Lennart Poettering [Thu, 3 May 2012 12:46:29 +0000 (14:46 +0200)]
service: default to Type=dbus if BusName= is specified
Lennart Poettering [Thu, 3 May 2012 12:42:49 +0000 (14:42 +0200)]
service: warn if a dbus name is specified but the service type is not dbus
Lennart Poettering [Thu, 3 May 2012 00:28:26 +0000 (02:28 +0200)]
vconsole: fix error messages
https://bugs.freedesktop.org/show_bug.cgi?id=49192
Lennart Poettering [Thu, 3 May 2012 00:25:49 +0000 (02:25 +0200)]
update TODO
Kay Sievers [Mon, 30 Apr 2012 13:30:57 +0000 (15:30 +0200)]
man: fix typo
Lennart Poettering [Mon, 30 Apr 2012 08:56:46 +0000 (10:56 +0200)]
units: make sure /var is writable before initializing random seed
Lennart Poettering [Mon, 30 Apr 2012 08:49:39 +0000 (10:49 +0200)]
TODO
Lennart Poettering [Sun, 29 Apr 2012 12:26:07 +0000 (14:26 +0200)]
unit: add new dependency type RequiresMountsFor=
RequiresMountsFor= is a shortcut for adding requires and after
dependencies to all mount units neeed for the specified paths.
This solves a couple of issues regarding dep loop cycles for encrypted
swap.
Lennart Poettering [Thu, 26 Apr 2012 18:32:26 +0000 (20:32 +0200)]
service: document new Type=idle
Lennart Poettering [Thu, 26 Apr 2012 18:32:07 +0000 (20:32 +0200)]
update TODO
Kay Sievers [Sun, 29 Apr 2012 22:46:38 +0000 (00:46 +0200)]
update TODO
Lucas De Marchi [Thu, 26 Apr 2012 16:36:02 +0000 (13:36 -0300)]
udev: use container_of() macro
Lucas De Marchi [Thu, 26 Apr 2012 16:36:01 +0000 (13:36 -0300)]
util: introduce container_of() macro
This macro comes from kernel and it's useful for unwrapping structs
inside another one. The generated code is actually the same to the one
where this logic is used in udev, but using this macro is much cleaner
and less error prone.
Kay Sievers [Thu, 26 Apr 2012 15:43:48 +0000 (17:43 +0200)]
libudev: prefix log macros with 'udev_'
Kay Sievers [Thu, 26 Apr 2012 00:03:31 +0000 (02:03 +0200)]
update TODO
Kay Sievers [Wed, 25 Apr 2012 23:49:15 +0000 (01:49 +0200)]
udev: use only oom_score_adj interface
Kay Sievers [Wed, 25 Apr 2012 23:48:24 +0000 (01:48 +0200)]
udev: docs - it's 'GUdev'
Michal Schmidt [Wed, 25 Apr 2012 09:58:27 +0000 (11:58 +0200)]
core: add NOP jobs, job type collapsing
Two of our current job types are special:
JOB_TRY_RESTART, JOB_RELOAD_OR_START.
They differ from other job types by being sensitive to the unit active state.
They perform some action when the unit is active and some other action
otherwise. This raises a question: when exactly should the unit state be
checked to make the decision?
Currently the unit state is checked when the job becomes runnable. It's more
sensible to check the state immediately when the job is added by the user.
When the user types "systemctl try-restart foo.service", he really intends
to restart the service if it's running right now. If it isn't running right
now, the restart is pointless.
Consider the example (from Bugzilla[1]):
sleep.service takes some time to start.
hello.service has After=sleep.service.
Both services get started. Two jobs will appear:
hello.service/start waiting
sleep.service/start running
Then someone runs "systemctl try-restart hello.service".
Currently the try-restart operation will block and wait for
sleep.service/start to complete.
The correct result is to complete the try-restart operation immediately
with success, because hello.service is not running. The two original
jobs must not be disturbed by this.
