Don't handle master-ready at the beginning of stage2, but instead while
scheduling (and then possibly delaying the scheduling of stage2).
This seems more idiomatic:
When inside a stage and your part is done: call schedule-next-stage.
That is, always schedule the next stage, not the current one.
schedule-next-stage then might delay to really scheduling until the
device is ready for the next state.
Fixes: 85ac903bb8
(cherry picked from commit 7bbc090387)
During stage2, if the slave detected that it would need to wait for
the master, it would return FALSE (which removes the g-idle-handler).
However, it would not clear the activation-source, so later, when
the master becomes ready, its attempt to schedule stage2 again would
result in an error-log and the idle-handler would not be scheduled
again.
Fixes: 85ac903bb8https://bugzilla.redhat.com/show_bug.cgi?id=1268797https://bugzilla.redhat.com/show_bug.cgi?id=1183444
(cherry picked from commit c5210b322d)
When activating for example a team device which is to be enslaved to a
bridge, nm_device_activate_stage1_device_prepare() will postpone
stage 2.
In that case, we didn't register the "master-ready" of the team
device and thus never progressed the slave from stage2.
Reproduce:
# nmcli connection delete t-br0
# nmcli connection delete t-team0
nmcli connection add type bridge con-name t-br0 autoconnect no ifname i-br0 ip4 192.168.177.100/24 gw4 192.168.177.1
nmcli connection add type team con-name t-team0 autoconnect no ifname i-team0
nmcli connection modify id t-team0 connection.master i-br0 connection.slave-type bridge
nmcli connection up t-team0
For libnm library, "nm-dbus-interface.h" contains defines like the D-Bus
paths of NetworkManager. It is desirable to have this header usable without
having a dependency on "glib.h", for example for a QT application. For that,
commit c0852964a8 removed that dependancy.
For libnm-glib library, the analog to "nm-dbus-interface.h" is
"NetworkManager.h", and the same applies there. Commit
159e827a72 removed that include.
However, that broke build on PackageKit [1] which expected to get the
version macros by including "NetworkManager.h". So at least for libnm-glib,
we need to preserve old behavior so that a user including
"NetworkManager.h" gets the version macros, but not "glib.h".
Extract the version macros to a new header file "nm-version-macros.h".
This header doesn't include "glib.h" and can be included from
"NetworkManager.h". This gives as previous behavior and a glib-free
include.
For libnm we still don't include "nm-version-macros.h" to "nm-dbus-interface.h".
Very few users will actually need the version macros, but not using
libnm.
Users that use libnm, should just include (libnm's) "NetworkManager.h" to
get all headers.
As a special case, a user who doesn't want to use glib/libnm, but still
needs both "nm-dbus-interface.h" and "nm-version-macros.h", can include
them both separately.
[1] https://github.com/hughsie/PackageKit/issues/85
Fixes: 4545a7fe96
(cherry picked from commit 7bf10a75db)
Depending on the network and the values of the 'dad_transmits' and
'retrans_timeout_ms' sysctls, DAD for the IPv6LL address can take
quite a while. Obviously there must be some upper bound on the
timeout, but 5 seconds seems a bit low. In this case it was observed
to be ~12 seconds but the sysctl values are unknown. In any case,
we can't predict the network/config so being a bit more lenient here
makes sense.
https://bugzilla.redhat.com/show_bug.cgi?id=1101809
(cherry picked from commit 5b374a4a9f)
If SIM in a modem is locked, ModemManager can't initialize SupportedIpFamilies
and NetworkManager will set the property to 0. ModemManager then updates the
property after the modem is unlocked, but NetworkManager did not watch changes
to the property. And that resulted in a connection failure:
(ttyUSB1): Failed to connect 'O2 Internet': Connection requested IPv4 but IPv4 is unsuported by the modem.
(ttyUSB1): device state change: prepare -> failed (reason 'modem-init-failed') [40 120 28]
https://bugzilla.redhat.com/show_bug.cgi?id=1263959
(cherry picked from commit eecb4c46cc)
NetworkManager set wpa_supplicant's fragment_size option to 1300. But if MTU
was lower, wpa_supplicant failed with "l2_packet_send - sendto: Message too
long" due to fragmentation of EAP-TLS or EAP-PEAP packets.
Actually, MTU has to be 14 bytes bigger than the "fragment_size" parameter.
