Instead of returning an enum of return codes, make them return error
codes like kdbus does internally.
Also, document this behaviour so that clients can stick to it.
(Also rework bus-control.c to always have to functions for dbus1 vs.
kernel implementation of the various calls.)
Since we want to retain the ability to break kernel ←→ userspace ABI
after the next release, let's not make use by default of kdbus, so that
people with future kernels will not suddenly break with current systemd
versions.
kdbus support is left in all builds but must now be explicitly requested
at runtime (for example via setting $DBUS_SESSION_BUS). Via a configure
switch the old behaviour can be restored. In fact, we change autogen.sh
to do this, so that git builds (which run autogen.sh) get kdbus by
default, but tarball builds (which ue the configure defaults) do not get
it, and hence this stays out of the distros by default.
This reverts commit adcf4c81c5.
We have a better solution for the problem of making two processes run in
the same namespace, and --listener is not needed hence and should be
dropped.
Conflicts:
man/systemd-socket-proxyd.xml
* library support for setns() system call was added to glibc
version 2.14 (setns() call is use in src/machine/machinectl.c
and src/libsystemd-bus-container.c)
* utf8 validation call are already exported (via sd-utf8.c file) -
commit - 369c583b3f
Message handler callbacks can be simplified drastically if the
dispatcher automatically replies to method calls if errors are returned.
Thus: add an sd_bus_error argument to all message handlers. When we
dispatch a message handler and it returns negative or a set sd_bus_error
we send this as message error back to the client. This means errors
returned by handlers by default are given back to clients instead of
rippling all the way up to the event loop, which is desirable to make
things robust.
As a side-effect we can now easily turn the SELinux checks into normal
function calls, since the method call dispatcher will generate the right
error replies automatically now.
Also, make sure we always pass the error structure to all property and
method handlers as last argument to follow the usual style of passing
variables for return values as last argument.
This patch converts PID 1 to libsystemd-bus and thus drops the
dependency on libdbus. The only remaining code using libdbus is a test
case that validates our bus marshalling against libdbus' marshalling,
and this dependency can be turned off.
This patch also adds a couple of things to libsystem-bus, that are
necessary to make the port work:
- Synthesizing of "Disconnected" messages when bus connections are
severed.
- Support for attaching multiple vtables for the same interface on the
same path.
This patch also fixes the SetDefaultTarget() and GetDefaultTarget() bus
calls which used an inappropriate signature.
As a side effect we will now generate PropertiesChanged messages which
carry property contents, rather than just invalidation information.
When a service exits succesfully and has RemainAfterExit set, its hold
on the console (in m->n_on_console) wasn't released since the unit state
didn't change.
This tool applies hardware specific settings to network devices before they
are announced via libudev.
Settings that will probably eventually be supported are MTU, Speed,
DuplexMode, WakeOnLan, MACAddress, MACAddressPolicy (e.g., 'hardware',
'synthetic' or 'random'), Name and NamePolicy (replacing our current
interface naming logic). This patch only introduces support for
Description, as a proof of concept.
Some of these settings may later be overriden by a network management
daemon/script. However, these tools should always listen and wait on libudev
before touching a device (listening on netlink is not enough). This is no
different from how things used to be, as we always supported changing the
network interface name from udev rules, which does not work if someone
has already started using it.
The tool is configured by .link files in /etc/net/links/ (with the usual
overriding logic in /run and /lib). The first (in lexicographical order)
matching .link file is applied to a given device, and all others are ignored.
The .link files contain a [Match] section with (currently) the keys
MACAddress, Driver, Type (see DEVTYPE in udevadm info) and Path (this
matches on the stable device path as exposed as ID_PATH, and not the
unstable DEVPATH). A .link file matches a given device if all of the
specified keys do. Currently the keys are treated as plain strings,
but some limited globbing may later be added to the keys where it
makes sense.
Example:
/etc/net/links/50-wireless.link
[Match]
MACAddress=98:f2:e4:42:c6:92
Path=pci-0000:02:00.0-bcma-0
Type=wlan
[Link]
Description=The wireless link
We now treat passno as boleans in the generators, and don't need this any more. fsck itself
is able to sequentialize checks on the same local media, so in the common case the ordering
is redundant.
It is still possible to force an order by using .d fragments, in case that is desired.
rename old versions to ascii_*
Do not take into account zerowidth characters, but do consider double-wide characters.