To fix this we introduce two new concepts:
- a new job type: JOB_NOP
A JOB_NOP job does not do anything to the unit. It does not pull in any
dependencies. It is always immediately runnable. When installed to a unit,
it sits in a special slot (u->nop_job) where it never conflicts with
the installed job (u->job) of a different type. It never merges with jobs
of other types, but it can merge into an already installed JOB_NOP job.
- "collapsing" of job types
When a job of one of the two special types is added, the state of the unit
is checked immediately and the job type changes:
JOB_TRY_RESTART -> JOB_RESTART or JOB_NOP
JOB_RELOAD_OR_START -> JOB_RELOAD or JOB_START
Should a job type JOB_RELOAD_OR_START appear later during job merging, it
collapses immediately afterwards.
Collapsing actually makes some things simpler, because there are now fewer
job types that are allowed in the transaction.
[1] Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=753586
Lennart Poettering [Wed, 25 Apr 2012 14:49:02 +0000 (16:49 +0200)]
timedated: introduce systemd-timedated-ntp.target which is controlled by timedated's NTP setting
We shouldn't hardcode the name of the NTP implementation in the
timedated mechanism, especially since Fedora currently switched from NTP
to chrony.
This patch introduces a new target that is enabled/disabled instead of
the actual NTP implementation. The various NTP implementations should
then add .wants/ symlinks to their services and BindTo back to the
target, so that their implementations are started/stopped jointly with
the target.
https://bugzilla.redhat.com/show_bug.cgi?id=815748
Lennart Poettering [Wed, 25 Apr 2012 13:11:20 +0000 (15:11 +0200)]
nspawn: add --read-only switch
Lennart Poettering [Wed, 25 Apr 2012 13:08:00 +0000 (15:08 +0200)]
nspawn: bind mount /etc/resolv.conf from the host by default
Lennart Poettering [Tue, 24 Apr 2012 14:56:06 +0000 (16:56 +0200)]
shutdown: don't complain if we cannot lock memory, to make container shutdowns clean
Lennart Poettering [Tue, 24 Apr 2012 14:42:42 +0000 (16:42 +0200)]
remount: consolidate remount-api-vfs and remount-rootfs into one
This has the advantage of removing a confusing warning by mount if the
root directory is not listed in fstab.
Lennart Poettering [Tue, 24 Apr 2012 12:59:49 +0000 (14:59 +0200)]
update TODO
Lennart Poettering [Tue, 24 Apr 2012 12:28:00 +0000 (14:28 +0200)]
service: introduce Type=idle and use it for gettys
Type=idle is much like Type=simple, however between the fork() and the
exec() in the child we wait until PID 1 informs us that no jobs are
left.
This is mostly a cosmetic fix to make gettys appear only after all boot
output is finished and complete.
Note that this does not impact the normal job logic as we do not delay
the completion of any jobs. We just delay the invocation of the actual
binary, and only for services that otherwise would be of Type=simple.
Lennart Poettering [Tue, 24 Apr 2012 11:53:31 +0000 (13:53 +0200)]
manager: drop MountAuto= and SwapAuto= options
The ability to set MountAuto=no and SwapAuto=no was useful during the
adoption phase of systemd, so that distributions could stick to their
classic mount scripts a bit longer. It is about time to get rid of it
now.
Lennart Poettering [Tue, 24 Apr 2012 11:14:27 +0000 (13:14 +0200)]
man: rework nspawn man page to suggest yum --installroot instead of mock
Lennart Poettering [Tue, 24 Apr 2012 11:12:29 +0000 (13:12 +0200)]
readahead: rather than checking for virtualization in the C code, use ConditionVirtualization= in the unit
Michal Schmidt [Tue, 24 Apr 2012 09:21:03 +0000 (11:21 +0200)]
transaction: cancel jobs non-recursively on isolate
Recursive cancellation of jobs would trigger OnFailure actions of
dependent jobs. This is not desirable when isolating.
