When multiple address are assigned to an interface and the kernel must
decide which one should be used to communicate with a given IP, it
chooses the most specific one in the same subnet as the
destination. In case there are multiple addresses in the same subnet,
the primary address is choosen, which is basically the first one that
was added.
With commit 7197425137 ("device: expose NMIP4Config:addresses in
stable/defined sort order") we sorted all the addresses before
committing the configuration, with the side effect that the order no
longer respected the one in the user configuration.
Instead, change the sort function to keep the subnet order unchanged.
(cherry picked from commit e02752c2ed)
Since commit 9fafb382db, we would
explicitly set libnl's socket buffer size to 4*getpagesize().
That is also the default of libnl itself. Additionally, we would
workaround too small buffers by increasing the buffer size up to 512K.
A too small buffer causes messages to be lost. Usually, that only
results in a cache-resync, which isn't too bad. Lost messages are however
a problem if the lost message was an ACK that we were waiting for.
However, it is rather unlikely to happen, because it's expected that
the buffer size gets adjusted already when the cache is filled initially,
before any other requests are pending.
Still, let's increase the default buffer size to 32K, hoping that this
initial value is already large enough to avoid the problem altogether.
Note that iproute2 also uses a buffer size of 32K [1] [2].
Alternatively, we could use MSG_PEEK like systemd does [3]. However,
that requires two syscalls per message.
[1] https://patchwork.ozlabs.org/patch/592178/
[2] https://git.kernel.org/cgit/linux/kernel/git/shemminger/iproute2.git/tree/lib/libnetlink.c?id=f5f760b81250630da23a4021c30e802695be79d2#n274
[3] cd66af2274/src/libsystemd/sd-netlink/netlink-socket.c (L323)
(cherry picked from commit ed82b6bcb3)
Autoconnect property doesn't really matter for the applied
connection. Whitelist it from the properties and allow changing
it during reapply.
(cherry picked from commit d29839c430)
We don't want to enable MSG_PEEK due to the overhead. But when we detect
that we just lost a message due to MSG_TRUNC, increase the buffer size and
retry.
See-also: 55ea6e6b6c
(cherry picked from commit 9fafb382db)
Plugins could be already failed or disconnected when the helper fires.
E.g. they could send in an invalid IP4Config that would cause them to
fail and then follow with an IP6Config before they realize it's of no
use. We'd hit an assertion failure in that case, because the
NMVpnConnection would already be cleaned up.
(cherry picked from commit bf70ed2e86)
Otherwise its path remains visible on D-Bus despite the object is gone,
making libnm sad and grumpy:
libnm-WARNING **: no object known for /org/freedesktop/NetworkManager/AccessPoint/666
(cherry picked from commit d0c01cc79d)
When going to sleep, we unmanage devices setting the unmanaged flags
immediately but delaying the state transition (because we do it from
another state transition). The signal handler can be executed after
the wake and, especially, after we have already re-managed the device,
making the device unmanaged again.
Detect such situation and force the state to UNMANAGED (which will
also clear any pending state change), so that later we manage the
device again and it will try to activate any available connection.
Fixes: 81ea812362
(cherry picked from commit 3cc06c3db679c1ff2f61a301396393300d36adbb)
The interaction between the manager state and connectivity check code
is tricky. When there is an active connection with a default route and
NMConnectivity reports full connectivity, we set the CONNECTED_GLOBAL
state. However, if the connectivity check hasn't run yet, we stay in
CONNECTED_SITE state. If there are also other connections that are
activating, the state is set to CONNECTING.
This is a problem, because in CONNECTING we never run the connectivity
check and thus we fail to recognize that there is full connectivity
until a periodic check is run.
To solve this, schedule the connectivity check every time there is an
active connection with default route, even if other connection are
still activating, so that the check result can make the state progress
to CONNECTED_GLOBAL.
(cherry picked from commit 084da69a30)
For example for tun devices we get a lot of
(tun7): hw-addr: failed reading current MAC address
warnings. Just be silent about it. We log when something
changes, we don't need to log when we fail to obtain
a MAC address.
Thereby, refactor nm_device_update_hw_address() to return early.
(cherry picked from commit 0e0018c801)
Conflicts:
src/devices/nm-device.c
The unmanaged flags PLATFORM_INIT indicates whether UDEV is done
initializing the device. We should not handle IP config changes
before that pointer.
