chiark / gitweb /
Lennart Poettering [Mon, 14 Mar 2011 04:37:14 +0000 (05:37 +0100)]
socket: use 777 as default mode for sockets
Lennart Poettering [Mon, 14 Mar 2011 04:36:43 +0000 (05:36 +0100)]
nspawn: we don't want a network namespace
Lennart Poettering [Mon, 14 Mar 2011 03:48:14 +0000 (04:48 +0100)]
main: log to the console in a container
Lennart Poettering [Mon, 14 Mar 2011 03:08:12 +0000 (04:08 +0100)]
main: don't parse /proc/cmdline in containers
Lennart Poettering [Mon, 14 Mar 2011 03:07:52 +0000 (04:07 +0100)]
util: add detect_container()
Lennart Poettering [Mon, 14 Mar 2011 02:28:16 +0000 (03:28 +0100)]
nspawn: reset environment and load login shell
Lennart Poettering [Mon, 14 Mar 2011 02:28:00 +0000 (03:28 +0100)]
nspawn: reset umask if needed
Lennart Poettering [Mon, 14 Mar 2011 02:27:28 +0000 (03:27 +0100)]
nspawn: define MS_MOVE manually if needed
Lennart Poettering [Mon, 14 Mar 2011 02:12:04 +0000 (03:12 +0100)]
nspawn: improve exit warning
Lennart Poettering [Mon, 14 Mar 2011 02:10:09 +0000 (03:10 +0100)]
core: move abstract namespace sockets to /dev/.run
Now that we have /dev/.run there's no need to use abstract namespace
sockets. So, let's move things to /dev/.run, to make things more easily
discoverable and improve compat with chroot() and fs namespacing.
Lennart Poettering [Mon, 14 Mar 2011 01:40:36 +0000 (02:40 +0100)]
nspawn: add simple chroot(1) like tool to execute commands in a namespace container
Lennart Poettering [Mon, 14 Mar 2011 01:36:00 +0000 (02:36 +0100)]
util: detect CLONE_NEWPID namespaces, and cache results
Lennart Poettering [Mon, 14 Mar 2011 01:34:18 +0000 (02:34 +0100)]
git: ignore generated policy file
Lennart Poettering [Mon, 14 Mar 2011 01:33:51 +0000 (02:33 +0100)]
util: return exit status in wait_for_terminate_and_warn()
Lennart Poettering [Mon, 14 Mar 2011 01:33:23 +0000 (02:33 +0100)]
util: properly identify pty devices by their major
Kay Sievers [Sat, 12 Mar 2011 13:31:44 +0000 (14:31 +0100)]
build-sys: add org.freedesktop.systemd1.policy to CLEANFILES
ERROR: files left in build directory after distclean:
./src/org.freedesktop.systemd1.policy
make[1]: *** [distcleancheck] Error 1
Lennart Poettering [Sat, 12 Mar 2011 00:15:30 +0000 (01:15 +0100)]
polkit: autogenerate polkit policy with correct paths
Lennart Poettering [Sat, 12 Mar 2011 00:03:13 +0000 (01:03 +0100)]
systemctl: support remote and privileged systemctl access via SSH and pkexec
This adds support for executing systemctl operations remotely or as
privileged user while still running systemctl itself unprivileged and
locally.
This currently requires a D-Bus patch to work properly.
https://bugs.freedesktop.org/show_bug.cgi?id=35230
Kay Sievers [Fri, 11 Mar 2011 14:55:20 +0000 (15:55 +0100)]
gnome-ask-password-agent.vala: check for VALA_0_12
<juergbi> VALA_0_X is defined up to and including the current version
<juergbi> so VALA_0_10 is defined in 0.10 and 0.11 and 0.12
<juergbi> VALA_0_12 is defined in 0.11 and 0.12 (and later versions)
<juergbi> reverse the branches and use VALA_0_12 as conditional
<kay> juergbi: changing api but keep the defines?
<kay> how do we support vala 0.14 then?
<kay> you keep that model with the old defines?