Fixes https://bugzilla.redhat.com/show_bug.cgi?id=798328
Michal Schmidt [Sun, 22 Apr 2012 23:24:04 +0000 (01:24 +0200)]
job: serialize jobs properly
Jobs were not preserved correctly over a daemon-reload operation.
A systemctl process waiting for a job completion received a job removal
signal. The job itself changed its id. The job timeout started ticking all
over again.
This fixes the deficiencies.
Michal Schmidt [Sun, 22 Apr 2012 08:54:58 +0000 (10:54 +0200)]
transaction: abort does not need to use recursive deletion
Recursion is unnecessary, because we're deleting all transaction jobs
anyway. And the recursive deletion produces debug messages that are
pointless in transaction abort.
Michal Schmidt [Sun, 22 Apr 2012 00:09:04 +0000 (02:09 +0200)]
transaction: fix detection of cycles involving installed jobs
A transaction can be acyclic, but when it's added to installed jobs,
a cycle may result.
transaction_verify_order_one() attempts to detect these cases, but it
fails because the installed jobs often have the exact generation number
that makes them look as if they were walked already.
Fix it by resetting the generation numbers of all installed jobs before
detecting cycles.
An alternative fix could be to add the generation counter to the
Manager and use it instead of starting always from 1 in
transaction_activate(). But I prefer not having to worry about it
wrapping around.
Michal Schmidt [Sat, 21 Apr 2012 19:40:40 +0000 (21:40 +0200)]
transaction: improve readability
The functions looked complicated with the nested loops with breaks,
continues, and "while (again)".
Here using goto actually makes them easier to understand.
Also correcting the comment about redundant jobs.
Lennart Poettering [Sun, 22 Apr 2012 13:50:52 +0000 (15:50 +0200)]
shutdown: don't try to shut down DM devices in a container
Lennart Poettering [Sun, 22 Apr 2012 13:47:38 +0000 (15:47 +0200)]
hostname: if there's already a hostname set when PID 1 is invoked, don't complain
Lennart Poettering [Sun, 22 Apr 2012 13:39:26 +0000 (15:39 +0200)]
job: the status messages are proper sentences, hence end them with a full stop
Lennart Poettering [Sun, 22 Apr 2012 13:35:35 +0000 (15:35 +0200)]
units: remount file systems only if /etc/fstab actually exists
Lennart Poettering [Sun, 22 Apr 2012 13:33:43 +0000 (15:33 +0200)]
mount: don't fail if fstab doesn't exist
Lennart Poettering [Sun, 22 Apr 2012 13:22:52 +0000 (15:22 +0200)]
transaction: downgrade warnings about masked units
Lennart Poettering [Sun, 22 Apr 2012 13:22:27 +0000 (15:22 +0200)]
transaction: add missing emacs and license headers
Lennart Poettering [Sun, 22 Apr 2012 13:07:35 +0000 (15:07 +0200)]
machine-id: fix spelling
Lennart Poettering [Sun, 22 Apr 2012 12:48:46 +0000 (14:48 +0200)]
util: unify getenv() logic for other PID
Lennart Poettering [Sun, 22 Apr 2012 12:48:21 +0000 (14:48 +0200)]
nspawn: add --uuid= switch to allow setting the machine id for the container
Lennart Poettering [Sun, 22 Apr 2012 11:39:18 +0000 (13:39 +0200)]
units: don't try to load kernel modules if CAP_SYS_MODULE is missing
Lennart Poettering [Sun, 22 Apr 2012 11:38:46 +0000 (13:38 +0200)]
units: skip root fsck if the root directory is writable
Lennart Poettering [Sun, 22 Apr 2012 11:37:24 +0000 (13:37 +0200)]
nspawn: add -b switch to automatically look for an init binary
Lennart Poettering [Sun, 22 Apr 2012 00:45:39 +0000 (02:45 +0200)]
default to v102 everywhere, instead of vt100, to synchronize with agetty
Lennart Poettering [Sun, 22 Apr 2012 00:30:13 +0000 (02:30 +0200)]
login: assing /dev/console logins to seat0
Lennart Poettering [Sun, 22 Apr 2012 00:15:22 +0000 (02:15 +0200)]
container: spawn a getty instead of a sulogin in a container
Lennart Poettering [Sun, 22 Apr 2012 00:03:08 +0000 (02:03 +0200)]
loginctl: avoid segfault for kill-session and kill-user commands
As reported by Mantas Mikulėnas.