Import needed utf8 helper code from glib.
v3: rebase ontop of utf8 restructuring work
[zj: tweak the algorithm a bit, move new code to separate file]
This adds a lightweight scheme how to define interfaces in static fixed
arrays which then can be easily registered on a bus connection. This
makes it much easier to write bus services.
This automatically handles implementation of the Properties,
ObjectManager, and Introspection bus interfaces.
The cgroup attribute memory.soft_limit_in_bytes is unlikely to stay
around in the kernel for good, so let's not expose it for now. We can
readd something like it later when the kernel guys decided on a final
API for this.
This adds a simple generator that is capable of automatically
discovering certain GPT partitions by their type UUID and mount/enable
them. This currently covers swap partitions and /home partitions, but is
expected to grow more features soon.
This currently doesn't handle LUKS encrypted /home.
This enables all swap partitions of type
0657fd6da4ab43c484e50933c84b4f4f, if found.
This mounts the first partition of type 933ac7e12eb44f13b8440e14e2aef915
as /home, if it is found.
So far, we would show up to 128 bytes from a message, simply
cutting of the rest. With multiline messages, it is quite common
for a message to be longer than that, and this model doesn't really
work anymore.
A new limit is added: up to 3 lines will be shown, unless --full is
used (c.f. first line below). The limit for bytes is extended to 300
bytes. An ellipsis will always be used, if some form of truncation
occurs. If the tail of the message is cut off, either because of
length or line limit, dots will be shown at the end of the last
line. If this last line is short, the dots will be simply appended. If
the last line is too long for that, it will be ellipsized with dots at
the very end.
Note that the limits are in bytes, not characters, and we suck at
outputting unicode strings (c.f. last three lines below).
Aug 11 10:46:21 fedora python[67]: test message
line
line...
Aug 11 10:50:47 fedora python[76]: test message word word word word word word word word word word word wor...
Aug 11 10:55:11 fedora python[83]: xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx...
Aug 11 11:03:21 fedora python[90]: ąąąąąąąąąąąąąąąąąąąąąąąąąąąąąą...
Aug 11 11:03:53 fedora python[97]: aąąąąąąąąąąąąąąąąąąąąąąąąąąąąąą...
Aug 11 11:25:45 fedora python[121]: aąąąąąąąąąąąąąąąąąąąąąąąąąąąąąąąąąą�...
"systemctl set-log-level" is a command for analysis and tracing hence
"systemd-analyze" should be the better home for it, thus allowing us to
make the overly large "systemctl" a bit smaller.
It's an analysis command and its format is explicitly not covered by any
stability guarantees, hence move away from systemctl and into
systemd-analyze, minimizing the already large interface of systemctl a
bit.
This patch also adds auto-paging to the various systemd-analyze commands
where that makes sense
$ systemctl --user status hoohoo
hoohoo.service
Loaded: loaded (/home/zbyszek/.config/systemd/user/hoohoo.service; static)
Active: inactive (dead)
start condition failed at Tue 2013-06-25 18:08:42 EDT; 1s ago
ConditionPathExists=/tmp/hoo was not met
Full information is exported over D-Bus:
[(condition, trigger, negate, param, state),...]
where state is one of "failed" (<0), "untested" (0), "OK" (>0).
I've decided to use 0 for "untested", because it might be useful to
differentiate different types of failure later on, without breaking
compatibility.
systemctl shows the failing condition, if there was a non-trigger
failing condition, or says "none of the trigger conditions were met",
because there're often many trigger conditions, and they must all
fail for the condition to fail, so printing them all would consume
a lot of space, and bring unnecessary attention to something that is
quite low-level.
A few asserts are replaced with 'return -EINVAL'. I think that
assert should not be used to check argument in public functions.
Fields in struct sd_journal are rearranged to make it less
swiss-cheesy.
Hi,
I redid the boot ID look up to use enumerate_unique.
This is quite fast if the cache is warm but painfully slow if
it isn't. It has a slight chance of returning the wrong order if
realtime clock jumps around.
This one has to do n searches for every boot ID there is plus
a sort, so it depends heavily on cache hotness. This is in contrast
to the other way of look-up through filtering by a MESSAGE_ID,
which only needs about 1 seek + whatever amount of relative IDs
you want to walk.
I also have a linked-list + (in-place) mergesort version of this
patch, which has pretty much the same runtime. But since this one
is using libc sorting and armortized allocation, I prefer this
one.
To summarize: The MESSAGE_ID way is a *lot* faster but can be
incomplete due to rotation, while the enumerate+sort will find
every boot ID out there but will be painfully slow for large
journals and cold caches.