<juergbi> VALA_0_10 corresponds to valaversion >= 0.10
Kay Sievers [Fri, 11 Mar 2011 14:41:37 +0000 (15:41 +0100)]
gnome-ask-password-agent: check for vala 0.10 instead of libnotify
Kay Sievers [Fri, 11 Mar 2011 13:09:10 +0000 (14:09 +0100)]
gnome-ask-password-agent: restore removed libnotify0.6 support
Lennart Poettering [Fri, 11 Mar 2011 00:51:45 +0000 (01:51 +0100)]
gnome-ask-password-agent: fix path to watch
Lennart Poettering [Fri, 11 Mar 2011 00:06:53 +0000 (01:06 +0100)]
units: move the last flag files to /dev/.run
Lennart Poettering [Thu, 10 Mar 2011 23:52:13 +0000 (00:52 +0100)]
util: close all fds before freezing execution
Lennart Poettering [Thu, 10 Mar 2011 23:45:06 +0000 (00:45 +0100)]
dbus: timeout connection setup
Lennart Poettering [Thu, 10 Mar 2011 22:01:42 +0000 (23:01 +0100)]
main: properly handle -b boot option
Andrey Borzenkov [Thu, 10 Mar 2011 14:39:02 +0000 (17:39 +0300)]
pam: do not leak file descriptor if flock fails
If flock fails, fd is not returned to caller so it cannot clean up.
Andrey Borzenkov [Thu, 10 Mar 2011 17:31:18 +0000 (20:31 +0300)]
man: trivial typo in systemd(1)
Lennart Poettering [Wed, 9 Mar 2011 22:59:27 +0000 (23:59 +0100)]
main: refuse system to be started in a chroot
Lennart Poettering [Wed, 9 Mar 2011 22:59:07 +0000 (23:59 +0100)]
main: don't check if /usr really is a mount point, since it is fine if it is passed pre-mounted to us from the initrd
Lennart Poettering [Wed, 9 Mar 2011 22:58:17 +0000 (23:58 +0100)]
condition: take a timestamp and store last result of conditions
Lennart Poettering [Wed, 9 Mar 2011 21:45:47 +0000 (22:45 +0100)]
dev: use /dev/.run/systemd as runtime directory, instead of /dev/.systemd
Lennart Poettering [Wed, 9 Mar 2011 21:13:24 +0000 (22:13 +0100)]
machine-id: move machine-id-setup to /sbin
Lennart Poettering [Wed, 9 Mar 2011 19:15:44 +0000 (20:15 +0100)]
pkconfig: export full search path as .pc variable
Lennart Poettering [Wed, 9 Mar 2011 19:12:30 +0000 (20:12 +0100)]
selinux: bump up error level when in non-enforcing mode
Andrey Borzenkov [Wed, 9 Mar 2011 19:03:29 +0000 (20:03 +0100)]
dbus: fix dbus assert due to uninitialized error
Add missing dbus_error_init() in UNKNOWN_OBJECT case. Fixes assertion
systemd[1]: Caught <ABRT>, dumped core as pid 6256.
systemd[1]: Freezing execution.
Core was generated by `/bin/systemd systemd.unit=graphical.target'.
Program terminated with signal 6, Aborted.
#0 0x00007f8966ec81db in raise () from /lib64/libpthread.so.0
(gdb) bt
#0 0x00007f8966ec81db in raise () from /lib64/libpthread.so.0
#1 0x000000000040823b in crash (sig=6) at src/main.c:120
#2 <signal handler called>
#3 0x00007f896613c075 in raise () from /lib64/libc.so.6
#4 0x00007f896613d806 in abort () from /lib64/libc.so.6
#5 0x00007f89672ac8a5 in _dbus_abort () at dbus-sysdeps.c:94
#6 0x00007f89672a37b5 in _dbus_warn_check_failed (
format=0x7f89672b35d8 "arguments to %s() were incorrect, assertion
\"%s\" failed in file %s line %d.\nThis is normally a bug in some
application using the D-Bus library.\n") at dbus-internals.c:289
#7 0x0000000000441500 in bus_unit_message_handler (connection=0x21b6090,
message=0x21b6760, data=0x1f3e870) at src/dbus-unit.c:572
#8 0x00007f8967299f11 in _dbus_object_tree_dispatch_and_unlock (
tree=0x219d660, message=0x21b6760) at dbus-object-tree.c:858
#9 0x00007f896728bca2 in dbus_connection_dispatch (connection=0x21b6090)
at dbus-connection.c:4688
#10 0x000000000043befa in bus_dispatch (m=0x1f3e870) at src/dbus.c:547
#11 0x000000000041056d in manager_loop (m=0x1f3e870) at src/manager.c:2344
#12 0x0000000000409515 in main (argc=<value optimized out>,
argv=<value optimized out>) at src/main.c:1229
Lennart Poettering [Wed, 9 Mar 2011 19:01:53 +0000 (20:01 +0100)]
dbus: properly generate UnknownInterface, UnknownProperty and PropertyReadOnly errors
Lennart Poettering [Wed, 9 Mar 2011 18:48:02 +0000 (19:48 +0100)]
drop unnecessary suffix NULs as gcc adds them anyway
Lennart Poettering [Wed, 9 Mar 2011 01:51:45 +0000 (02:51 +0100)]
mount: use /dev/.run as an early boot alias for /var/run
During early boot, mount a tmpfs to /dev/.run and then bind mount it to
/var/run as soon as /var is available.