http://lists.freedesktop.org/archives/systemd-devel/2012-April/004993.html
Lennart Poettering [Sat, 21 Apr 2012 23:59:11 +0000 (01:59 +0200)]
util: fix tty_is_vc_resolve() in a container where /sys/class/tty/console/active is misleading
Lennart Poettering [Sat, 21 Apr 2012 23:01:54 +0000 (01:01 +0200)]
log: include syslog identifier in default log properties, to avoid comm truncation
Lennart Poettering [Sat, 21 Apr 2012 23:01:22 +0000 (01:01 +0200)]
nspawn: be more careful when initializing the hostname from the directory name
Lennart Poettering [Sat, 21 Apr 2012 23:00:57 +0000 (01:00 +0200)]
mount-setup: ignore common container bind mounts
Lennart Poettering [Sat, 21 Apr 2012 22:32:27 +0000 (00:32 +0200)]
update TODO
Lennart Poettering [Sat, 21 Apr 2012 22:32:13 +0000 (00:32 +0200)]
nspawn: make /dev/kmsg unavailable in the container, but allow access to /proc/kmsg
Lennart Poettering [Sat, 21 Apr 2012 13:47:38 +0000 (15:47 +0200)]
watchdog: fix default configuration fragment for watchdog
Kay Sievers [Fri, 20 Apr 2012 22:28:02 +0000 (00:28 +0200)]
man: doc-sync - properly delete no longer existing stuff on remote server
Michal Schmidt [Fri, 20 Apr 2012 12:44:00 +0000 (14:44 +0200)]
transaction: add starting requirements for JOB_RESTART
While having a Requires= dependency between units, the dependency is started
automatically on "systemctl start", but it's not started on "systemctl
restart".
JOB_RESTART jobs did not pull the dependencies for starting into the
transaction.
https://bugzilla.redhat.com/show_bug.cgi?id=802770
Note that the other bug noted in comment #2 has been fixed already by avoiding
the deletion of anchor jobs.
Michal Schmidt [Fri, 20 Apr 2012 10:28:31 +0000 (12:28 +0200)]
dbus-job: allow multiple bus clients
Merging of jobs can result in more than one client being interested in a job.
Michal Schmidt [Fri, 20 Apr 2012 08:02:05 +0000 (10:02 +0200)]
transaction: remove checks for installed
Transactions cannot contain installed jobs anymore. Remove the now pointless
checks.
Michal Schmidt [Fri, 20 Apr 2012 07:38:43 +0000 (09:38 +0200)]
transaction: rework merging with installed jobs
Previously transactions could reference installed jobs. It made some issues
difficult to fix.
This sets new rules for jobs:
A job cannot be both a member of a transaction and installed. When jobs are
created, they are linked to a transaction. The whole transaction is constructed
(with merging of jobs within, etc.). When it's complete, all the jobs are
unlinked from it one by one and let to install themselves. It is during the
installation when merging with previously installed jobs (from older
transactions) is contemplated.
Merging with installed jobs has different rules than merging within a
transaction:
- An installed conflicting job gets cancelled. It cannot be simply deleted,
because someone might be waiting for its completion on DBus.
- An installed, but still waiting, job can be safely merged into.
- An installed and running job can be tricky. For some job types it is safe to
just merge. For the other types we merge anyway, but put the job back into
JOB_WAITING to allow it to run again. This may be suboptimal, but it is not
currently possible to have more than one installed job for a unit.