This avoids codepaths that require the permanent MAC address of the
device. We should not freeze the permanent MAC address before
UDEV initialized the device, for two reasons:
- getting the permanent MAC address using ethtool is racy as
UDEV might still rename the interface.
- freezing a fake permanent MAC address should only happen after
UDEV is done configuring the MAC address of software devices.
#0 0x000055555568bc7a in nm_device_update_permanent_hw_address (self=self@entry=0x555555f0fb70 [NMDeviceVeth], force_freeze=force_freeze@entry=1) at src/devices/nm-device.c:11817
#1 0x000055555568c443 in nm_device_get_permanent_hw_address_full (self=self@entry=0x555555f0fb70 [NMDeviceVeth], force_freeze=force_freeze@entry=1, out_is_fake=out_is_fake@entry=0x0)
at src/devices/nm-device.c:12227
#2 0x000055555568cb06 in nm_device_get_permanent_hw_address (self=self@entry=0x555555f0fb70 [NMDeviceVeth]) at src/devices/nm-device.c:12237
#3 0x000055555568cb50 in spec_match_list (self=0x555555f0fb70 [NMDeviceVeth], specs=0x555555a5c000 = {...}) at src/devices/nm-device.c:12294
#4 0x00005555556a4ee6 in spec_match_list (device=0x555555f0fb70 [NMDeviceVeth], specs=0x555555a5c000 = {...}) at src/devices/nm-device-ethernet.c:1461
#5 0x00005555556978db in nm_device_spec_match_list (self=self@entry=0x555555f0fb70 [NMDeviceVeth], specs=0x555555a5c000 = {...}) at src/devices/nm-device.c:12277
#6 0x000055555558e187 in _match_section_infos_lookup (match_section_infos=0x555555a5d500, keyfile=0x555555a46f80, property=property@entry=0x555555793123 "ipv4.route-metric", device=device@entry=0x555555f0fb70 [NMDeviceVeth], out_value=out_value@entry=0x7fffffffe018) at src/nm-config-data.c:1169
#7 0x00005555555922ca in nm_config_data_get_connection_default (self=0x555555a548c0 [NMConfigData], property=property@entry=0x555555793123 "ipv4.route-metric", device=device@entry=0x555555f0fb70 [NMDeviceVeth]) at src/nm-config-data.c:1234
#8 0x00005555556790cd in _get_ipx_route_metric (self=self@entry=0x555555f0fb70 [NMDeviceVeth], is_v4=is_v4@entry=1) at src/devices/nm-device.c:1142
#9 0x000055555567912e in nm_device_get_ip4_route_metric (self=self@entry=0x555555f0fb70 [NMDeviceVeth]) at src/devices/nm-device.c:1161
#10 0x000055555567da6c in ip4_config_merge_and_apply (self=self@entry=0x555555f0fb70 [NMDeviceVeth], config=config@entry=0x0, commit=commit@entry=0, out_reason=out_reason@entry=0x0)
at src/devices/nm-device.c:4787
#11 0x000055555567e0fb in update_ip4_config (self=self@entry=0x555555f0fb70 [NMDeviceVeth], initial=initial@entry=0) at src/devices/nm-device.c:9532
#12 0x0000555555693acd in queued_ip4_config_change (user_data=0x555555f0fb70) at src/devices/nm-device.c:9651
#13 0x00007ffff4c966ba in g_main_context_dispatch (context=0x555555a46af0) at gmain.c:3154
#14 0x00007ffff4c966ba in g_main_context_dispatch (context=context@entry=0x555555a46af0) at gmain.c:3769
#15 0x00007ffff4c96a70 in g_main_context_iterate (context=0x555555a46af0, block=block@entry=1, dispatch=dispatch@entry=1, self=<optimized out>) at gmain.c:3840
#16 0x00007ffff4c96d92 in g_main_loop_run (loop=0x555555a47400) at gmain.c:4034
#17 0x000055555558372a in main (argc=<optimized out>, argv=<optimized out>) at src/main.c:411
(cherry picked from commit 31ca7962f8)
Before the link is initialized, that is before UDEV completed
initializing the device, we should not evaluate the user-settings
unmanaged flags.
