| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
During shutdown, we will need to still iterate the main loop
to do a coordinated shutdown. Currently we do not, and we just
exit, leaving a lot of objects hanging.
If we are going to fix that, we need during shutdown tell
NMDBusManager to reject all future operations.
Note that property getters and "GetManagerObjects" call is not
blocked. It continues to work.
Certainly for some operations, we want to allow them to be called even
during shutdown. However, these have to opt-in.
This also fixes an uglyness, where nm_dbus_manager_start() would
get the set-property-handler and the @manager as user-data. However,
NMDBusManager will always outlife NMManager, hence, after NMManager
is destroyed, the user-data would be a dangling pointer. Currently
that is not an issue, because
- we always leak NMManager
- we don't run the mainloop during shutdown
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When the D-Bus name is already taken, NM crashes in the following
way. That's because disposed object are not unexported when quitting
and so they linger in the bus-manager's list of exported objects,
causing an invalid access when a neighboring item is accessed. Instead
of just clearing the path, fully unexport the object.
The behavior of not forcefully exporting objects on quit was added in
f9ee20a7b2bd ("core: explicitly unexport objects when we're done with
them"), but such behavior doesn't seem to be needed by the stated
goal.
<error> [1524062008.1886] bus-manager: fatal failure to acquire D-Bus service "org.freedesktop.NetworkManager" (3). Service already taken
<trace> [1524062008.2327] config: state: success writing state file "/var/lib/NetworkManager/NetworkManager.state"
<trace> [1524062008.2338] dns-mgr: stopping...
<info> [1524062008.2344] exiting (error)
<debug> [1524062008.2628] disposing NMManager singleton (0xce587e0)
<trace> [1524062008.2640] dns-mgr: disposing
<debug> [1524062008.2651] disposing NMDnsManager singleton (0xceb8b50)
<debug> [1524062008.2666] disposing NMFirewallManager singleton (0xceb62b0)
<debug> [1524062008.2709] disposing NMHostnameManager singleton (0xce7b370)
<trace> [1524062008.2722] dbus-object[0xce70f40]: unexport: "/org/freedesktop/NetworkManager/AgentManager"
==16381== Invalid write of size 8
==16381== at 0x42F511: c_list_unlink_stale (c-list.h:158)
==16381== by 0x42F511: c_list_unlink (c-list.h:171)
==16381== by 0x42F511: _nm_dbus_manager_obj_unexport (nm-dbus-manager.c:1135)
==16381== by 0x4C5E35: nm_dbus_object_unexport (nm-dbus-object.c:165)
==16381== by 0x5C01E9: dispose (nm-agent-manager.c:1634)
==16381== by 0x6636F37: g_object_unref (gobject.c:3303)
==16381== by 0x4BDC89: _nm_singleton_instance_destroy (nm-core-utils.c:138)
==16381== by 0x400FA85: _dl_fini (in /usr/lib64/ld-2.27.so)
==16381== by 0x7F806AB: __run_exit_handlers (in /usr/lib64/libc-2.27.so)
==16381== by 0x7F807DB: exit (in /usr/lib64/libc-2.27.so)
==16381== by 0x41DA34: main (main.c:463)
==16381== Address 0xce706a0 is 48 bytes inside a block of size 176 free'd
==16381== at 0x4C2EDAC: free (vg_replace_malloc.c:530)
==16381== by 0x6ACA3E1: g_free (gmem.c:194)
==16381== by 0x6AE2572: g_slice_free1 (gslice.c:1136)
==16381== by 0x66550AE: g_type_free_instance (gtype.c:1943)
==16381== by 0x4505F8: dispose (nm-manager.c:6867)
==16381== by 0x6636F37: g_object_unref (gobject.c:3303)
==16381== by 0x4BDC89: _nm_singleton_instance_destroy (nm-core-utils.c:138)
==16381== by 0x400FA85: _dl_fini (in /usr/lib64/ld-2.27.so)
==16381== by 0x7F806AB: __run_exit_handlers (in /usr/lib64/libc-2.27.so)
==16381== by 0x7F807DB: exit (in /usr/lib64/libc-2.27.so)
==16381== by 0x41DA34: main (main.c:463)
==16381== Block was alloc'd at
==16381== at 0x4C2DBAB: malloc (vg_replace_malloc.c:299)
==16381== by 0x6ACA2D5: g_malloc (gmem.c:99)
==16381== by 0x6AE1E36: g_slice_alloc (gslice.c:1025)
==16381== by 0x6AE247C: g_slice_alloc0 (gslice.c:1051)
==16381== by 0x6654E09: g_type_create_instance (gtype.c:1848)
==16381== by 0x66376C7: g_object_new_internal (gobject.c:1799)
==16381== by 0x6638E14: g_object_new_with_properties (gobject.c:1967)
==16381== by 0x66399D0: g_object_new (gobject.c:1639)
==16381== by 0x5D6F18: nm_settings_new (nm-settings.c:1897)
==16381== by 0x4514B4: constructed (nm-manager.c:6489)
==16381== by 0x66378FA: g_object_new_internal (gobject.c:1839)
==16381== by 0x6638E14: g_object_new_with_properties (gobject.c:1967)
https://github.com/NetworkManager/NetworkManager/pull/96
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
The GObject property "path" exists for the sole reasons so that
other components can connect to the "notify::path" signal.
However, notifications are blocked by g_object_freeze_notify(),
and especially for NMDBusObject we want to make use of that to
combine multiple PropertiesChanged events into one.
This blocking of the signal is not desired for the case where
we wait for "notify::path". Convert that to a signal, which
will not be blocked.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
into class
I dislike the static hash table to cache the integer counter for
numbered paths. Let's instead cache the counter at the class instance
itself -- since the class contains the information how the export
path should be exported.
However, we cannot use a plain integer field inside the class structure,
because the class is copied between derived classes. For example,
NMDeviceEthernet and NMDeviceBridge both get a copy of the NMDeviceClass
instance. Hence, the class doesn't contain the counter directly, but
a pointer to one counter that can be shared between sibling classes.
|
|
|
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
|