This makes it possible for programs involved in early boot to put
runtime data in /dev/.run which later on will show up in /var/run like
any other.
This can be used to solve the early-boot D-Bus problem: D-Bus may start
up with its socket bound to /dev/.run/dbus/system_bus_socket and after
/var it will also be available under the traditional name
/var/run/dbus/system_bus_socket.
This also is intended to be used as a better place for systemd, mount,
mdadm, blkid, plymouth, bootchart and dracut runtime data, which is
currently stored in various places in /dev/.xxx.
Andrey Borzenkov [Tue, 8 Mar 2011 18:47:29 +0000 (19:47 +0100)]
build-sys: add Mandriva distribution support
This merges several separate patches that I carry as part of
Mandriva systemd RPM. They touch those parts that are very
unlikely to be changed in near future and do not impose any
functionality change for systemd core. I also think it is
useful for troubleshooting to have real distribution name in
system logs, espicially when someone reports problem upstream.
The patch looks bigger than sum of replaced patches because
- previous patches were applied on top of distro=fedora, now
I need to add all those bits for distro=mandriva as well
- part of patch was done as spec file magic, but it seems more
logical to ship all these bits together
Lennart Poettering [Tue, 8 Mar 2011 18:41:19 +0000 (19:41 +0100)]
build-sys: prepare v20
Lennart Poettering [Tue, 8 Mar 2011 17:51:20 +0000 (18:51 +0100)]
service: use uppercase SYSV prefix to make it easier to recognize
Lennart Poettering [Tue, 8 Mar 2011 17:41:23 +0000 (18:41 +0100)]
service: prefix description with LSB only if script has LSB header, use 'SysV:' otherwise
Lennart Poettering [Tue, 8 Mar 2011 02:24:42 +0000 (03:24 +0100)]
unit: don't accidently create ordering links to targets when default deps are off for either target and unit
Lennart Poettering [Tue, 8 Mar 2011 02:08:15 +0000 (03:08 +0100)]
mount: support less cumbersome x-systemd-xxx mount options
Lennart Poettering [Tue, 8 Mar 2011 02:04:47 +0000 (03:04 +0100)]
unit: distuingish mandatory from triggering conditions
Lennart Poettering [Tue, 8 Mar 2011 01:27:43 +0000 (02:27 +0100)]
dbus: return DBUS_ERROR_UNKNOWN_OBJECT when an object is unknown
Lennart Poettering [Tue, 8 Mar 2011 00:47:30 +0000 (01:47 +0100)]
systemctl: when forwarding is-enabled to chkconfig hardcode runlevel 3
Lennart Poettering [Tue, 8 Mar 2011 00:44:19 +0000 (01:44 +0100)]
job: introduce new job result code 'skipped' to use when pre conditions of job did not apply
https://bugzilla.redhat.com/show_bug.cgi?id=682245
https://bugzilla.redhat.com/show_bug.cgi?id=635780
Lennart Poettering [Tue, 8 Mar 2011 00:43:41 +0000 (01:43 +0100)]
job: convert job type as early as we can, to simplify things a bit
Zbigniew Jędrzejewski-Szmek [Mon, 7 Mar 2011 18:23:04 +0000 (19:23 +0100)]
Keep emacs configuration in one configuration file.
On 07.03.2011 19:04, Lennart Poettering wrote:
>
> Oh, I had assumed that .vimrc trick would allow project-wide vim
> modelines without having to edit each and every single file.
>
> Currently, every file does contain an emacs modeline at the top. It
> would be fair I guess to add a vim modeline to all those files too, even
> though it's not necessarily pretty.
Hi,
maybe it makes sense to go in the opposite direction: add .dir-locals.el
in the top directory. The settings are actually identical in all .[ch]
files.
Lennart Poettering [Mon, 7 Mar 2011 20:56:09 +0000 (21:56 +0100)]
syslog: make sure the kmsg bridge is always pulled in and never terminated automatically
Since we want to replace the bridge dynamically by the real syslog
implementation such as rsyslog we need to make sure that the the bridge
stays running right to the moment rsyslog is up so that we process
messages enqueued by other processes started before rsyslog, so that
those clients don't stay stuck.
