Broken by 25387cd1ff
When an activation request comes in via D-Bus for a slave, the
slave device's priv->master is set in stage1 in master_ready_cb().
Then nm_device_bring_up() is called on the slave, which triggers
link_changed_cb() and device_link_changed(). That then executes
this code:
if (priv->master)
nm_device_enslave_slave (priv->master, self, NULL);
which enslaves the slave, but due to the NULL will not configure
the slave.
This code was only meant to be run for externally triggered
master/slave changes.
NetworkManager[30304]: <info> (virbr0): bridge port virbr0-nic was detached
NetworkManager[30304]: (devices/nm-device.c:962):nm_device_release_one_slave: runtime check failed: (reason == NM_DEVICE_STATE_REASON_NONE)
NetworkManager[30304]: <info> (virbr0-nic): released from master virbr0
If the slave is removed, then the master is already cleaned up so NM
doesn't need to do anything. 5dd48f fixed that but forgot to update
the !configure case, causing the warning but no other problems.
Fixes: 5dd48f7527
When the device decides it needs re-auth during IP config and returns
to the NEED_AUTH state, make sure we clean up any half-done IP operations
since they will be re-started after auth is completed and the
IP_CONFIG state is re-entered.
https://bugzilla.gnome.org/show_bug.cgi?id=741342
platform/nm-linux-platform.c: In function 'setup':
platform/nm-linux-platform.c:4364:2: error: 'object' undeclared (first use in this function)
object = nl_cache_get_first (priv->link_cache);
^
Fixes 2b8060b9b3
This reverts commit efd09845c4.
It turns out that the socket space might not be the only buffer that may get
too full. 128K ought to be enough for it and we should resynchronize with the
kernel now if needed.
Kernel can return ENOBUFS in variety of reasons. If that happens, we know we've
lost events and should pick up kernel state.
Simple reproducer that triggers an ENOBUFS condition no matter how big our
netlink socket buffer is:
ip link add bridge0 type bridge
for i in seq $(0 1023); do ip link add dummy$i type dummy; \
ip link set dummy$i master bridge0; done
ip link del bridge0
We assume that in nm_nl_cache_search() and correctly set that in
get_kernel_object(), but we rtnl_link_alloc_cache() can initialize the cache
with devices of other families.
The consequence is that we don't notify when the bridge changes to IFF_UP as we
fail to match and remove the old downed object from the cache:
nm_device_bring_up(): [0xf506c0] (bridge0): bringing up device.
nm_platform_link_set_up(): link: setting up 'bridge0' (12)
link_change_flags(): link: change 12: flags set 'up' (1)
get_kernel_object(): get_kernel_object for link: bridge0 (12, family 7)
log_link(): signal: link added: 12: bridge0 <UP> mtu 1500 bridge driver 'bridge' udi '/sys/devices/virtual/net/bridge0'
get_kernel_object(): get_kernel_object for link: bridge0 (12, family 7)
log_link(): signal: link changed: 12: bridge0 <UP> mtu 1500 bridge driver 'bridge' udi '/sys/devices/virtual/net/bridge0'
log_link(): signal: link changed: 12: bridge0 <UP> mtu 1500 bridge driver 'bridge' udi '/sys/devices/virtual/net/bridge0'
(bridge0): device not up after timeout!
(bridge0): preparing device
If we didn't start IPv4 and IPv6, but they're allowed to fail, progress
the activation without failing it. Also, progress assumed connections to
check-ip with whatever configuration that is available.
https://bugzilla.redhat.com/show_bug.cgi?id=1141264
VPN connections always return true for nm_connection_need_secrets(), but the
documented behavior of GetSecrets() is just to return any secrets we have
(otherwise nmcli c --show-secrets would not be useful for VPN connections).
Since f32075d2fc, we remove the kernel
added IPv4 device route, and re-add it with appropriate metric.
This could potentially replace existing, conflicting routes. Be more
careful and only take any action when we don't have a conflicting
route and when we add the address for the first time.
The motivation for this was libreswan which might install a VPN route
for a subnet that we also have configured on an interface. But the route
conflict could happen easily for other reasons, for example if you
configure a conflicting route manually.
Don't replace the device route if we have any indication that
a conflict could arise.
https://bugzilla.gnome.org/show_bug.cgi?id=723178
Coverity:
Defect type: CONSTANT_EXPRESSION_RESULT
/src/NetworkManagerUtils.c:726: result_independent_of_operands: "18446744073709551615UL /* 9223372036854775807L * 2UL + 1UL */ < (gulong)sleep_duration_msec * 1000UL" is always false regardless of the values of its operands. This occurs as the logical first operand of '?:'.
because either GSM or CDMA is present. It is checked just above.
Coverity:
Defect type: DEADCODE
src/devices/bluetooth/nm-device-bt.c:312: dead_error_line: Execution cannot reach this statement: "fallback_prefix = dcgettext...".
