Fix problem when switch provider from initscript to nm

Problem:

After `tests_bridge_initscripts.yml` passed, the `tests_bridge_nm.yml`
will fail with NetworkManager 1.18.

Root cause:

 1. The `absent` and `down` action of initscript provider will not
    remove the bridge interface which fail the assertion in
    `tests_bridge_nm.yml`.
 2. In initscript mode, network role will create ifcfg file with
    `NM_CONTROLLED=no` instructing NetworkManager to mark the bridge as
    unmanaged. The follow up `down` and `absent` action of initscript
    provider will not change the NetworkManager's understanding on
    unmanaged state of this interface.

Fixes:
 1. We cannot change existing behaviour of initscript on not deleting
    interface in `down` and `absent` action. So we change the test
    function `tests/playbooks/down_profile.yml` to delete the interface
    manually via `ip link del <ifname>` command.

 2. Use `NM.Client.reload_connections_async()` to reload the
    configuration for nm provider on NetworkManager 1.18.

Previous test infrastructure is running each test file in a brand new VM
or container which cause this problem not been found before.

Dedicate test case `tests/tests_switch_provider.yml` included.

Signed-off-by: Gris Ge <fge@redhat.com>

.github/run_test.sh

@@ -8,10 +8,13 @@ C8_CONTAINER_IMAGE="quay.io/linux-system-roles/c8-network-role"
 C7_CONTAINER_IMAGE="quay.io/linux-system-roles/c7-network-role"
 PODMAN_OPTS="--systemd=true --privileged"
 
+# FIXME: test_wireless has been removed because is not supported on CI
+# container EL7. We need to add this back for EL8 or greater.
 read -r -d '' TEST_FILES << EOF || :
 tests_802_1x_nm.yml
 tests_bond_nm.yml
 tests_auto_gateway_nm.yml
+tests_bridge_initscripts.yml
 tests_bridge_nm.yml
 tests_change_indication_on_repeat_run.yml
 tests_dummy_nm.yml
@@ -22,6 +25,7 @@ tests_ethtool_ring_nm.yml
 tests_ipv6_disabled_nm.yml
 tests_ipv6_nm.yml
 tests_vlan_mtu_nm.yml
+tests_switch_provider.yml
 EOF
 
 EXEC_PATH=$(dirname "$(realpath "$0")")

library/network_connections.py

@@ -2055,6 +2055,16 @@ class Cmd_nm(Cmd):
                 len(self.connections) * DEFAULT_ACTIVATION_TIMEOUT
             )
 
+        # On NetworkManger 1.18, If user switch from initscripts provider where
+        # NM_CONTROLLED=no defined in ifcfg-ethX file, NetworkManager daemon will treat
+        # that interface as strictly unmanaged, even the follow up deletion of
+        # ifcfg-ethX file cannot change the NetworManager's unmanaged state of this
+        # interface. This will prevent any follow up "nm" provider action on this
+        # interface.  To solve that, we instruct NetworkManager to reload the
+        # configuration.
+        if self._nm_provider.get_client_version().startswith("1.18."):
+            self._nm_provider.reload_configuration()
+
     def rollback_transaction(self, idx, action, error):
         Cmd.rollback_transaction(self, idx, action, error)
         self.on_failure()

module_utils/network_lsr/nm/provider.py

@@ -60,3 +60,31 @@ class NetworkManagerProvider:
     def get_connections(self):
         nm_client = client.get_client()
         return nm_client.get_connections()
+
+    def get_client_version(self):
+        nm_client = client.get_client()
+        return nm_client.get_version()
+
+    def reload_configuration(self):
+        timeout = 10
+        nm_client = client.get_client()
+        main_loop = client.get_mainloop(timeout)
+        logging.debug("Reloading configuration with timeout %s", timeout)
+        nm_client.reload_connections_async(
+            main_loop.cancellable, _reload_config_callback, main_loop
+        )
+        main_loop.run()
+
+
+def _reload_config_callback(nm_client, result, main_loop):
+    try:
+        success = nm_client.reload_connections_finish(result)
+    except client.GLib.Error as e:
+        logging.warn("Failed to reload configuration: %s", e)
+        main_loop.quit()
+        return
+    if success:
+        logging.debug("Reloading configuration finished")
+    else:
+        logging.warn("Failed to reload configuration, no error message")
+    main_loop.quit()

tests/ensure_provider_tests.py

@@ -132,6 +132,7 @@ NM_CONDITIONAL_TESTS = {
 IGNORE = [
     # checked by tests_regression_nm.yml
     "playbooks/tests_checkpoint_cleanup.yml",
+    "playbooks/tests_switch_provider.yml",
 ]
 
 RUN_PLAYBOOK_WITH_INITSCRIPTS = """# SPDX-License-Identifier: BSD-3-Clause

tests/playbooks/down_profile+delete_interface.yml

@@ -0,0 +1,7 @@
+# SPDX-License-Identifier: BSD-3-Clause
+---
+- import_playbook: down_profile.yml
+- name: Delete the interface when the network provider is initscripts
+  hosts: all
+  tasks:
+    - include_tasks: tasks/delete_interface.yml

tests/playbooks/tests_switch_provider.yml

@@ -0,0 +1,66 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# This file was generated by ensure_provider_tests.py
+---
+# set network provider and gather facts
+- hosts: all
+  name: Switch initscripts provider to nm
+  vars:
+    interface: LSR-TST-br34
+  tasks:
+    - name: set fact to use initscripts network_provider
+      set_fact:
+        network_provider: initscripts
+      tags:
+        - always
+    - name: "Create test bridge {{ interface }} via initscripts provider"
+      include_role:
+        name: linux-system-roles.network
+      vars:
+        network_connections:
+          - name: "{{ interface }}"
+            state: up
+            type: bridge
+            ip:
+              dhcp4: false
+              auto6: false
+    - include_tasks: tasks/assert_device_present.yml
+    - name: "Take the profile {{ interface }} down via initscripts provider"
+      include_role:
+        name: linux-system-roles.network
+      vars:
+        network_connections:
+          - name: "{{ interface }}"
+            persistent_state: absent
+            state: down
+            type: bridge
+    # The initscripts should not remove the interface for down/absent
+    - include_tasks: tasks/assert_device_present.yml
+    - name: set fact to use nm network_provider
+      set_fact:
+        network_provider: nm
+      tags:
+        - always
+    - name: "Create test bridge {{ interface }} via nm provider"
+      include_role:
+        name: linux-system-roles.network
+      vars:
+        network_connections:
+          - name: "{{ interface }}"
+            state: up
+            type: bridge
+            ip:
+              dhcp4: false
+              auto6: false
+    - include_tasks: tasks/assert_device_present.yml
+    - name: "Remove bridge {{ interface }} via nm provider"
+      include_role:
+        name: linux-system-roles.network
+      vars:
+        network_connections:
+          - name: "{{ interface }}"
+            state: down
+            type: bridge
+    # NetworkManager should not remove pre-exist interface for down/absent
+    - include_tasks: tasks/assert_device_present.yml
+    - include_tasks: tasks/delete_interface.yml
+    - include_tasks: tasks/assert_device_absent.yml

tests/tests_switch_provider.yml

@@ -0,0 +1,10 @@
+# SPDX-License-Identifier: BSD-3-Clause
+---
+- hosts: all
+  name: Run playbook 'playbooks/tests_switch_provider.yml'
+
+- import_playbook: playbooks/tests_switch_provider.yml
+  when:
+    # The test requires or should run with NetworkManager, therefore it cannot
+    # run on RHEL/CentOS 6
+    - ansible_distribution_major_version != '6'