mirror of
https://gitlab.freedesktop.org/NetworkManager/NetworkManager
synced 2024-09-06 17:14:02 +00:00
297d4985ab
Previously, we used the generated GDBusInterfaceSkeleton types and glued
them via the NMExportedObject base class to our NM types. We also used
GDBusObjectManagerServer.
Don't do that anymore. The resulting code was more complicated despite (or
because?) using generated classes. It was hard to understand, complex, had
ordering-issues, and had a runtime and memory overhead.
This patch refactors this entirely and uses the lower layer API GDBusConnection
directly. It replaces the generated code, GDBusInterfaceSkeleton, and
GDBusObjectManagerServer. All this is now done by NMDbusObject and NMDBusManager
and static descriptor instances of type GDBusInterfaceInfo.
This adds a net plus of more then 1300 lines of hand written code. I claim
that this implementation is easier to understand. Note that previously we
also required extensive and complex glue code to bind our objects to the
generated skeleton objects. Instead, now glue our objects directly to
GDBusConnection. The result is more immediate and gets rid of layers of
code in between.
Now that the D-Bus glue us more under our control, we can address issus and
bottlenecks better, instead of adding code to bend the generated skeletons
to our needs.
Note that the current implementation now only supports one D-Bus connection.
That was effectively the case already, although there were places (and still are)
where the code pretends it could also support connections from a private socket.
We dropped private socket support mainly because it was unused, untested and
buggy, but also because GDBusObjectManagerServer could not export the same
objects on multiple connections. Now, it would be rather straight forward to
fix that and re-introduce ObjectManager on each private connection. But this
commit doesn't do that yet, and the new code intentionally supports only one
D-Bus connection.
Also, the D-Bus startup was simplified. There is no retry, either nm_dbus_manager_start()
succeeds, or it detects the initrd case. In the initrd case, bus manager never tries to
connect to D-Bus. Since the initrd scenario is not yet used/tested, this is good enough
for the moment. It could be easily extended later, for example with polling whether the
system bus appears (like was done previously). Also, restart of D-Bus daemon isn't
supported either -- just like before.
Note how NMDBusManager now implements the ObjectManager D-Bus interface
directly.
Also, this fixes race issues in the server, by no longer delaying
PropertiesChanged signals. NMExportedObject would collect changed
properties and send the signal out in idle_emit_properties_changed()
on idle. This messes up the ordering of change events w.r.t. other
signals and events on the bus. Note that not only NMExportedObject
messed up the ordering. Also the generated code would hook into
notify() and process change events in and idle handle, exhibiting the
same ordering issue too.
No longer do that. PropertiesChanged signals will be sent right away
by hooking into dispatch_properties_changed(). This means, changing
a property in quick succession will no longer be combined and is
guaranteed to emit signals for each individual state. Quite possibly
we emit now more PropertiesChanged signals then before.
However, we are now able to group a set of changes by using standard
g_object_freeze_notify()/g_object_thaw_notify(). We probably should
make more use of that.
Also, now that our signals are all handled in the right order, we
might find places where we still emit them in the wrong order. But that
is then due to the order in which our GObjects emit signals, not due
to an ill behavior of the D-Bus glue. Possibly we need to identify
such ordering issues and fix them.
Numbers (for contrib/rpm --without debug on x86_64):
- the patch changes the code size of NetworkManager by
- 2809360 bytes
+ 2537528 bytes (-9.7%)
- Runtime measurements are harder because there is a large variance
during testing. In other words, the numbers are not reproducible.
Currently, the implementation performs no caching of GVariants at all,
but it would be rather simple to add it, if that turns out to be
useful.
Anyway, without strong claim, it seems that the new form tends to
perform slightly better. That would be no surprise.
$ time (for i in {1..1000}; do nmcli >/dev/null || break; echo -n .; done)
- real 1m39.355s
+ real 1m37.432s
$ time (for i in {1..2000}; do busctl call org.freedesktop.NetworkManager /org/freedesktop org.freedesktop.DBus.ObjectManager GetManagedObjects > /dev/null || break; echo -n .; done)
- real 0m26.843s
+ real 0m25.281s
- Regarding RSS size, just looking at the processes in similar
conditions, doesn't give a large difference. On my system they
consume about 19MB RSS. It seems that the new version has a
slightly smaller RSS size.
