Up to now, the "include" directory contained (only) header files that were
used project-wide by libs, core, clients, et al.
Since the directory now also contains a non-header file, the "include"
name is misleading. Instead of adding yet another directory that is
project-wide, with non-header-only content, rename the "include"
directory to "shared".
Parent MAC can be NULL if the interface has gone, fix the following
failed assertion:
[devices/nm-device-vlan.c:107] parent_hwaddr_changed(): (vlan1): parent hardware address changed
nm_device_set_hw_addr: assertion 'addr != NULL' failed
While at it, improve logging by printing the new MAC address.
Fixes: e6d7fee5a6
Make it possible to change ethernet.mtu and
ethernet.cloned-mac-address properties of tun/tap devices
(cloned-mac-address is meaningful only for taps).
When adding a very first item for a certain id, don't yet create a
hash table but store the first item inplace. This optimizes storage
for the case where we have only one item for a certain id.
Downsides:
- Add some additional overhead to manage the index
- The NMPCacheId struct grows to 16 bytes, affecting
hashing performance for all object types.
Still do it, based on the assumption that it doesn't matter
for a low number of interfaces. But the O(1) access time matters
when having lots of interfaces.
When deleting an object, we allow failure to delete a non-existing object.
Thus, the only thing we care about is whether the object is no longer
present after deletion. Adjust the return values to reflect that.
The idea was allowing pending delayed-actions and process them in an
idle handler. We dont want to do that, because whenever platform code
returns, we want to have no pending actions -- because otherwise the
platform cache might be in an inconsistent state.
Just drop it.
We would not expect that delayed_action_handle_all() is called recursively.
Assert against that. If we ever happen to call it recursively, we would
need to take care of properly avoiding infinite loops or deep call
stacks.
Track pending netlink requests and properly report the resulting errno
back.
Currently we send only requests in do_request_link() and do_request_all().
These callers don't actually care about the result, all they care that the
request is answered before returning back from platform code to the caller.
Thus, up to now the tracking of the sequence number was pretty simple.
Later we also want to get the errno from a request, thus rework sending
requests to also remember about outstanding sequence numbers and
properly track muliple parallel requests.
Later the synchronous actions (e.g. add-link) will also be handled
via the asynchronous socket.
Calling nl_recvmsgs_default() leads dirctly to recvmsgs() from
"nl.c". This functions reads messages (recvmsg) in a loop and
invokes the callbacks.
Later we want to merge nlh and nlh_event, meaning that we must
anticipate parsing unrelated messages while waiting for an ACK.
While that would be possible by registering different callbacks
and letting them interact, it is actually more complicated. Just
assume full control over the message parsing.
Basically, copy recvmsgs() to event_handler_recvmsg(). For now just
copy the function and do little adjustment (to show the similarity
to the original). Cleanup follows.
The nm_device_master_add_slave() also modifies slave's master property which
impacts the ability to enslave. When called in reaction to external
master property change we now no longer call enslave_slave which used to queue
the recheck previously:
# nmcli c add type bridge ifname br0
# ip link add dummy0 type dummy
# ip link set dummy0 up
# ip link set dummy0 master br0 # We should recheck for assumed connection
# here, since dummy0 can now be assumed.
We only need to do that when we're replacing the master with a different
one. Just after the link creation is has no master and we'd remove it
from the master device here.
Randomly choosing one between the two link creation methods (platform
and external) makes it more difficult to immediately detect when a bug
in introduced. Always execute them using both methods to have a better
code coverage, but still try a third time with a random one to test
different combinations of methods.
Fixes a crash if we can't read the ATM index. We need the ATM
index, and we can't do anything with the device before we have it,
so don't bother creating one if we we can't get it.
NetworkManager[9662]: <error> [1449678770.705541] [nm-device-adsl.c:607] constructor(): (atmtcp0): error reading ATM device index
(NetworkManager:9662): GLib-GObject-CRITICAL **: object NMDeviceAdsl 0x1e8f880 finalized while still in-construction
(NetworkManager:9662): GLib-GObject-CRITICAL **: Custom constructor for class NMDeviceAdsl returned NULL (which is invalid). Please use GInitable instead.
**
NetworkManager-adsl:ERROR:nm-atm-manager.c:121:adsl_add: assertion failed: (device)
At some point the platform changed to no longer ask the kernel for
interfaces when one wasn't in its cache, but to wait for netlink
events to be notified of the new interface. That broke some assumptions
that the ADSL code was making, causing a crash.
Rework the ADSL br2684 interface to clean up a couple of things
(get rid of 'disposed', consolidate dispose/deactivate cleanup) and
watch for the br2684 interface to show up with a periodic timeout.
Only do so on user initiated changes. Fixes this:
# ip link add br0 type bridge
# ip addr add 2001:DB8::666/64 dev br0
# ip link set br0 up # A generated connection is assumed
# ip link del br0 # The device and its address are removed.
# The address removal triggers an update
# of the connection's ipv6 settings,
# which causes the NMDevice to reappear.
# ip link add br0 type bridge # The new plink is associated with
# the NMDevice, managed by NM
Fixes autoconnect after the device is realized again:
# nmcli c add type team
# nmcli c up team
# nmcli d dis nm-team # autoconnect is blocked
# nmcli c del team # the is unrealized
# nmcli c add type team # the device is realized again, not
# activating with the new connection