Note this also fixes a bug where the anchor job could be deleted during merging
within the transaction.
Michal Schmidt [Fri, 20 Apr 2012 00:04:01 +0000 (02:04 +0200)]
job: separate job_install()
Let the jobs install themselves.
Michal Schmidt [Thu, 19 Apr 2012 21:56:26 +0000 (23:56 +0200)]
transaction: remove a couple of asserts
We already asserted these facts in the previous loop.
Michal Schmidt [Fri, 20 Apr 2012 00:11:14 +0000 (02:11 +0200)]
transaction: remove the anchor link
tr->anchor_job is sufficient.
Michal Schmidt [Fri, 20 Apr 2012 00:48:24 +0000 (02:48 +0200)]
transaction: avoid garbage collecting the anchor job
Make sure the anchor job is never considered garbage, even if it has no links
leading to it (this will be allowed in the next patch).
Michal Schmidt [Thu, 19 Apr 2012 23:20:14 +0000 (01:20 +0200)]
transaction: simplify transaction_find_jobs_that_matter_to_anchor()
Michal Schmidt [Fri, 20 Apr 2012 08:33:37 +0000 (10:33 +0200)]
transaction: change the linking of isolate jobs to the anchor
When isolating, the JOB_STOP jobs have no parent job, so they are all peers
of the real anchor job. This is a bit odd.
Link them from the anchor job.
Michal Schmidt [Thu, 19 Apr 2012 22:33:26 +0000 (00:33 +0200)]
transaction: maintain anchor_job
Track which job is the anchor in the transaction.
Michal Schmidt [Wed, 18 Apr 2012 16:15:49 +0000 (18:15 +0200)]
transaction: do not add installed jobs to the transaction
Do not attempt to optimize away the job creation by refering to installed jobs.
We do not want to disturb installed jobs until commiting the transaction.
(A later patch to job merging will make the separation of transaction jobs and
installed jobs complete.)
Michal Schmidt [Wed, 18 Apr 2012 13:21:24 +0000 (15:21 +0200)]
job: jobs shouldn't need to know about transaction anchors
Let the transactions maintain their own anchor links.
Michal Schmidt [Tue, 17 Apr 2012 23:39:20 +0000 (01:39 +0200)]
job: job_new() can find the manager from the unit
Michal Schmidt [Thu, 19 Apr 2012 21:54:11 +0000 (23:54 +0200)]
manager: split transaction.[ch]
manager.c takes care of the main loop, unit management, signal handling, ...
transaction.c computes transactions.
After split:
manager.c: 65 KB
transaction.c: 40 KB
Michal Schmidt [Fri, 20 Apr 2012 08:22:07 +0000 (10:22 +0200)]
manager: Transaction as an object
This makes it obvious that transactions are short-lived. They are created in
manager_add_job() and destroyed after the application of jobs.
It also prepares for a split of the transaction code to a new source.
Michal Schmidt [Fri, 20 Apr 2012 08:21:37 +0000 (10:21 +0200)]
job: job_uninstall()
Split the uninstallation of the job from job_free() into a separate function.
Adjust the callers.
job_free() now only works on unlinked and uninstalled jobs. This enforces clear
thinking about job lifetimes.
Michal Schmidt [Thu, 19 Apr 2012 21:23:04 +0000 (23:23 +0200)]
manager: simplify transaction_abort()
This is equivalent.
Michal Schmidt [Thu, 19 Apr 2012 21:20:34 +0000 (23:20 +0200)]
job: allow job_free() only on already unlinked jobs
job_free() is IMO too helpful when it unlinks the job from the transaction.
The callers should ensure the job is already unlinked before freeing.
The added assertions check if anyone gets it wrong.