- 19356 RSS
+ 18660 RSS
|
||
|---|---|---|
| clients | ||
| contrib | ||
| data | ||
| dispatcher | ||
| docs | ||
| examples | ||
| introspection | ||
| libnm | ||
| libnm-core | ||
| libnm-glib | ||
| libnm-util | ||
| m4 | ||
| man | ||
| po | ||
| shared | ||
| src | ||
| tools | ||
| vapi | ||
| .dir-locals.el | ||
| .gitignore | ||
| .travis.yml | ||
| AUTHORS | ||
| autogen.sh | ||
| ChangeLog | ||
| config.h.meson | ||
| configure.ac | ||
| CONTRIBUTING | ||
| COPYING | ||
| linker-script-binary.ver | ||
| linker-script-devices.ver | ||
| linker-script-settings.ver | ||
| MAINTAINERS | ||
| Makefile.am | ||
| Makefile.examples | ||
| Makefile.glib | ||
| Makefile.vapigen | ||
| meson.build | ||
| meson_options.txt | ||
| meson_post_install.py | ||
| NetworkManager.pc.in | ||
| NEWS | ||
| README | ||
| TODO | ||
| valgrind.suppressions | ||
| zanata.xml | ||
****************** 2008-12-11: NetworkManager core daemon has moved to git.freedesktop.org! git clone git://git.freedesktop.org/git/NetworkManager/NetworkManager.git ****************** Networking that Just Works -------------------------- NetworkManager attempts to keep an active network connection available at all times. The point of NetworkManager is to make networking configuration and setup as painless and automatic as possible. NetworkManager is intended to replace default route, replace other routes, set IP addresses, and in general configure networking as NM sees fit (with the possibility of manual override as necessary). In effect, the goal of NetworkManager is to make networking Just Work with a minimum of user hassle, but still allow customization and a high level of manual network control. If you have special needs, we'd like to hear about them, but understand that NetworkManager is not intended for every use-case. NetworkManager will attempt to keep every network device in the system up and active, as long as the device is available for use (has a cable plugged in, the killswitch isn't turned on, etc). Network connections can be set to 'autoconnect', meaning that NetworkManager will make that connection active whenever it and the hardware is available. "Settings services" store lists of user- or administrator-defined "connections", which contain all the settings and parameters required to connect to a specific network. NetworkManager will _never_ activate a connection that is not in this list, or that the user has not directed NetworkManager to connect to. How it works: The NetworkManager daemon runs as a privileged service (since it must access and control hardware), but provides a D-Bus interface on the system bus to allow for fine-grained control of networking. NetworkManager does not store connections or settings, it is only the mechanism by which those connections are selected and activated. To store pre-defined network connections, two separate services, the "system settings service" and the "user settings service" store connection information and provide these to NetworkManager, also via D-Bus. Each settings service can determine how and where it persistently stores the connection information; for example, the GNOME applet stores its configuration in GConf, and the system settings service stores it's config in distro-specific formats, or in a distro- agnostic format, depending on user/administrator preference. A variety of other system services are used by NetworkManager to provide network functionality: wpa_supplicant for wireless connections and 802.1x wired connections, pppd for PPP and mobile broadband connections, DHCP clients for dynamic IP addressing, dnsmasq for proxy nameserver and DHCP server functionality for internet connection sharing, and avahi-autoipd for IPv4 link-local addresses. Most communication with these daemons occurs, again, via D-Bus. Why doesn't my network Just Work? Driver problems are the #1 cause of why NetworkManager sometimes fails to connect to wireless networks. Often, the driver simply doesn't behave in a consistent manner, or is just plain buggy. NetworkManager supports _only_ those drivers that are shipped with the upstream Linux kernel, because only those drivers can be easily fixed and debugged. ndiswrapper, vendor binary drivers, or other out-of-tree drivers may or may not work well with NetworkManager, precisely because they have not been vetted and improved by the open-source community, and because problems in these drivers usually cannot be fixed. Sometimes, command-line tools like 'iwconfig' will work, but NetworkManager will fail. This is again often due to buggy drivers, because these drivers simply aren't expecting the dynamic requests that NetworkManager and wpa_supplicant make. Driver bugs should be filed in the bug tracker of the distribution being run, since often distributions customize their kernel and drivers. Sometimes, it really is NetworkManager's fault. If you think that's the case, please file a bug at http://bugzilla.gnome.org and choose the NetworkManager component. Attaching the output of /var/log/messages or /var/log/daemon.log (wherever your distribution directs syslog's 'daemon' facility output) is often very helpful, and (if you can get) a working wpa_supplicant config file helps enormously.