The reason is, that evaluating it likely involves looking at the
permanent MAC address, which might use the wrong fake MAC address
(before UDEV set the right one). Also, it might use the wrong ifname
to lookup the permanent MAC address via ethtool.
(cherry picked from commit c0d249b733)
The permanent MAC address of an NMDevice shall not change as
long as the device is realized. That is, we read it only once
and don't change it afterwards.
There are two issues that this commit tries to mitigate:
(1) users are advised to use UDEV to rename interfaces. As we lookup
the permenent MAC address using ethtool (which uses the interface
name), there is a race where we could read the permanent MAC
address using the wrong interface name. We should wait until
UDEV finished initializing the device and until the interface
name is stable (see rh#1388286).
This commit still cannot avoid the race of ethtool entirely. It only
tries to avoid ethtool until UDEV has done its work. That is, until we
expect the interface name no longer to change.
(2) some device types, don't have a permanent MAC address so we fall
back to use the currently set address (fake). Again, users are advised
to use UDEV to configure the MAC addresses on such software devices.
Thus, we should not get the fake MAC address until UDEV initialized
the device.
This patch actually doesn't solve the problem at all yet.
The reason is that a regular caller of nm_device_get_permanent_hw_address() can
not afford to wait until UDEV settled. Thus, any user who requests the
permanent MAC address before the link is initialized, runs into the
problems above.
In a next step, we shall revisit such calls to nm_device_get_permanent_hw_address()
and delay them until the link is initialized.
(cherry picked from commit 7b7c653c4f)
Conflicts:
src/devices/nm-device.c
src/nm-manager.c
We repeatedly call nm_device_update_hw_address() to reset the cached
MAC address of the device. However, we don't allow changing the address
length once it is set.
Multiple entities (initial, current and permanent MAC address) are all
checked to have the same address length. Changing the length would be a
very strange thing (and probably indicate a bug somewhere else).
Just don't allow that.
(cherry picked from commit cbea1f9f23)
Now that we persist the fake permanent address across
restart of NetworkManager, we want to consider fake
addresses as good enough in most cases.
(cherry picked from commit 416164aa29)
The default wired connection is already generated allowing the use of a fake
address, but for the state file and the device matching specs only non-fake
addresses are used. Let's allow fake addresses consistently, so that default
wired connections work properly in containers (where the veth address is
considered fake) as well.
Also, it would really be a better idea to use ifnames everywhere instead, but
that would change the format of the state file.
(cherry picked from commit bcb685c4cb)
_nm_utils_hwaddr_length() did a validation of the string
and returned the length of the address. In all cases where
we were interested in that, we also either want to validate
the address, get the address in binary form, or canonicalize
the address.
We can avoid these duplicate checks, by using _nm_utils_hwaddr_aton()
which both does the parsing and returning the length.
(cherry picked from commit e5fe5a4c03)
Since commit 1495853e01, LOGD_HW is renamed to
LOGD_PLATFORM. Remove the internal usage of the deprecated name.
Backport this patch to nm-1-4 because it avoids follow-up conflicts with
future backports. Also, it is a really trivial change, meaning, we only
replace the obsolte LOGD_HW name.
(cherry picked from commit 64951f07fb)
Conflicts:
src/devices/nm-device.c
src/devices/wwan/nm-modem.c
nm_settings_connection_set_autoconnect_blocked_reason() must be called
on the settings-connection. Fixes the following:
GLib-GObject-WARNING **: invalid cast from 'NMSimpleConnection' to 'NMSettingsConnection'
Fixes: 06da353242
(cherry picked from commit 7034ea7aa3)
Depending on how arguments are passed to the called function,
this could lead to a crash.
Maybe not on 32 bit machines where the size of the pointer is
the size of an int.
Maybe not on x86_64, where the arguments are passed in registers.
Fixes: b88c309167
(cherry picked from commit 548a5440e9)
The @assoc_cancellable was never initialized and thus ineffective; fix
this.
