 years agountis: allow both console and plymouth ask password agents in boot transaction to...
Lennart Poettering [Mon, 29 Aug 2011 19:48:51 +0000 (21:48 +0200)]
untis: allow both console and plymouth ask password agents in boot transaction to allow dynamically disabled plymouth

10 years agoplymouth: don't enable ask-password agent of plymouth is disabled
Lennart Poettering [Mon, 29 Aug 2011 17:45:22 +0000 (19:45 +0200)]
plymouth: don't enable ask-password agent of plymouth is disabled

10 years agoexecute: fix bus serialization for commands
Lennart Poettering [Mon, 29 Aug 2011 17:44:52 +0000 (19:44 +0200)]
execute: fix bus serialization for commands