Andrey Borzenkov [Mon, 7 Mar 2011 18:25:15 +0000 (19:25 +0100)]
add .vimrc to .gitignore
Lennart Poettering [Mon, 7 Mar 2011 16:38:15 +0000 (17:38 +0100)]
mount: make /media a tmpfs
Lennart Poettering [Mon, 7 Mar 2011 01:17:23 +0000 (02:17 +0100)]
update TODO
Andrey Borzenkov [Sun, 6 Mar 2011 16:17:02 +0000 (19:17 +0300)]
add org.freedesktop.DBus.Properies.Set method
The patch adds framework for processing Set requests on properties
and as first consumers allows setting of LogTarget and LogLevel on
interface org.freedesktop.systemd1.Manager of org.freedesktop.systemd1.
The code should be generic enough to allow processing of any
property on other objects and interfaces as well.
No systemctl visible interface is added so far. The LogTarget and
LogLevel are meant for debugging in the first place; user interface
can be added if other usage emerge.
Set on systemwide systemd is restricted to root; I am not sure
how session level access is controlled.
Lennart Poettering [Fri, 4 Mar 2011 13:54:15 +0000 (16:54 +0300)]
service: drop redundant check in sysv_translate_facility
We check for _r with an assert in the function header anyway, no need to
check it again.
Lennart Poettering [Fri, 4 Mar 2011 23:30:54 +0000 (00:30 +0100)]
update TODO
Lennart Poettering [Fri, 4 Mar 2011 21:50:56 +0000 (22:50 +0100)]
machine-id: generate /etc/machine-id 0444 by default
Lennart Poettering [Fri, 4 Mar 2011 21:28:13 +0000 (22:28 +0100)]
machine-id: typo fix
Lennart Poettering [Fri, 4 Mar 2011 20:53:19 +0000 (21:53 +0100)]
main: introduce /etc/machine-id
This is supposed to play the same roles /var/lib/dbus/machine-id,
however fixes a couple of problems:
- It is available during early boot since it is stored in /etc
- Removes the ID from the D-Bus context and moves it into a system
context, thus hopefully lowering hesitation by people to use it.
- It is generated at installation time. If the file is empty at boot
time it will be mounted over with a randomly generated ID, which is
not saved to disk. This is useful to support state-less machines with
no transient or writable /etc configuration.
Lennart Poettering [Fri, 4 Mar 2011 18:48:38 +0000 (19:48 +0100)]
systemctl: fix exit code when directing is-enabled to chkconfig
Lennart Poettering [Fri, 4 Mar 2011 18:48:12 +0000 (19:48 +0100)]
main: minor optimization
Lennart Poettering [Fri, 4 Mar 2011 18:47:43 +0000 (19:47 +0100)]
dbus: add 'Tainted' property to Manager object
Lennart Poettering [Fri, 4 Mar 2011 17:34:56 +0000 (18:34 +0100)]
dbus: expose distribution name and features on manager object as properties
Lennart Poettering [Fri, 4 Mar 2011 13:13:47 +0000 (14:13 +0100)]
man: document changed EnvironmentFile= behaviour
Lennart Poettering [Fri, 4 Mar 2011 04:07:01 +0000 (05:07 +0100)]
main: add link to wiki page with longer explanation of the /usr madness
Lennart Poettering [Fri, 4 Mar 2011 02:58:52 +0000 (03:58 +0100)]
s/seperate/separate/
Lennart Poettering [Fri, 4 Mar 2011 02:44:43 +0000 (03:44 +0100)]
execute: load environment files at time of execution, not when we load the service configuration
https://bugzilla.redhat.com/show_bug.cgi?id=661282
Lennart Poettering [Fri, 4 Mar 2011 00:49:37 +0000 (01:49 +0100)]
path: after installing inotify watches, recheck file again to fix race
Lennart Poettering [Thu, 3 Mar 2011 23:22:00 +0000 (00:22 +0100)]
path: don't use IN_ATTRIB on parent dirs when watching a file, since those cannot be removed without emptying the dir first anyway and we need IN_ATTRIB only to detect the link count dropping to 0
Lennart Poettering [Thu, 3 Mar 2011 22:55:30 +0000 (23:55 +0100)]
kill: always send SIGCONT after SIGTERM
When we kill a process to terminate it make sure to send SIGCONT to
ensure it is unpaused and processes the signal.