Coverity:
Defect type: UNINIT
src/NetworkManagerUtils.c:1906: uninit_use_in_call: Using uninitialized value "tp.tv_nsec" when calling "monotonic_timestamp_get".
src/NetworkManagerUtils.c:1879: uninit_use_in_call: Using uninitialized value "tp.tv_nsec" when calling "monotonic_timestamp_get".
src/NetworkManagerUtils.c:1852: uninit_use_in_call: Using uninitialized value "tp.tv_nsec" when calling "monotonic_timestamp_get".
src/NetworkManagerUtils.c:1825: uninit_use_in_call: Using uninitialized value "tp.tv_nsec" when calling "monotonic_timestamp_get".
Coverity:
Defect type: CONSTANT_EXPRESSION_RESULT
src/NetworkManagerUtils.c:1978: same_on_both_sides: "(v1->diff_result & NM_SETTING_DIFF_RESULT_IN_B) != (v1->diff_result & NM_SETTING_DIFF_RESULT_IN_B)" is always false regardless of the values of its operands because those operands are identical. This occurs as the logical operand of if.
In several cases, connection uuids are generated based on
some strings. Change the algorithm, to prefix the hashed
identifier differently for each setting type. This makes
collisions very unlikely.
Also, change the algorithm, to create proper Variant3 UUIDs.
This is a behavioral change, but it only affects code places
that were added since nm-0-9-10 and were not yet part of
a stable release.
There are different types (variants) of UUIDs defined.
Especially variants 3 and 5 are name based variants (rfc4122).
The way we create our UUIDs in nm_utils_uuid_generate_from_string()
however does not create them according to RFC and does not set
the flags to indicate the variant.
Modify the signature of nm_utils_uuid_generate_from_string() to accept
a "uuid_type" argument, so that we later can add other algorithms without
breaking API.
nm_device_removed() calls nm_device_release_one_slave() in order to
fix up NetworkManager's master/slave bookkeeping, but we don't want to
tell the kernel to actually unslave the device, since that would
happen automatically anyway if the device was really removed, and
shouldn't happen if the device is just being removed from NM's device
list.
(In particular, don't remove all libvirt-created virtual network
devices from virbr0 when NetworkManager exits.)
nm_device_get_hw_address() may return NULL and nm_platform_link_get_type may
return NM_LINK_TYPE_NONE. While it might be a good idea to check for such cases
at the init time it seems easier to just ignore it and prevent blowing up in
subsequent deactivation.
A quick test case:
# while :; do ip link add moo0 type veth peer moo1; ip link del moo0 ; done
Yields:
NetworkManager:ERROR:devices/nm-device-ethernet.c:268:constructor:
assertion failed: (link_type == NM_LINK_TYPE_ETHERNET ||
link_type == NM_LINK_TYPE_VETH)
nm_device_set_hw_addr: assertion 'addr != NULL' failed
https://bugzilla.gnome.org/show_bug.cgi?id=740992
For IPv4, iproute for example defaults to a metric of 0.
Hence, the name NM_PLATFORM_ROUTE_METRIC_DEFAULT was misleading.
Also add a NM_PLATFORM_ROUTE_METRIC_DEFAULT_IP4 define for completeness.
https://bugzilla.gnome.org/show_bug.cgi?id=740780
We recently changed default values for route metrics. Revise that
again and increase the space between the default values.
No strong reason to do this, but it seems better to have larger
gaps and make use of the available range.
The parent device is just the device that happens to be the best device.
It does not mean, that its route metric should be inherited to the VPN
connection.
This also makes the resulting route metric for VPN connections much
more predictable: now it is either ipv4.route-metric from the connection,
or it falls back to NM_VPN_ROUTE_METRIC_DEFAULT (10).
Also read the command line argument as G_OPTION_ARG_INT64 type. Otherwise, on
32-bit integers, you could not enter any number larger then G_MAXINT32.
Signed-off-by: Thomas Haller <thaller@redhat.com>
For IPv4 addresses, the kernel automatically adds a route when
configuring an IP address. Unfortunately, there is no way to control
this behavior or to set the route metric.
Fix this, by adding our own route and removing the kernel provided
one.
Note that this adds a major change in that we no longer call
nm_ip4_config_commit() for assumed devices.
https://bugzilla.gnome.org/show_bug.cgi?id=723178
Signed-off-by: Thomas Haller <thaller@redhat.com>
NMIPRoute is used by NMSettingIPConfig, but also
NMIPConfig. In the former case, default routes are (still)
disallowed. But in the NMIPConfig use-case, it can make sense
to expose default routes as NMIPRoute instances.
Relax the restriction on the NMIPRoute API to allow this
future change.
No code actually supports having NMIPRoute instances with
prefix length zero (default routes). Up to now, all such uses
would be a bug.
https://bugzilla.gnome.org/show_bug.cgi?id=739969
Signed-off-by: Thomas Haller <thaller@redhat.com>
The previous nm-dhcp-systemd code for logging the lease expiry time,
and exporting that value to D-Bus was clamping the value to
G_MAXUINT32-1, but that's unnecessary on x86_64, and incorrect on x86
(since time_t is signed).
Correctly adding a value to the current time and not overflowing seems
to be more-or-less impossible without having separate cases for 4- and
8-byte time_t. Since this was basically just for logging purposes
anyway, just log the number of seconds rather than the timestamp, and
then we don't have to worry about sizeof(time_t).