Ideally, wpa_supplicant would take MTU in the account and adjust the
fragmentation limit accordingly. See discussion in
http://lists.shmoo.com/pipermail/hostap/2015-August/033546.htmlhttps://bugzilla.gnome.org/show_bug.cgi?id=755145
(cherry picked from commit 94bbe7465f)
When a new link is detected, NM tries to generate a default "Wired
connection" in nm_settings_device_added(), but if the link has not
been initialized by udev yet the function returns early because
priv->unmanaged_flags = UNMANAGED_PLATFORM_INIT.
To be sure that a default connection is created is such situation, we
need to call again nm_settings_device_added() after link
initialization.
https://bugzilla.redhat.com/show_bug.cgi?id=1254089
(cherry picked from commit b3b0b46250)
If at the moment when spawning nm-iface-helper dhcp4/slaac
did not yet complete, we would not enable it.
That is wrong. If the connection indicates to use dhcp4/slaac,
it should be used by nm-iface-helper without considering the
current state on the device.
https://bugzilla.redhat.com/show_bug.cgi?id=1260243
(cherry picked from commit b0815813fa)
If DHCP fails for an assumed connection, NetworkManager would
transition the device to the FAILED and then to the ACTIVATED state
(because it is assumed); hence if the DHCP server goes temporarily
down the device will go into a permanent state without IP
configuration.
Fix this and try DHCP again after some time when the connection
is an assumed one.
https://bugzilla.redhat.com/show_bug.cgi?id=1246496
This would cause the ip_vti0 generic device (that appears upon insertion of
ip_vti module during libreswan ipsec stack init) to go managed and brought UP.
Without addresses assigned the device would cause all the VPN traffic to
disappear in the oblivion.
(cherry picked from commit 1c46ddf196)
When the DHCPv6 lease received from the server contains multiple
addresses, dhclient generates a new BOUND event for each of
them. Instead of overwriting the previous IP6 configuration for each
BOUND event, we should try to detect if the new configuration belongs
to the same lease and merge its addresses with the existing one in
such case.
This allows NetworkManager to configure multiple addresses on an
interface via DHCPv6.
https://bugzilla.gnome.org/show_bug.cgi?id=681764https://bugzilla.redhat.com/show_bug.cgi?id=1244293
(cherry picked from commit 1d6e8e8da7)
Don't disconnect the device when the DHCP renewal fails and there are
already configured static IP addresses on the device. Instead, keep
the device up and try DHCP again after some time.
https://bugzilla.redhat.com/show_bug.cgi?id=1168388
(cherry picked from commit abc96ecdfd)
The parent of a link (IFLA_LINK) can be in another network namespace and
thus invisible to NM.
This requires the netlink attribute IFLA_LINK_NETNSID which is supported
by recent versions of kernel and libnl.
In this case, set the parent field to NM_PLATFORM_LINK_OTHER_NETNS
and properly handle this special case.
(cherry picked from commit 790a0713d2)
The logging macros _LOGD(), etc. are specific to each
file as they format the message according to their context.
Still, they were cumbersome to define and their implementation
was repeated over and over (slightly different at times).
Move the declaration of these macros to "nm-logging.h".
The source file now only needs to define _NMLOG(), and either
_NMLOG_ENABLED() or _NMLOG_DOMAIN.
This reduces code duplication and encourages a common implementation
and usage of these macros.
(cherry picked from commit ad7cdfc766)
Don't clear NMDevice @master in nm_device_cleanup() if the device link
is still enslaved because this causes an inconsistent state in which
the slave in included in the @slaves field of master device but
@master of slave device is NULL.
In such state, if the master link gets deleted, NM receives a change
event for each slave and a deletion event for the master; the change
events should also remove slaves from @slaves of master device, but
since their @master field is NULL the removal can't be performed.
Later, when the master deletion event is received, @slaves is not empty
in dispose() of NMDevice and the following assertion is triggered:
dispose: runtime check failed: (priv->slaves == NULL)
https://bugzilla.redhat.com/show_bug.cgi?id=1243371
(cherry picked from commit b557f91a1b)
If linklocal6_start() finds that the device already has an IPv6
link-local address and returns NM_ACT_STAGE_RETURN_FINISH, we don't
need to fill @out_config with an empty configuration. A non-NULL
@out_config is required only for ret = NM_ACT_STAGE_RETURN_SUCCESS.