Michal Schmidt [Tue, 17 Apr 2012 20:54:23 +0000 (22:54 +0200)]
manager: fix comment
Michal Schmidt [Tue, 17 Apr 2012 20:53:35 +0000 (22:53 +0200)]
tmpfiles: fix error message
Kay Sievers [Fri, 20 Apr 2012 06:22:18 +0000 (08:22 +0200)]
docs: remove duplicated install hook
Lennart Poettering [Fri, 20 Apr 2012 10:39:36 +0000 (12:39 +0200)]
log: fix LOG_TARGET_JOURNAL_OR_KMSG
Lennart Poettering [Fri, 20 Apr 2012 10:39:09 +0000 (12:39 +0200)]
mount-setup: don't log with LOG_ERROR if a mount that doesn't matter fails
Lennart Poettering [Fri, 20 Apr 2012 10:37:57 +0000 (12:37 +0200)]
main: log to the journal in container mode, by default
Kay Sievers [Fri, 20 Apr 2012 01:25:36 +0000 (03:25 +0200)]
docs: hook-up gtk-doc to 'make check'
Kay Sievers [Thu, 19 Apr 2012 17:00:35 +0000 (19:00 +0200)]
docs: rebase html documentation for online and local use
Javier Jardón [Thu, 19 Apr 2012 14:06:25 +0000 (15:06 +0100)]
docs: do not generate tmpl files
Its not needed as all the documentation is inline source code,
not in separate *.tmpl files
Sven Anders [Thu, 19 Apr 2012 10:03:28 +0000 (12:03 +0200)]
fix typo in src/shared/install.c
Kay Sievers [Wed, 18 Apr 2012 18:31:16 +0000 (20:31 +0200)]
update TODO
Kay Sievers [Wed, 18 Apr 2012 14:29:43 +0000 (16:29 +0200)]
udev: units - reference systemd-* units
Kay Sievers [Wed, 18 Apr 2012 14:13:37 +0000 (16:13 +0200)]
build-sys: move dev-setup to label.la
Kay Sievers [Wed, 18 Apr 2012 11:37:45 +0000 (13:37 +0200)]
remove MS_* which can not be combined with current kernel code
MS_BIND|MS_MOVE can not be combined:
do_mount()
else if (flags & MS_BIND)
do_loopback(&path, dev_name, flags & MS_REC);
[...]
else if (flags & MS_MOVE)
do_move_mount(&path, dev_name);
MS_REMOUNT|MS_UNBINDABLE can not be combined:
do_mount()
if (flags & MS_REMOUNT)
do_remount(&path, flags & ~MS_REMOUNT, mnt_flags, data_page);
[...]
else if (flags & (MS_SHARED | MS_PRIVATE | MS_SLAVE | MS_UNBINDABLE))
do_change_type(&path, flags);
Kay Sievers [Tue, 17 Apr 2012 20:25:24 +0000 (22:25 +0200)]
udev: unify /dev static symlink setup
Kay Sievers [Tue, 17 Apr 2012 17:02:23 +0000 (19:02 +0200)]
udev: docs - updated index page to match common gtk style
Kay Sievers [Tue, 17 Apr 2012 16:47:49 +0000 (18:47 +0200)]
update TODO
Kay Sievers [Tue, 17 Apr 2012 16:42:09 +0000 (18:42 +0200)]
silence a bunch of gcc warnings
Kay Sievers [Tue, 17 Apr 2012 15:03:38 +0000 (17:03 +0200)]
build-sys: silence gtk-doc check, remove gtk-doc options from configure and distcheck
Javier Jardón [Tue, 17 Apr 2012 14:17:27 +0000 (15:17 +0100)]
build-sys: make gtk-doc dependency optional
Kay Sievers [Tue, 17 Apr 2012 14:54:45 +0000 (16:54 +0200)]
update TODO
Kay Sievers [Tue, 17 Apr 2012 14:05:03 +0000 (16:05 +0200)]
selinux: unify systemd and udev code
Kay Sievers [Tue, 17 Apr 2012 12:43:28 +0000 (14:43 +0200)]
mkdir: do not use alloca() in a loop