A watchdog notification may be handled after the watchdog timer was stopped
while stopping the service. As a result the timer is restarted and the
service may be restarted as well.
The watchdog timestamp is initially set during startup in
service_enter_start_post() and cleared when the timer is stopped. Therefore
it can be used as an indication if the timer should be reset.
For services without ExecStop= the state SERVICE_STOP is never entered. as
a result the watchdog timer is not stopped and the service is restarted (if
it is configuered to restart).
Stopping the watchdog timer for SERVICE_STOP_SIGTERM as well fixes this.
In the x32 ABI, syscall numbers start at 0x40000000. Mask that bit on
x32 for lookups in the syscall_names array and syscall_filter and ensure
that syscall.h is parsed correctly.
[zj: added SYSCALL_TO_INDEX, INDEX_TO_SYSCALL macros.]
This allows us to print simple performance data of all parts of the boot now:
- firmware
- boot loader
- kernel
- initrd
- userspace
This only works for bootloaders which support passing TSC data via EFI
variables. As of now that's only gummiboot.
In commit 246aa6d (core: add bus API and systemctl commands for altering
cgroup parameters during runtime), when rewriting unit_add_one_default_cgroup
to prefered style, the check of strduped b->controller was incorrectly
changed to check the containing structure. Correct it.
This makes sure that a service is not indefinitely restarted in a tight
loop if it fails before it is able to process its socket.
This corrects the breakage introduced with
8d1b002a2e. Shame on me.
We no longer allow early-boot init scripts, however in late boot the
syslog socket and local mounts are established anyway, so let's simplify
our dep graph a bit.
If $syslog doesn't resolve to syslog.target anymore there's no reason to
keep syslog.target around anymore. Let's remove it.
Note that many 3rd party service unit files order themselves after
syslog.target. These will be dangling dependencies now, which should be
unproblematic, however.
New file output.h with output flags and modes.
--full parameter also for cgls and loginctl.
Include 'all' parameter in flags (show_cgroup_by_path, show_cgroup,
show_cgroup_and_extra, show_cgroup_and_extra_by_spec).
get_process_cmdline with max_length == 0 will not ellipsize output.
Replace LINE_MAX with 0 in some calls of get_process_cmdline.
[zj: Default to --full when under pager for clgs.
Drop '-f' since it wasn't documented and didn't actually work.
Reindent a bit.
]
Systemd should not introduce any new facilities. Distributions which still
need to support their non-standard/legacy facilities should add them as
patches to their packaging.
The following facilities are no longer recognized:
$x-display-manager
$mail-transfer-agent
$mail-transport-agent
$mail-transfer-agent
$smtp
$null
This target is no longer available:
mail-transfer-agent.target
MNT_FORCE is honoured by NFS and FUSE and allows unmounting of the FS
even if consumers still use it. For our brute-force loop we rely on
EBUSY being reported as long as a file system is still used by a
loopback device or suchlike. Hence, drop MNT_FORCE to make EBUSY
reliable.
For all unit files foobar.service we will now read
foobar.service.d/*.conf, too. This may be used to override certain unit
settings without having to edit unit files directly.
This makes it really easy to change specific settings for services
without having to edit any unit file:
mkdir /etc/systemd/system/avahi-daemon.service.d/
echo -e '[Service]\nNice=99' > /etc/systemd/system/avahi-daemon.service.d/nice.conf
systemctl daemon-reload
For all other object mehtods there are already counterparts on the
manager object, as they help us reduce round-trips. So let's complete
this, and reduce complexity on the client side a bit.
As a side effect this also makes "systemctl snapshot" without arguments
work again.
As it turns out the bus properties for timer units wre really broken,
so let's clean this up for good and properly add calendar timer
serialization. We really should get that right before finalizing the
bus API documentation in the wiki...
Adding UNIT= to log lines allows them to be shown
in 'systemctl status' output, etc.
A new set of macros and functions is added. This allows for less
verbose notation than using log_struct() explicitly.
The set of logging functions is expanded to take a pair of arguments
(e.g. "UNIT=" and the RHS) which add an extra line to the structured
log entry. This can be used to add macros which add a different
identifier later on.
This simplifies the upstream system code quite a bit. If downstream distributions want to maintain compatibility with their old configuration files, they are welcome to do so, but need to maintain this as patches downstream. The burden needs to be on the distributions to maintain differences here. Our suggestion however is to just convert the old configuration files on upgrade, as multiple distributions already do.
Michael Olbrich
Zbigniew Jędrzejewski-Szmek
Eelco Dolstra
Lennart Poettering
Kay Sievers
Jonathan Callen
Mirco Tischler
Nestor Ovroy
Lukas Nykryn
Michał Bartoszkiewicz
Michal Sekletar
Thomas Hindoe Paaboel Andersen