Furthermore, we only cancel it in nm_supplicant_interface_disconnect()
as we expect that clients call the function before destroying the
interface. Don't assume this and also cancel it in dispose().
https://bugzilla.redhat.com/show_bug.cgi?id=1383628
(cherry picked from commit 0539725aef)
Some TTY drivers or devices appear to ignore port speed and always
report zero. Technically this means the port is hung up and control
lines should be disconnected, but with USB devices many of the serial
port attributes are meaningless and ignored by some devices.
pppd requires the port's speed to be greater than zero, and will
exit immediately when that is not the case, even though these
modems will work fine. Passing an explicit speed to pppd in this
case works around the issue, as pppd attempts to set that speed
on the port and doesn't actually care if that operation fails.
https://bugzilla.redhat.com/show_bug.cgi?id=1281731
(cherry picked from commit 01de14b1ddcd011ebc2f4676e5950b9ec890c698)
The 'device-added' and 'device-removed' signals indicate when the
value of the 'Devices' property changes. The property only returns
realized devices and so if a device unrealizes we should emit the
removed signal for it.
Fixes: 5da37a129chttps://bugzilla.gnome.org/show_bug.cgi?id=771324
(cherry picked from commit cdedd2b53e)
Some drivers (brcmfmac) don't change the MAC address right away.
NetworkManager works around that by waiting synchronously until
the address changes (commit 1a85103765).
wpa_supplicant on the other hand, only re-reads the MAC address
when changing state from DISABLED to ENABLED, which happens when
the interface comes up.
That is a bug in wpa_supplicant and the driver, but we can work-around by
waiting until the MAC address actually changed before setting the interface
IFF_UP. Also note, that there is still a race in wpa_supplicant which might
miss a change to DISABLED state altogether.
https://bugzilla.gnome.org/show_bug.cgi?id=770504https://bugzilla.redhat.com/show_bug.cgi?id=1374023
(cherry picked from commit 32f7c1d4b9)
An empty 802-11-wireless-security.proto is equivalent to
'wpa,rsn'. Previously we added the two protocols when reading the
connection and the variables were missing, with the result that an
empty value would be read as 'wpa,rsn' at the next restart. This is
harmless but makes the two connections appear as different, with bad
effects when 'monitor-connection-files' is enabled.
Ensure that the original value persists after a write/read cycle.
https://bugzilla.gnome.org/show_bug.cgi?id=770907
(cherry picked from commit 00c4e7e73a)
brcmfmac and possibly other drivers don't change the MAC address
right away, but instead the result is delayed. That is problematic
because we cannot continue activation before the MAC address is
settled.
Add a hack to workaround the issue by waiting until the MAC address
changed.
The previous attempt to workaround this was less intrusive: we would
just refresh the link once and check the result. But that turns out
not to be sufficent for all cases. Now, wait and poll.
https://bugzilla.gnome.org/show_bug.cgi?id=770456https://bugzilla.redhat.com/show_bug.cgi?id=1374023
(cherry picked from commit 1a85103765)
A D-Bus signal is asynchronous and it can happen that nm-dhcp-helper
emits the "Event" signal before the server is able to register a handler:
NM_DHCP_HELPER=/usr/libexec/nm-dhcp-helper
nmcli general logging level TRACE
for i in `seq 1 500`; do $NM_DHCP_HELPER & done
journalctl -u NetworkManager --since '1 min ago' | grep "didn't have associated interface" | wc -l
499
Avoid that, by calling the synchronous D-Bus method "Notify".
Interestingly, this race seem to exist since 2007.
Actually, we called g_dbus_connection_signal_subscribe() from inside
GDBusServer:new-connection signal. So it is not clear how such a race
could exist. I was not able to reproduce it by putting a sleep
before g_dbus_connection_signal_subscribe(). On the other hand, there
is bug rh#1372854 and above reproducer which strongly indicates that
events can be lost under certain circumstances.
Now we instead g_dbus_connection_register_object() from the
new-connection signal. According to my tests there was no more race
as also backed by glib's documentation. Still, keep a simple retry-loop
in nm-dhcp-helper just to be sure.
https://bugzilla.redhat.com/show_bug.cgi?id=1372854https://bugzilla.redhat.com/show_bug.cgi?id=1373276
(cherry picked from commit 2856a658b3)
Don't exit(1) from fatal_error() because that skips destroying
local variables in main(). Just return regularly.
(cherry picked from commit bb489163db)
It's not "signal-handles", as it currently tracks the registration ID of
type int. Rename it, it is effectively the list of connections that we
track.
(cherry picked from commit 2dd3a5245f)