Lennart Poettering [Thu, 3 Mar 2011 22:03:26 +0000 (23:03 +0100)]
readahead: disable collector automatically on read-only media
Lennart Poettering [Thu, 3 Mar 2011 21:17:09 +0000 (22:17 +0100)]
sysctl: use scandir() instead of nftw() to guarantee systematic ordering
Lennart Poettering [Thu, 3 Mar 2011 15:29:50 +0000 (16:29 +0100)]
dirent: support DT_UNKNOWN where necessary for compat with reiserfs
Lennart Poettering [Tue, 1 Mar 2011 22:44:26 +0000 (23:44 +0100)]
README: refer to profusion if people ask for consulting/engineering services
Lennart Poettering [Tue, 1 Mar 2011 21:19:08 +0000 (22:19 +0100)]
systemctl: add condreload alias for compat with ALTLinux
Lennart Poettering [Tue, 1 Mar 2011 21:18:49 +0000 (22:18 +0100)]
man: fix chkconfig syntax to use --level 3
Lennart Poettering [Tue, 1 Mar 2011 00:33:38 +0000 (01:33 +0100)]
build-sys: prepare v19
Lennart Poettering [Tue, 1 Mar 2011 00:02:35 +0000 (01:02 +0100)]
update TODO
Lennart Poettering [Mon, 28 Feb 2011 22:24:33 +0000 (23:24 +0100)]
ask-password: make a confused gcc shut up
Harald Hoyer [Mon, 21 Feb 2011 14:32:18 +0000 (15:32 +0100)]
ratelimit: removed n_printed
Removed n_printed and renamed n_printed to num.
This is not a logging rate limiter anymore.
Harald Hoyer [Mon, 21 Feb 2011 14:32:17 +0000 (15:32 +0100)]
Spelling Corrections
Just some lame spelling corrections with no functionality.
Zbigniew Jędrzejewski-Szmek [Mon, 28 Feb 2011 10:25:53 +0000 (11:25 +0100)]
systemctl: always null-terminate the password
Alexey Shabalin [Mon, 28 Feb 2011 13:02:27 +0000 (16:02 +0300)]
build-sys: update ALTLinux support
Please apply patch for update ALTLinux support.
Lennart Poettering [Mon, 28 Feb 2011 21:04:53 +0000 (22:04 +0100)]
udev: don't ignore non-encrypted block devices with no superblock
They might be encrypted disks with no LUKS header.
https://bugzilla.redhat.com/show_bug.cgi?id=679842
Lennart Poettering [Mon, 28 Feb 2011 20:47:35 +0000 (21:47 +0100)]
udev: expose ttyUSB devices too
Lennart Poettering [Mon, 28 Feb 2011 20:46:34 +0000 (21:46 +0100)]
udev: mark hvc devices for exposure in systemd
Lennart Poettering [Fri, 25 Feb 2011 01:56:27 +0000 (02:56 +0100)]
cryptsetup: add a terse help
Lennart Poettering [Fri, 25 Feb 2011 01:52:07 +0000 (02:52 +0100)]
agent: don't print warnings if a password was removed or timed out
Lennart Poettering [Fri, 25 Feb 2011 01:51:48 +0000 (02:51 +0100)]
systemctl: shutdown agent explicitly so that it can reset the tty properly
Lennart Poettering [Fri, 25 Feb 2011 00:49:10 +0000 (01:49 +0100)]
service: never clean up a service that still has a process in it
Lennart Poettering [Fri, 25 Feb 2011 00:47:31 +0000 (01:47 +0100)]
label: udev might be making changes in /dev while we iterate through it
Also, there are most likely dead symlinks in there, so let's ignore
ENOENT when we relabel.
https://bugzilla.redhat.com/show_bug.cgi?id=680169
Lennart Poettering [Thu, 24 Feb 2011 02:24:51 +0000 (03:24 +0100)]
systemctl: properly handle job results
Lennart Poettering [Thu, 24 Feb 2011 02:24:23 +0000 (03:24 +0100)]
job: also trigger on-failure dependencies when jobs faile due to dependencies, timeout
Lennart Poettering [Thu, 24 Feb 2011 02:23:43 +0000 (03:23 +0100)]
job: don't access j->type when j is already freed
Lennart Poettering [Thu, 24 Feb 2011 02:23:14 +0000 (03:23 +0100)]
job: when cancelling jobs, make sure to propagate this properly to depending jobs
Lennart Poettering [Thu, 24 Feb 2011 01:41:34 +0000 (02:41 +0100)]
job: start job timeout on enqueuing not when we start to process a job
This practically reverses
23e1e0c4a3bc93fb841ac9575953ed5d0ef5282d.
This makes it easier to timeout individual jobs so that this timeout can
be relied on, since blocked dependencies will not influence the timeout.
Lennart Poettering [Thu, 24 Feb 2011 01:38:19 +0000 (02:38 +0100)]
unit: increase default timeout to 3min
Lennart Poettering [Thu, 24 Feb 2011 01:38:02 +0000 (02:38 +0100)]
logger: leave the logger longer around and allow it do handle more connections