Fixes: 396dc2b3b4
(cherry picked from commit 54d10f0ec1)
During merge_and_apply(), we merge several NMIPxConfig into a new
@composite. For 'ignore-auto-routes' and 'ignore-auto-dns', we want to
prevent certain routes/dns-settings to be merged.
But it is wrong to reject settings from all partial configs.
For example, especially ext_ipx_config and vpn_ipx_config contain
routes that we still must merge and preserve.
This was recently changed by 79630c11e5
and previously by ab6548c621. But it was
wrong for a long time already.
Also note, that nm_ip4_config_merge() now also ignores NIS, WINS, and
dns-options.
https://bugzilla.gnome.org/show_bug.cgi?id=752546
Fixes: 79630c11e5
(cherry picked from commit 43e6f6a178)
The new flags are not yet used, so there is no change in functionality.
The flags NM_IP_CONFIG_MERGE_NO_ROUTES and NM_IP_CONFIG_MERGE_NO_DNS go
together with the 'ignore-auto-routes' and 'ignore-auto-dns' setting.
Note that for IPv4, NM_IP_CONFIG_MERGE_NO_DNS also ignores NIS, WINS, and dns-options.
This is different from current other places that handle 'ignore-auto-dns'
and only care about nameservers, domains, and searches.
(cherry picked from commit f7a8962dd4)
The MTU of 0 means default, not zero-length packets:
<warn> (wlp3s0): Lowering IPv6 MTU (1472) to match device MTU (0)
<warn> (wlp3s0): IPv6 MTU (0) smaller than 1280, adjusting
<warn> (wlp3s0): Raising device MTU (0) to match IPv6 MTU (1280)
<error> [1437068831.306733] [platform/nm-linux-platform.c:2440] sysctl_set(): platform-linux: sysctl: failed to set '/proc/sys/net/ipv6/conf/wlp3s0/mtu' to '1472': (22) Invalid argument
Reported-by: Jan Alexander Steffens <jan.steffens@gmail.com>
https://bugzilla.gnome.org/show_bug.cgi?id=752508
(cherry picked from commit a92d8b0c67)
Try to set the MTU of the parent Ethernet interface to match the
requested PPP MTU and MRU. This allows the negotiation of a PPP MTU
and MRU greater than 1492.
(cherry picked from commit 1d3eff45d2)
Commit d51975e changed, that we treat assumed and non-assumed
connections the same with respect to the default route. This is
certainly wrong, if we have an nm-generated-assumed connection
at hand. In this case, NM just generated a connection based on what
was configured on the system. Looking at that result and re-enforcing
the default-route is wrong.
We want to manage the default-route for assumed, persistent connections.
If the connection was assumed and generated, we do not.
This commit reverts d51975ed for nm-generated-assumed connection and
restores the previous behavior.
https://bugzilla.redhat.com/show_bug.cgi?id=1244483
Fixes: d51975ed92
(cherry picked from commit bebeff69e8)
Rename "default_route.v4_configure_first_time" to "v4_commit_first_time".
For one, the name "commit" matches better to the @commit variable in ip4_config_merge_and_apply()
and ip6_config_merge_and_apply(). Then, we don't need this information
only for default-routes, so move the variable out of the @default_route
struct.
(cherry picked from commit ad03cdbc73)
The metered property of a NMDevice that reaches the activated state is
copied from the active connection and if its value is 'unknown' some
heuristics are used to guess the actual value.
When the connection is torn down the metered property is reset to
'unknown'.
(cherry picked from commit a86255a043)
This makes wifi preferred to wwan (the modem and bluetooth device types
to be specific) by default, so that users that care about being
connected at all times can keep both enabled with auto-connect. As wifi
is usually unmetered and often faster than wwan, it makes sense to
prefer it. This is also how pretty much every smart-phone in the world
behaves, so it aligns better with user expectations too.
https://bugzilla.gnome.org/show_bug.cgi?id=744754
(cherry picked from commit 3efb6740fc)
In most cases, when syncing routes, we should only remove routes
that were configured by us previously. Otherwise, there is a race
that we can remove routes added externally.
Now, when applying IP configuration for a device, only do a full-sync
at the first time when we activate the device. Later on, only remove
routes that were added by us.
(cherry picked from commit 8b015826e6)
We currently don't manage a veth inside a container despite we should because
it's an externally configured software interface and thus waits for IFF_UP.
Given veths are prevented from being managed outside of a container by an udev
rule anyway it's safe to lift the external IFF_UP requirement for them.
(cherry picked from commit bcc79cc0eb)