You choose :P
Jan
As of kmod v14, it is possible to export the static node information from
/lib/modules/`uname -r`/modules.devname in tmpfiles.d(5) format.
Use this functionality to let systemd-tmpfilesd create the static device nodes
at boot, and drop the functionality from systemd-udevd.
As an effect of this we can move from systemd-udevd to systemd-tmpfiles-setup-dev:
* the conditional CAP_MKNOD (replaced by checking if /sys is mounted rw)
* ordering before local-fs-pre.target (see 89d09e1b5c)
In order to prepare things for the single-writer cgroup scheme, let's
make logind use systemd's own primitives for cgroup management.
Every login user now gets his own private slice unit, in which his sessions
live in a scope unit each. Also, add user@$UID.service to the same
slice, and implicitly start it on first login.
"Scope" units are very much like service units, however with the
difference that they are created from pre-existing processes, rather
than processes that systemd itself forks off. This means they are
generated programmatically via the bus API as transient units rather
than from static configuration read from disk. Also, they do not provide
execution-time parameters, as at the time systemd adds the processes to
the scope unit they already exist and the parameters cannot be applied
anymore.
The primary benefit of this new unit type is to create arbitrary cgroups
for worker-processes forked off an existing service.
This commit also adds a a new mode to "systemd-run" to run the specified
processes in a scope rather then a transient service.
Transient units can be created via the bus API. They are configured via
the method call parameters rather than on-disk files. They are subject
to normal GC. Transient units currently may only be created for
services (however, we will extend this), and currently only ExecStart=
and the cgroup parameters can be configured (also to be extended).
Transient units require a unique name, that previously had no
configuration file on disk.
A tool systemd-run is added that makes use of this functionality to run
arbitrary command lines as transient services:
$ systemd-run /bin/ping www.heise.de
Will cause systemd to create a new transient service and run ping in it.
This introduces two bus calls to make runtime changes to selected bus
properties, optionally with persistence.
This currently hooks this up only for three cgroup atributes, but this
brings the infrastructure to add more changable attributes.
This allows setting multiple attributes at once, and takes an array
rather than a dictionary of properties, in order to implement simple
resetting of lists using the same approach as when they are sourced from
unit files. This means, that list properties are appended to by this
call, unless they are first reset via assigning the empty list.
Replace the very generic cgroup hookup with a much simpler one. With
this change only the high-level cgroup settings remain, the ability to
set arbitrary cgroup attributes is removed, so is support for adding
units to arbitrary cgroup controllers or setting arbitrary paths for
them (especially paths that are different for the various controllers).
This also introduces a new -.slice root slice, that is the parent of
system.slice and friends. This enables easy admin configuration of
root-level cgrouo properties.
This replaces DeviceDeny= by DevicePolicy=, and implicitly adds in
/dev/null, /dev/zero and friends if DeviceAllow= is used (unless this is
turned off by DevicePolicy=).
- This changes all logind cgroup objects to use slice objects rather
than fixed croup locations.
- logind can now collect minimal information about running
VMs/containers. As fixed cgroup locations can no longer be used we
need an entity that keeps track of machine cgroups in whatever slice
they might be located. Since logind already keeps track of users,
sessions and seats this is a trivial addition.
- nspawn will now register with logind and pass various bits of metadata
along. A new option "--slice=" has been added to place the container
in a specific slice.
- loginctl gained commands to list, introspect and terminate machines.
- user.slice and machine.slice will now be pulled in by logind.service,
since only logind.service requires this slice.
In order to prepare for the kernel cgroup rework, let's introduce a new
unit type to systemd, the "slice". Slices can be arranged in a tree and
are useful to partition resources freely and hierarchally by the user.
Each service unit can now be assigned to one of these slices, and later
on login users and machines may too.
Slices translate pretty directly to the cgroup hierarchy, and the
various objects can be assigned to any of the slices in the tree.
Also reworded a few debug messages for brevity, and added a log
statement which prints out the filter at debug level:
Journal filter: (((UNIT=sys-module-configfs.device AND _PID=1) OR (COREDUMP_UNIT=sys-module-configfs.device AND MESSAGE_ID=fc2e22bc6ee647b6b90729ab34a250b1) OR _SYSTEMD_UNIT=sys-module-configfs.device) AND _BOOT_ID=4e3c518ab0474c12ac8de7896fe6b154)
We want to allow clients to process an sd_bus_message on a different
thread than it was received on. Since unreffing a bus message might
readd some of its memfds to the memfd cache add some minimal locking
around the cache.
Zbigniew Jędrzejewski-Szmek